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 EC_GROUP_to_params which creates an OSSL_PARAM array
29 from a given EC_GROUP.
33 * Added support for pluggable (provider-based) TLS signature algorithms.
34 This enables TLS 1.3 authentication operations with algorithms embedded
35 in providers not included by default in OpenSSL. In combination with
36 the already available pluggable KEM and X.509 support, this enables
37 for example suitable providers to deliver post-quantum or quantum-safe
38 cryptography to OpenSSL users.
42 * Added support for Hybrid Public Key Encryption (HPKE) as defined
43 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
44 Message Layer Security (MLS) and other IETF specifications.
45 HPKE can also be used by other applications that require
46 encrypting "to" an ECDH public key. External APIs are defined in
47 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
51 * Add support for certificate compression (RFC8879), including
52 library support for Brotli and Zstandard compression.
56 * Add the ability to add custom attributes to PKCS12 files. Add a new API
57 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
58 for a user specified callback and optional argument.
59 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
60 added to the existing STACK_OF attrs.
64 * Major refactor of the libssl record layer
68 * Added a new BIO_s_dgram_mem() to read/write datagrams to memory
72 * Add a mac salt length option for the pkcs12 command.
76 * Add more SRTP protection profiles from RFC8723 and RFC8269.
80 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
82 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
84 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
85 supported and enabled.
89 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
90 to the list of ciphersuites providing Perfect Forward Secrecy as
91 required by SECLEVEL >= 3.
93 *Dmitry Belyavskiy, Nicola Tuveri*
95 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
96 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
97 SSL_get0_iana_groups() function-like macro, retrieves the list of
98 supported groups sent by the peer.
99 The function SSL_client_hello_get_extension_order() populates
100 a caller-supplied array with the list of extension types present in the
101 ClientHello, in order of appearance.
105 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
106 to make it possible to use empty passphrase strings.
110 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
111 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
112 of 160 bits and above and less than 224 bits were previously accepted by
113 default but are now no longer allowed. By default TLS compression was
114 already disabled in previous OpenSSL versions. At security level 2 it cannot
119 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
124 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
125 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
126 will need to load the legacy crypto provider.
129 * CCM8 cipher suites in TLS have been downgraded to security level zero
130 because they use a short authentication tag which lowers their strength.
134 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
139 * Add X.509 certificate codeSigning purpose and related checks on key usage and
140 extended key usage of the leaf certificate according to the CA/Browser Forum.
144 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
145 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
146 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
147 X.509 version 3 if the certificate information includes X.509 extensions.
151 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
152 such as adding a trace facility for debugging certificate chain building.
156 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
157 in particular supporting requests for central key generation, generalized
158 polling, and various types of genm/genp exchanges defined in CMP Updates.
162 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
163 like correcting the TLS and proxy support and adding tracing for debugging.
167 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
171 * `CMS_add0_cert()` and `CMS_add1_cert()` no more throw an error
172 if a certificate to be added is already present.
173 * `CMS_sign_ex()` and `CMS_sign()` now ignore any duplicate certificates
174 in their `certs` argument and no longer throw an error for them.
178 * Fixed and extended `util/check-format.pl` for checking adherence to the
179 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
180 The checks are meanwhile more complete and yield fewer false positives.
184 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
185 sending and receiving multiple messages in a single call. An implementation
186 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
190 * The `SSL_CERT_PATH` and `SSL_CERT_URI` environment variables are introduced.
191 `SSL_CERT_URI` can be used to specify a URI for a root certificate store. The
192 `SSL_CERT_PATH` environment variable specifies a delimiter-separated list of
193 paths which are searched for root certificates.
195 The existing `SSL_CERT_DIR` environment variable is deprecated.
196 `SSL_CERT_DIR` was previously used to specify either a delimiter-separated
197 list of paths or an URI, which is ambiguous. Setting `SSL_CERT_PATH` causes
198 `SSL_CERT_DIR` to be ignored for the purposes of determining root certificate
199 directories, and setting `SSL_CERT_URI` causes `SSL_CERT_DIR` to be ignored
200 for the purposes of determining root certificate stores.
204 * Support for loading root certificates from the Windows certificate store
205 has been added. The support is in the form of a store which recognises the
206 URI string of `org.openssl.winstore://`. This store is enabled by default and
207 can be disabled using the new compile-time option `no-winstore`.
211 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
212 kernel versions that support KTLS have a known bug in CCM processing. That
213 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
214 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
219 * Zerocopy KTLS sendfile() support on Linux.
223 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
224 decryption as a protection against Bleichenbacher-like attacks.
225 The RSA decryption API will now return a randomly generated deterministic
226 message instead of an error in case it detects an error when checking
227 padding during PKCS#1 v1.5 decryption. This is a general protection against
228 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
230 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
231 on the RSA decryption context.
238 ### Changes between 3.0 and 3.1.0 [xx XXX xxxx]
240 * Add FIPS provider configuration option to enforce the
241 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
242 The option '-ems-check' can optionally be supplied to
243 'openssl fipsinstall'.
247 * The FIPS provider includes a few non-approved algorithms for
248 backward compatibility purposes and the "fips=yes" property query
249 must be used for all algorithm fetches to ensure FIPS compliance.
251 The algorithms that are included but not approved are Triple DES ECB,
252 Triple DES CBC and EdDSA.
256 * Added support for KMAC in KBKDF.
260 * RNDR and RNDRRS support in provider functions to provide
261 random number generation for Arm CPUs (aarch64).
265 * s_client and s_server apps now explicitly say when the TLS version
266 does not include the renegotiation mechanism. This avoids confusion
267 between that scenario versus when the TLS version includes secure
268 renegotiation but the peer lacks support for it.
272 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
274 *Tomasz Kantecki, Andrey Matyukov*
276 * The various OBJ_* functions have been made thread safe.
280 * Parallel dual-prime 1536/2048-bit modular exponentiation for
281 AVX512_IFMA capable processors.
283 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
285 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
286 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
287 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
288 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
289 `OPENSSL_NO_DEPRECATED_3_1`.
291 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
292 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
293 definitions for these functions regardless of whether
294 `OPENSSL_NO_DEPRECATED_3_1` is defined.
296 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
297 functions regardless of whether they are using them. It is recommended that
298 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
302 * When generating safe-prime DH parameters set the recommended private key
303 length equivalent to minimum key lengths as in RFC 7919.
307 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
308 maximum size that is smaller or equal to the digest length to comply with
309 FIPS 186-4 section 5. This is implemented by a new option
310 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
311 `rsa_pss_saltlen` parameter, which is now the default. Signature
312 verification is not affected by this change and continues to work as before.
319 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
320 listed here are only a brief description.
321 The migration guide contains more detailed information related to new features,
322 breaking changes, and mappings for the large list of deprecated functions.
324 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
326 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
328 * Fixed NULL dereference during PKCS7 data verification.
330 A NULL pointer can be dereferenced when signatures are being
331 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
332 algorithm used for the signature is known to the OpenSSL library but
333 the implementation of the hash algorithm is not available the digest
334 initialization will fail. There is a missing check for the return
335 value from the initialization function which later leads to invalid
336 usage of the digest API most likely leading to a crash.
339 PKCS7 data is processed by the SMIME library calls and also by the
340 time stamp (TS) library calls. The TLS implementation in OpenSSL does
341 not call these functions however third party applications would be
342 affected if they call these functions to verify signatures on untrusted
347 * Fixed X.400 address type confusion in X.509 GeneralName.
349 There is a type confusion vulnerability relating to X.400 address processing
350 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
351 but the public structure definition for GENERAL_NAME incorrectly specified
352 the type of the x400Address field as ASN1_TYPE. This field is subsequently
353 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
356 When CRL checking is enabled (i.e. the application sets the
357 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
358 pass arbitrary pointers to a memcmp call, enabling them to read memory
359 contents or enact a denial of service.
364 * Fixed NULL dereference validating DSA public key.
366 An invalid pointer dereference on read can be triggered when an
367 application tries to check a malformed DSA public key by the
368 EVP_PKEY_public_check() function. This will most likely lead
369 to an application crash. This function can be called on public
370 keys supplied from untrusted sources which could allow an attacker
371 to cause a denial of service attack.
373 The TLS implementation in OpenSSL does not call this function
374 but applications might call the function if there are additional
375 security requirements imposed by standards such as FIPS 140-3.
378 *Shane Lontis, Tomáš Mráz*
380 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
382 An invalid pointer dereference on read can be triggered when an
383 application tries to load malformed PKCS7 data with the
384 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
386 The result of the dereference is an application crash which could
387 lead to a denial of service attack. The TLS implementation in OpenSSL
388 does not call this function however third party applications might
389 call these functions on untrusted data.
394 * Fixed Use-after-free following BIO_new_NDEF.
396 The public API function BIO_new_NDEF is a helper function used for
397 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
398 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
399 be called directly by end user applications.
401 The function receives a BIO from the caller, prepends a new BIO_f_asn1
402 filter BIO onto the front of it to form a BIO chain, and then returns
403 the new head of the BIO chain to the caller. Under certain conditions,
404 for example if a CMS recipient public key is invalid, the new filter BIO
405 is freed and the function returns a NULL result indicating a failure.
406 However, in this case, the BIO chain is not properly cleaned up and the
407 BIO passed by the caller still retains internal pointers to the previously
408 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
409 then a use-after-free will occur. This will most likely result in a crash.
412 *Viktor Dukhovni, Matt Caswell*
414 * Fixed Double free after calling PEM_read_bio_ex.
416 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
417 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
418 data. If the function succeeds then the "name_out", "header" and "data"
419 arguments are populated with pointers to buffers containing the relevant
420 decoded data. The caller is responsible for freeing those buffers. It is
421 possible to construct a PEM file that results in 0 bytes of payload data.
422 In this case PEM_read_bio_ex() will return a failure code but will populate
423 the header argument with a pointer to a buffer that has already been freed.
424 If the caller also frees this buffer then a double free will occur. This
425 will most likely lead to a crash.
427 The functions PEM_read_bio() and PEM_read() are simple wrappers around
428 PEM_read_bio_ex() and therefore these functions are also directly affected.
430 These functions are also called indirectly by a number of other OpenSSL
431 functions including PEM_X509_INFO_read_bio_ex() and
432 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
433 internal uses of these functions are not vulnerable because the caller does
434 not free the header argument if PEM_read_bio_ex() returns a failure code.
437 *Kurt Roeckx, Matt Caswell*
439 * Fixed Timing Oracle in RSA Decryption.
441 A timing based side channel exists in the OpenSSL RSA Decryption
442 implementation which could be sufficient to recover a plaintext across
443 a network in a Bleichenbacher style attack. To achieve a successful
444 decryption an attacker would have to be able to send a very large number
445 of trial messages for decryption. The vulnerability affects all RSA padding
446 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
449 *Dmitry Belyavsky, Hubert Kario*
451 * Fixed X.509 Name Constraints Read Buffer Overflow.
453 A read buffer overrun can be triggered in X.509 certificate verification,
454 specifically in name constraint checking. The read buffer overrun might
455 result in a crash which could lead to a denial of service attack.
456 In a TLS client, this can be triggered by connecting to a malicious
457 server. In a TLS server, this can be triggered if the server requests
458 client authentication and a malicious client connects.
463 * Fixed X.509 Policy Constraints Double Locking security issue.
465 If an X.509 certificate contains a malformed policy constraint and
466 policy processing is enabled, then a write lock will be taken twice
467 recursively. On some operating systems (most widely: Windows) this
468 results in a denial of service when the affected process hangs. Policy
469 processing being enabled on a publicly facing server is not considered
470 to be a common setup.
475 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
476 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
477 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
478 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
479 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
480 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
481 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
482 for legacy EC and SM2 keys is also changed similarly to honor the
483 equivalent conversion format flag as specified in the underlying
484 `EC_KEY` object being exported to a provider, when this function is
485 called through `EVP_PKEY_export()`.
489 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
491 * Fixed two buffer overflows in punycode decoding functions.
493 A buffer overrun can be triggered in X.509 certificate verification,
494 specifically in name constraint checking. Note that this occurs after
495 certificate chain signature verification and requires either a CA to
496 have signed the malicious certificate or for the application to continue
497 certificate verification despite failure to construct a path to a trusted
500 In a TLS client, this can be triggered by connecting to a malicious
501 server. In a TLS server, this can be triggered if the server requests
502 client authentication and a malicious client connects.
504 An attacker can craft a malicious email address to overflow
505 an arbitrary number of bytes containing the `.` character (decimal 46)
506 on the stack. This buffer overflow could result in a crash (causing a
510 An attacker can craft a malicious email address to overflow four
511 attacker-controlled bytes on the stack. This buffer overflow could
512 result in a crash (causing a denial of service) or potentially remote code
513 execution depending on stack layout for any given platform/compiler.
518 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
519 parameters in OpenSSL code.
520 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
521 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
522 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
523 Using these invalid names may cause algorithms to use slower methods
524 that ignore the CRT parameters.
528 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
533 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
534 data to be signed before signing the certificate.
538 * Added RIPEMD160 to the default provider.
542 * Ensured that the key share group sent or accepted for the key exchange
543 is allowed for the protocol version.
547 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
549 * OpenSSL supports creating a custom cipher via the legacy
550 EVP_CIPHER_meth_new() function and associated function calls. This function
551 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
552 to use the new provider mechanism in order to implement custom ciphers.
554 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
555 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
556 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
557 and decryption initialisation functions). Instead of using the custom cipher
558 directly it incorrectly tries to fetch an equivalent cipher from the
559 available providers. An equivalent cipher is found based on the NID passed to
560 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
561 given cipher. However it is possible for an application to incorrectly pass
562 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
563 is used in this way the OpenSSL encryption/decryption initialisation function
564 will match the NULL cipher as being equivalent and will fetch this from the
565 available providers. This will succeed if the default provider has been
566 loaded (or if a third party provider has been loaded that offers this
567 cipher). Using the NULL cipher means that the plaintext is emitted as the
570 Applications are only affected by this issue if they call
571 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
572 encryption/decryption initialisation function. Applications that only use
573 SSL/TLS are not impacted by this issue.
578 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
583 * Fixed the linux-mips64 Configure target which was missing the
584 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
589 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
594 * Correctly handle a retransmitted ClientHello in DTLS
598 * Fixed detection of ktls support in cross-compile environment on Linux
602 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
607 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
608 report correct results in some cases
612 * Fix UWP builds by defining VirtualLock
616 * For known safe primes use the minimum key length according to RFC 7919.
617 Longer private key sizes unnecessarily raise the cycles needed to compute the
618 shared secret without any increase of the real security. This fixes a
619 regression from 1.1.1 where these shorter keys were generated for the known
624 * Added the loongarch64 target
628 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
629 only passed to the FIPS provider and not to the default or legacy provider.
633 * Fixed reported performance degradation on aarch64. Restored the
634 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
635 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
636 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
637 The new algorithm is still used for 32 bit targets.
641 * Added a missing header for memcmp that caused compilation failure on some
646 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
648 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
649 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
650 This issue makes the RSA implementation with 2048 bit private keys
651 incorrect on such machines and memory corruption will happen during
652 the computation. As a consequence of the memory corruption an attacker
653 may be able to trigger a remote code execution on the machine performing
656 SSL/TLS servers or other servers using 2048 bit RSA private keys running
657 on machines supporting AVX512IFMA instructions of the X86_64 architecture
658 are affected by this issue.
663 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
664 implementation would not encrypt the entirety of the data under some
665 circumstances. This could reveal sixteen bytes of data that was
666 preexisting in the memory that wasn't written. In the special case of
667 "in place" encryption, sixteen bytes of the plaintext would be revealed.
669 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
670 they are both unaffected.
673 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
675 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
677 * In addition to the c_rehash shell command injection identified in
678 CVE-2022-1292, further bugs where the c_rehash script does not
679 properly sanitise shell metacharacters to prevent command injection have been
682 When the CVE-2022-1292 was fixed it was not discovered that there
683 are other places in the script where the file names of certificates
684 being hashed were possibly passed to a command executed through the shell.
686 This script is distributed by some operating systems in a manner where
687 it is automatically executed. On such operating systems, an attacker
688 could execute arbitrary commands with the privileges of the script.
690 Use of the c_rehash script is considered obsolete and should be replaced
691 by the OpenSSL rehash command line tool.
694 *Daniel Fiala, Tomáš Mráz*
696 * Case insensitive string comparison no longer uses locales. It has instead
697 been directly implemented.
701 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
703 * Case insensitive string comparison is reimplemented via new locale-agnostic
704 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
705 comparison. The previous implementation had problems when the Turkish locale
710 * Fixed a bug in the c_rehash script which was not properly sanitising shell
711 metacharacters to prevent command injection. This script is distributed by
712 some operating systems in a manner where it is automatically executed. On
713 such operating systems, an attacker could execute arbitrary commands with the
714 privileges of the script.
716 Use of the c_rehash script is considered obsolete and should be replaced
717 by the OpenSSL rehash command line tool.
722 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
723 certificate on an OCSP response. The bug caused the function in the case
724 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
725 response (meaning a successful verification) even in the case where the
726 response signing certificate fails to verify.
728 It is anticipated that most users of `OCSP_basic_verify` will not use the
729 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
730 a negative value (indicating a fatal error) in the case of a certificate
731 verification failure. The normal expected return value in this case would be
734 This issue also impacts the command line OpenSSL "ocsp" application. When
735 verifying an ocsp response with the "-no_cert_checks" option the command line
736 application will report that the verification is successful even though it
737 has in fact failed. In this case the incorrect successful response will also
738 be accompanied by error messages showing the failure and contradicting the
739 apparently successful result.
744 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
745 AAD data as the MAC key. This made the MAC key trivially predictable.
747 An attacker could exploit this issue by performing a man-in-the-middle attack
748 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
749 that the modified data would still pass the MAC integrity check.
751 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
752 endpoint will always be rejected by the recipient and the connection will
753 fail at that point. Many application protocols require data to be sent from
754 the client to the server first. Therefore, in such a case, only an OpenSSL
755 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
757 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
758 sent in both directions. In this case both clients and servers could be
759 affected, regardless of the application protocol.
761 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
762 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
763 the handshake when using this ciphersuite.
765 The confidentiality of data is not impacted by this issue, i.e. an attacker
766 cannot decrypt data that has been encrypted using this ciphersuite - they can
769 In order for this attack to work both endpoints must legitimately negotiate
770 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
771 OpenSSL 3.0, and is not available within the default provider or the default
772 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
773 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
774 following must have occurred:
776 1) OpenSSL must have been compiled with the (non-default) compile time option
777 enable-weak-ssl-ciphers
779 2) OpenSSL must have had the legacy provider explicitly loaded (either
780 through application code or via configuration)
782 3) The ciphersuite must have been explicitly added to the ciphersuite list
784 4) The libssl security level must have been set to 0 (default is 1)
786 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
788 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
789 others that both endpoints have in common
794 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
795 occupied by the removed hash table entries.
797 This function is used when decoding certificates or keys. If a long lived
798 process periodically decodes certificates or keys its memory usage will
799 expand without bounds and the process might be terminated by the operating
800 system causing a denial of service. Also traversing the empty hash table
801 entries will take increasingly more time.
803 Typically such long lived processes might be TLS clients or TLS servers
804 configured to accept client certificate authentication.
807 *Hugo Landau, Aliaksei Levin*
809 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
810 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
811 statistics are no longer supported. For compatibility, these statistics are
812 still listed in the output but are now always reported as zero.
816 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
818 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
819 for non-prime moduli.
821 Internally this function is used when parsing certificates that contain
822 elliptic curve public keys in compressed form or explicit elliptic curve
823 parameters with a base point encoded in compressed form.
825 It is possible to trigger the infinite loop by crafting a certificate that
826 has invalid explicit curve parameters.
828 Since certificate parsing happens prior to verification of the certificate
829 signature, any process that parses an externally supplied certificate may thus
830 be subject to a denial of service attack. The infinite loop can also be
831 reached when parsing crafted private keys as they can contain explicit
832 elliptic curve parameters.
834 Thus vulnerable situations include:
836 - TLS clients consuming server certificates
837 - TLS servers consuming client certificates
838 - Hosting providers taking certificates or private keys from customers
839 - Certificate authorities parsing certification requests from subscribers
840 - Anything else which parses ASN.1 elliptic curve parameters
842 Also any other applications that use the BN_mod_sqrt() where the attacker
843 can control the parameter values are vulnerable to this DoS issue.
848 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
849 to the list of ciphersuites providing Perfect Forward Secrecy as
850 required by SECLEVEL >= 3.
852 *Dmitry Belyavskiy, Nicola Tuveri*
854 * Made the AES constant time code for no-asm configurations
855 optional due to the resulting 95% performance degradation.
856 The AES constant time code can be enabled, for no assembly
857 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
861 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
866 * The negative return value handling of the certificate verification callback
867 was reverted. The replacement is to set the verification retry state with
868 the SSL_set_retry_verify() function.
872 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
874 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
875 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
876 verify a certificate supplied by a server. That function may return a
877 negative return value to indicate an internal error (for example out of
878 memory). Such a negative return value is mishandled by OpenSSL and will cause
879 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
880 success and a subsequent call to SSL_get_error() to return the value
881 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
882 returned by OpenSSL if the application has previously called
883 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
884 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
885 totally unexpected and applications may not behave correctly as a result. The
886 exact behaviour will depend on the application but it could result in
887 crashes, infinite loops or other similar incorrect responses.
889 This issue is made more serious in combination with a separate bug in OpenSSL
890 3.0 that will cause X509_verify_cert() to indicate an internal error when
891 processing a certificate chain. This will occur where a certificate does not
892 include the Subject Alternative Name extension but where a Certificate
893 Authority has enforced name constraints. This issue can occur even with valid
899 * Corrected a few file name and file reference bugs in the build,
900 installation and setup scripts, which lead to installation verification
901 failures. Slightly enhanced the installation verification script.
905 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
910 * Fixed PVK encoder to properly query for the passphrase.
914 * Multiple fixes in the OSSL_HTTP API functions.
918 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
919 OSSL_PARAM_INTEGER data type and return error on negative numbers
920 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
921 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
925 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
929 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
933 * Multiple threading fixes.
937 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
941 * Allow fetching an operation from the provider that owns an unexportable key
942 as a fallback if that is still allowed by the property query.
946 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
948 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
953 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
954 S390X capability vector to zero. This simplifies testing of different code
955 paths on S390X architecture.
959 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
960 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
961 SP 800-38D". The communication will fail at this point.
965 * The EC_GROUP_clear_free() function is deprecated as there is nothing
966 confidential in EC_GROUP data.
970 * The byte order mark (BOM) character is ignored if encountered at the
971 beginning of a PEM-formatted file.
975 * Added CMS support for the Russian GOST algorithms.
979 * Due to move of the implementation of cryptographic operations
980 to the providers, validation of various operation parameters can
981 be postponed until the actual operation is executed where previously
982 it happened immediately when an operation parameter was set.
984 For example when setting an unsupported curve with
985 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
986 fail but later keygen operations with the EVP_PKEY_CTX will fail.
988 *OpenSSL team members and many third party contributors*
990 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
991 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
992 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
993 instead to retrieve these algorithms from a provider.
997 * On build targets where the multilib postfix is set in the build
998 configuration the libdir directory was changing based on whether
999 the lib directory with the multilib postfix exists on the system
1000 or not. This unpredictable behavior was removed and eventual
1001 multilib postfix is now always added to the default libdir. Use
1002 `--libdir=lib` to override the libdir if adding the postfix is
1007 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1008 no longer interoperable with OpenSSL 1.1.1.
1012 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1013 function codes, this function can only cause problems for calling
1018 * Add a configurable flag to output date formats as ISO 8601. Does not
1019 change the default date format.
1023 * Version of MSVC earlier than 1300 could get link warnings, which could
1024 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1025 Support for this flag has been removed.
1029 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1030 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1031 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1032 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1033 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1037 * The signatures of the functions to get and set options on SSL and
1038 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1039 Some source code changes may be required.
1043 * The public definitions of conf_method_st and conf_st have been
1044 deprecated. They will be made opaque in a future release.
1046 *Rich Salz and Tomáš Mráz*
1048 * Client-initiated renegotiation is disabled by default. To allow it, use
1049 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1050 flag, or the "ClientRenegotiation" config parameter as appropriate.
1054 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1055 or modify relative pathname inclusion.
1059 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1060 validated. Please consult the README-FIPS and
1061 README-PROVIDERS files, as well as the migration guide.
1063 *OpenSSL team members and many third party contributors*
1065 * For the key types DH and DHX the allowed settable parameters are now different.
1069 * The openssl commands that read keys, certificates, and CRLs now
1070 automatically detect the PEM or DER format of the input files.
1072 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1074 * Added enhanced PKCS#12 APIs which accept a library context.
1078 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1082 * Added support for Kernel TLS (KTLS).
1084 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1086 * Support for RFC 5746 secure renegotiation is now required by default for
1087 SSL or TLS connections to succeed.
1091 * The signature of the `copy` functional parameter of the
1092 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1093 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1094 the signature of the `pub_decode` functional parameter of the
1095 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1096 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1100 * The error return values from some control calls (ctrl) have changed.
1104 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1108 * Many functions in the EVP_ namespace that are getters of values from
1109 implementations or contexts were renamed to include get or get0 in their
1110 names. Old names are provided as macro aliases for compatibility and
1115 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1116 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1117 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1122 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1125 * The output from the command line applications may have minor
1130 * The output from numerous "printing" may have minor changes.
1134 * Windows thread synchronization uses read/write primitives (SRWLock) when
1135 supported by the OS, otherwise CriticalSection continues to be used.
1139 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1140 work on read only BIO source/sinks that do not support these functions.
1141 This allows piping or redirection of a file BIO using stdin to be buffered
1142 into memory. This is used internally in OSSL_DECODER_from_bio().
1146 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1147 this function would return one of the values OSSL_STORE_INFO_NAME,
1148 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1149 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1150 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1151 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1152 using this function should be amended to handle the changed return value.
1156 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1157 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1158 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1159 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1160 contains more than one certificate identifier: This means that all
1161 certificates referenced there MUST be part of the validation chain.
1165 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1166 RC5, DESX and DES have been moved to the legacy provider.
1170 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1171 RIPEMD-160 have been moved to the legacy provider.
1175 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1180 * The deprecated functions EVP_PKEY_get0_RSA(),
1181 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1182 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1183 well as the similarly named "get1" functions behave differently in
1188 * A number of functions handling low-level keys or engines were deprecated
1189 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1190 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1191 EVP_PKEY_get0_siphash().
1195 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1196 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1197 will need to load the legacy crypto provider. This includes these PBE
1198 algorithms which use this KDF:
1199 - NID_pbeWithMD2AndDES_CBC
1200 - NID_pbeWithMD5AndDES_CBC
1201 - NID_pbeWithSHA1AndRC2_CBC
1202 - NID_pbeWithMD2AndRC2_CBC
1203 - NID_pbeWithMD5AndRC2_CBC
1204 - NID_pbeWithSHA1AndDES_CBC
1208 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1209 BIO_debug_callback() functions.
1213 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1214 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1218 * The RAND_METHOD APIs have been deprecated.
1222 * The SRP APIs have been deprecated.
1226 * Add a compile time option to prevent the caching of provider fetched
1227 algorithms. This is enabled by including the no-cached-fetch option
1228 at configuration time.
1232 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1233 count of PKCS12_DEFAULT_ITER.
1235 *Tomáš Mráz and Sahana Prasad*
1237 * The openssl speed command does not use low-level API calls anymore.
1241 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1244 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1246 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1250 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1251 providers may supply their own group implementations (using either the "key
1252 exchange" or the "key encapsulation" methods) which will automatically be
1253 detected and used by libssl.
1255 *Matt Caswell, Nicola Tuveri*
1257 * The undocumented function X509_certificate_type() has been deprecated;
1261 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1265 * Removed RSA padding mode for SSLv23 (which was only used for
1266 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1267 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1272 * Deprecated the obsolete X9.31 RSA key generation related functions.
1274 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1275 is not allowed to return a value > 1, this is no more taken as failure.
1277 *Viktor Dukhovni and David von Oheimb*
1279 * Deprecated the obsolete X9.31 RSA key generation related functions
1280 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1281 BN_X931_generate_prime_ex().
1285 * The default key generation method for the regular 2-prime RSA keys was
1286 changed to the FIPS 186-4 B.3.6 method.
1290 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1294 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1298 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1299 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1301 *Rich Salz, Richard Levitte, and David von Oheimb*
1303 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1307 * Deprecated `OCSP_parse_url()`.
1311 * Validation of SM2 keys has been separated from the validation of regular EC
1316 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1317 switches: a validation failure triggers an early exit, returning a failure
1318 exit status to the parent process.
1322 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1323 to ignore unknown ciphers.
1327 * The `-cipher-commands` and `-digest-commands` options
1328 of the command line utility `list` have been deprecated.
1329 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1333 * Added convenience functions for generating asymmetric key pairs:
1334 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1335 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1339 * All of the low-level EC_KEY functions have been deprecated.
1341 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1343 * Deprecated all the libcrypto and libssl error string loading
1348 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1349 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1354 * The `-crypt` option to the `passwd` command line tool has been removed.
1358 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1363 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1367 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1368 EVP_PKEY_get1_tls_encodedpoint().
1372 * The security callback, which can be customised by application code, supports
1373 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1374 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1378 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1379 interface. Their functionality remains unchanged.
1383 * Added new option for 'openssl list', '-providers', which will display the
1384 list of loaded providers, their names, version and status. It optionally
1385 displays their gettable parameters.
1389 * Removed EVP_PKEY_set_alias_type().
1393 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1394 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1398 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1399 parameters are still checked for type safety at compile time via helper
1404 * Remove the RAND_DRBG API
1406 *Paul Dale and Matthias St. Pierre*
1408 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1409 as well as actual hostnames.
1413 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1414 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1415 conversely, silently ignore DTLS protocol version bounds when configuring
1416 TLS-based contexts. The commands can be repeated to set bounds of both
1417 types. The same applies with the corresponding "min_protocol" and
1418 "max_protocol" command-line switches, in case some application uses both TLS
1421 SSL_CTX instances that are created for a fixed protocol version (e.g.
1422 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1423 attempts to apply bounds to these protocol versions would result in an
1424 error. Now only the "version-flexible" SSL_CTX instances are subject to
1425 limits in configuration files in command-line options.
1429 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1434 * Reworked the recorded ERR codes to make better space for system errors.
1435 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1436 given code is a system error (true) or an OpenSSL error (false).
1440 * Reworked the test perl framework to better allow parallel testing.
1442 *Nicola Tuveri and David von Oheimb*
1444 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1445 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1449 * 'Configure' has been changed to figure out the configuration target if
1450 none is given on the command line. Consequently, the 'config' script is
1451 now only a mere wrapper. All documentation is changed to only mention
1454 *Rich Salz and Richard Levitte*
1456 * Added a library context `OSSL_LIB_CTX` that applications as well as
1457 other libraries can use to form a separate context within which
1458 libcrypto operations are performed.
1462 * Added various `_ex` functions to the OpenSSL API that support using
1463 a non-default `OSSL_LIB_CTX`.
1467 * Handshake now fails if Extended Master Secret extension is dropped
1472 * Dropped interactive mode from the `openssl` program.
1476 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1478 *David von Oheimb and Shane Lontis*
1480 * Deprecated `EC_METHOD_get_field_type()`.
1484 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1485 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1486 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1490 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1494 * Add CAdES-BES signature verification support, mostly derived
1495 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1497 *Filipe Raimundo da Silva*
1499 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1503 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1504 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1508 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1512 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1513 EC_KEY_precompute_mult().
1517 * Deprecated EC_POINTs_mul().
1521 * Removed FIPS_mode() and FIPS_mode_set().
1525 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1529 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1530 EC_POINT_get_Jprojective_coordinates_GFp().
1534 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1535 arrays to be more easily constructed via a series of utility functions.
1536 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1537 the various push functions and finally convert to a passable OSSL_PARAM
1538 array using OSSL_PARAM_BLD_to_param().
1542 * The security strength of SHA1 and MD5 based signatures in TLS has been
1547 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1548 contain a provider side internal key.
1552 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1556 * Project text documents not yet having a proper file name extension
1557 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1558 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1562 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1563 have been converted to Markdown with the goal to produce documents
1564 which not only look pretty when viewed online in the browser, but
1565 remain well readable inside a plain text editor.
1567 To achieve this goal, a 'minimalistic' Markdown style has been applied
1568 which avoids formatting elements that interfere too much with the
1569 reading flow in the text file. For example, it
1571 * avoids [ATX headings][] and uses [setext headings][] instead
1572 (which works for `<h1>` and `<h2>` headings only).
1573 * avoids [inline links][] and uses [reference links][] instead.
1574 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1576 [ATX headings]: https://github.github.com/gfm/#atx-headings
1577 [setext headings]: https://github.github.com/gfm/#setext-headings
1578 [inline links]: https://github.github.com/gfm/#inline-link
1579 [reference links]: https://github.github.com/gfm/#reference-link
1580 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1581 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1583 *Matthias St. Pierre*
1585 * The test suite is changed to preserve results of each test recipe.
1586 A new directory test-runs/ with subdirectories named like the
1587 test recipes are created in the build tree for this purpose.
1591 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1592 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1593 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1595 *David von Oheimb, Martin Peylo*
1597 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1598 It supports arbitrary request and response content types, GET redirection,
1599 TLS, connections via HTTP(S) proxies, connections and exchange via
1600 user-defined BIOs (allowing implicit connections), persistent connections,
1601 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1602 The legacy OCSP-focused (and only partly documented) API
1603 is retained for backward compatibility, while most of it is deprecated.
1607 * Added `util/check-format.pl`, a tool for checking adherence to the
1608 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1609 The checks performed are incomplete and yield some false positives.
1610 Still the tool should be useful for detecting most typical glitches.
1614 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1615 If domain name resolution yields multiple IP addresses all of them are tried
1616 after `connect()` failures.
1620 * All of the low-level RSA functions have been deprecated.
1624 * X509 certificates signed using SHA1 are no longer allowed at security
1629 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1630 modified to use PKEY APIs. These commands are now in maintenance mode
1631 and no new features will be added to them.
1635 * The command line utility rsautl has been deprecated.
1639 * The command line utilities genrsa and rsa have been modified to use PKEY
1640 APIs. They now write PKCS#8 keys by default. These commands are now in
1641 maintenance mode and no new features will be added to them.
1645 * All of the low-level DH functions have been deprecated.
1647 *Paul Dale and Matt Caswell*
1649 * All of the low-level DSA functions have been deprecated.
1653 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1654 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1658 * Deprecated low-level ECDH and ECDSA functions.
1662 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1666 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1667 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1668 a new formulation to include all the things it can be used for,
1669 as well as words of caution.
1673 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1677 * All of the low-level HMAC functions have been deprecated.
1679 *Paul Dale and David von Oheimb*
1681 * Over two thousand fixes were made to the documentation, including:
1682 - Common options (such as -rand/-writerand, TLS version control, etc)
1683 were refactored and point to newly-enhanced descriptions in openssl.pod.
1684 - Added style conformance for all options (with help from Richard Levitte),
1685 documented all reported missing options, added a CI build to check
1686 that all options are documented and that no unimplemented options
1688 - Documented some internals, such as all use of environment variables.
1689 - Addressed all internal broken L<> references.
1693 * All of the low-level CMAC functions have been deprecated.
1697 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1698 functions have been deprecated.
1700 *Paul Dale and David von Oheimb*
1702 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1703 set of functions. The documentation mentioned negative values for some
1704 errors, but this was never the case, so the mention of negative values
1707 Code that followed the documentation and thereby check with something
1708 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1712 * All of the low-level cipher functions have been deprecated.
1714 *Matt Caswell and Paul Dale*
1716 * Removed include/openssl/opensslconf.h.in and replaced it with
1717 include/openssl/configuration.h.in, which differs in not including
1718 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1719 was added to include both.
1721 This allows internal hacks where one might need to modify the set
1722 of configured macros, for example this if deprecated symbols are
1723 still supposed to be available internally:
1725 #include <openssl/configuration.h>
1727 #undef OPENSSL_NO_DEPRECATED
1728 #define OPENSSL_SUPPRESS_DEPRECATED
1730 #include <openssl/macros.h>
1732 This should not be used by applications that use the exported
1733 symbols, as that will lead to linking errors.
1737 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1738 used in exponentiation with 512-bit moduli. No EC algorithms are
1739 affected. Analysis suggests that attacks against 2-prime RSA1024,
1740 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1741 difficult to perform and are not believed likely. Attacks against DH512
1742 are considered just feasible. However, for an attack the target would
1743 have to re-use the DH512 private key, which is not recommended anyway.
1744 Also applications directly using the low-level API BN_mod_exp may be
1745 affected if they use BN_FLG_CONSTTIME.
1750 * Most memory-debug features have been deprecated, and the functionality
1751 replaced with no-ops.
1755 * Added documentation for the STACK API.
1759 * Introduced a new method type and API, OSSL_ENCODER, to represent
1760 generic encoders. These do the same sort of job that PEM writers
1761 and d2i functions do, but with support for methods supplied by
1762 providers, and the possibility for providers to support other
1767 * Introduced a new method type and API, OSSL_DECODER, to represent
1768 generic decoders. These do the same sort of job that PEM readers
1769 and i2d functions do, but with support for methods supplied by
1770 providers, and the possibility for providers to support other
1775 * Added a .pragma directive to the syntax of configuration files, to
1776 allow varying behavior in a supported and predictable manner.
1777 Currently added pragma:
1781 This allows dollar signs to be a keyword character unless it's
1782 followed by a opening brace or parenthesis. This is useful for
1783 platforms where dollar signs are commonly used in names, such as
1784 volume names and system directory names on VMS.
1788 * Added functionality to create an EVP_PKEY from user data.
1792 * Change the interpretation of the '--api' configuration option to
1793 mean that this is a desired API compatibility level with no
1794 further meaning. The previous interpretation, that this would
1795 also mean to remove all deprecated symbols up to and including
1796 the given version, no requires that 'no-deprecated' is also used
1797 in the configuration.
1799 When building applications, the desired API compatibility level
1800 can be set with the OPENSSL_API_COMPAT macro like before. For
1801 API compatibility version below 3.0, the old style numerical
1802 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1803 For version 3.0 and on, the value is expected to be the decimal
1804 value calculated from the major and minor version like this:
1806 MAJOR * 10000 + MINOR * 100
1810 -DOPENSSL_API_COMPAT=30000 For 3.0
1811 -DOPENSSL_API_COMPAT=30200 For 3.2
1813 To hide declarations that are deprecated up to and including the
1814 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1815 given when building the application as well.
1819 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1820 access to certificate and CRL stores via URIs and OSSL_STORE
1823 This adds the following functions:
1825 - X509_LOOKUP_store()
1826 - X509_STORE_load_file()
1827 - X509_STORE_load_path()
1828 - X509_STORE_load_store()
1829 - SSL_add_store_cert_subjects_to_stack()
1830 - SSL_CTX_set_default_verify_store()
1831 - SSL_CTX_load_verify_file()
1832 - SSL_CTX_load_verify_dir()
1833 - SSL_CTX_load_verify_store()
1837 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1838 The presence of this system service is determined at run-time.
1842 * Added functionality to create an EVP_PKEY context based on data
1843 for methods from providers. This takes an algorithm name and a
1844 property query string and simply stores them, with the intent
1845 that any operation that uses this context will use those strings
1846 to fetch the needed methods implicitly, thereby making the port
1847 of application written for pre-3.0 OpenSSL easier.
1851 * The undocumented function NCONF_WIN32() has been deprecated; for
1852 conversion details see the HISTORY section of doc/man5/config.pod
1856 * Introduced the new functions EVP_DigestSignInit_ex() and
1857 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1858 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1859 pages for further details.
1863 * Over two thousand fixes were made to the documentation, including:
1864 adding missing command flags, better style conformance, documentation
1867 *Rich Salz, Richard Levitte*
1869 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1870 X25519, X448, Ed25519 and Ed448.
1874 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1879 * Deprecated the public definition of `ERR_STATE` as well as the function
1880 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1885 * Added ERR functionality to give callers access to the stored function
1886 names that have replaced the older function code based functions.
1888 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1889 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1890 ERR_peek_error_all() and ERR_peek_last_error_all().
1892 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1893 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1894 ERR_func_error_string().
1898 * Extended testing to be verbose for failing tests only. The make variables
1899 VERBOSE_FAILURE or VF can be used to enable this:
1901 $ make VF=1 test # Unix
1902 $ mms /macro=(VF=1) test ! OpenVMS
1903 $ nmake VF=1 test # Windows
1907 * Added the `-copy_extensions` option to the `x509` command for use with
1908 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1909 all extensions in the request are copied to the certificate or vice versa.
1911 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1913 * Added the `-copy_extensions` option to the `req` command for use with
1914 `-x509`. When given with the `copy` or `copyall` argument,
1915 all extensions in the certification request are copied to the certificate.
1919 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1920 they generate are by default RFC 5280 compliant in the following sense:
1921 There is a subjectKeyIdentifier extension with a hash value of the public key
1922 and for not self-signed certs there is an authorityKeyIdentifier extension
1923 with a keyIdentifier field or issuer information identifying the signing key.
1924 This is done unless some configuration overrides the new default behavior,
1925 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1929 * Added several checks to `X509_verify_cert()` according to requirements in
1930 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1931 (which may be done by using the CLI option `-x509_strict`):
1932 * The basicConstraints of CA certificates must be marked critical.
1933 * CA certificates must explicitly include the keyUsage extension.
1934 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1935 * The issuer name of any certificate must not be empty.
1936 * The subject name of CA certs, certs with keyUsage crlSign,
1937 and certs without subjectAlternativeName must not be empty.
1938 * If a subjectAlternativeName extension is given it must not be empty.
1939 * The signatureAlgorithm field and the cert signature must be consistent.
1940 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1941 must not be marked critical.
1942 * The authorityKeyIdentifier must be given for X.509v3 certs
1943 unless they are self-signed.
1944 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1948 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1949 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1953 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1954 used even when parsing explicit parameters, when loading a encoded key
1955 or calling `EC_GROUP_new_from_ecpkparameters()`/
1956 `EC_GROUP_new_from_ecparameters()`.
1957 This prevents bypass of security hardening and performance gains,
1958 especially for curves with specialized EC_METHODs.
1959 By default, if a key encoded with explicit parameters is loaded and later
1960 encoded, the output is still encoded with explicit parameters, even if
1961 internally a "named" EC_GROUP is used for computation.
1965 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1966 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1967 NULL. After this change, only the cofactor parameter can be NULL. It also
1968 does some minimal sanity checks on the passed order.
1973 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1974 An attack is simple, if the first CMS_recipientInfo is valid but the
1975 second CMS_recipientInfo is chosen ciphertext. If the second
1976 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1977 encryption key will be replaced by garbage, and the message cannot be
1978 decoded, but if the RSA decryption fails, the correct encryption key is
1979 used and the recipient will not notice the attack.
1980 As a work around for this potential attack the length of the decrypted
1981 key must be equal to the cipher default key length, in case the
1982 certificate is not given and all recipientInfo are tried out.
1983 The old behaviour can be re-enabled in the CMS code by setting the
1984 CMS_DEBUG_DECRYPT flag.
1988 * Early start up entropy quality from the DEVRANDOM seed source has been
1989 improved for older Linux systems. The RAND subsystem will wait for
1990 /dev/random to be producing output before seeding from /dev/urandom.
1991 The seeded state is stored for future library initialisations using
1992 a system global shared memory segment. The shared memory identifier
1993 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1994 the desired value. The default identifier is 114.
1998 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1999 when primes for RSA keys are computed.
2000 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2001 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2002 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2003 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2004 This avoids possible fingerprinting of newly generated RSA modules.
2008 * Correct the extended master secret constant on EBCDIC systems. Without this
2009 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2010 negotiate EMS will fail. Unfortunately this also means that TLS connections
2011 between EBCDIC systems with this fix, and EBCDIC systems without this
2012 fix will fail if they negotiate EMS.
2016 * Changed the library initialisation so that the config file is now loaded
2017 by default. This was already the case for libssl. It now occurs for both
2018 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2019 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2023 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2024 where the former acts as a replacement for `ERR_put_error()`, and the
2025 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2026 `ERR_raise_data()` adds more flexibility by taking a format string and
2027 an arbitrary number of arguments following it, to be processed with
2032 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2033 to check if a named provider is loaded and available. When called, it
2034 will also activate all fallback providers if such are still present.
2038 * Enforce a minimum DH modulus size of 512 bits.
2042 * Changed DH parameters to generate the order q subgroup instead of 2q.
2043 Previously generated DH parameters are still accepted by DH_check
2044 but DH_generate_key works around that by clearing bit 0 of the
2045 private key for those. This avoids leaking bit 0 of the private key.
2049 * Significantly reduce secure memory usage by the randomness pools.
2053 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2058 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2059 algorithms. An implementation of a key exchange algorithm can be obtained
2060 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2061 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2062 the older EVP_PKEY_derive_init() function. See the man pages for the new
2063 functions for further details.
2067 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2071 * Removed the function names from error messages and deprecated the
2076 * Removed NextStep support and the macro OPENSSL_UNISTD
2080 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2081 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2082 Also removed "export var as function" capability; we do not export
2083 variables, only functions.
2087 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2088 an error and 1 indicating success. In previous versions of OpenSSL this
2089 was a void type. If a key was set longer than the maximum possible this
2094 * Support SM2 signing and verification schemes with X509 certificate.
2098 * Use SHA256 as the default digest for TS query in the `ts` app.
2102 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2106 * Default cipher lists/suites are now available via a function, the
2107 #defines are deprecated.
2111 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2112 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2113 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2117 * Join the directories crypto/x509 and crypto/x509v3
2121 * Added command 'openssl kdf' that uses the EVP_KDF API.
2125 * Added command 'openssl mac' that uses the EVP_MAC API.
2129 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2130 as default directories. Also added the command 'openssl info'
2131 for scripting purposes.
2135 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2140 * Add prediction resistance to the DRBG reseeding process.
2144 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2145 mandated by IEEE Std 1619-2018.
2149 * Added newline escaping functionality to a filename when using openssl dgst.
2150 This output format is to replicate the output format found in the `*sum`
2151 checksum programs. This aims to preserve backward compatibility.
2153 *Matt Eaton, Richard Levitte, and Paul Dale*
2155 * Removed the heartbeat message in DTLS feature, as it has very
2156 little usage and doesn't seem to fulfill a valuable purpose.
2157 The configuration option is now deprecated.
2161 * Changed the output of 'openssl {digestname} < file' to display the
2162 digest name in its output.
2166 * Added a new generic trace API which provides support for enabling
2167 instrumentation through trace output.
2169 *Richard Levitte & Matthias St. Pierre*
2171 * Added build tests for C++. These are generated files that only do one
2172 thing, to include one public OpenSSL head file each. This tests that
2173 the public header files can be usefully included in a C++ application.
2175 This test isn't enabled by default. It can be enabled with the option
2176 'enable-buildtest-c++'.
2180 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2184 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2188 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2192 * Added KMAC to EVP_MAC.
2196 * Added property based algorithm implementation selection framework to
2201 * Added SCA hardening for modular field inversion in EC_GROUP through
2202 a new dedicated field_inv() pointer in EC_METHOD.
2203 This also addresses a leakage affecting conversions from projective
2204 to affine coordinates.
2206 *Billy Bob Brumley, Nicola Tuveri*
2208 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2209 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2210 those algorithms that were already supported through the EVP_PKEY API
2211 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2212 and scrypt are now wrappers that call EVP_KDF.
2216 * Build devcrypto engine as a dynamic engine.
2218 *Eneas U de Queiroz*
2220 * Add keyed BLAKE2 to EVP_MAC.
2224 * Fix a bug in the computation of the endpoint-pair shared secret used
2225 by DTLS over SCTP. This breaks interoperability with older versions
2226 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2227 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2228 interoperability with such broken implementations. However, enabling
2229 this switch breaks interoperability with correct implementations.
2231 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2232 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2236 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2240 * Changed the license to the Apache License v2.0.
2244 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2246 - Major releases (indicated by incrementing the MAJOR release number)
2247 may introduce incompatible API/ABI changes.
2248 - Minor releases (indicated by incrementing the MINOR release number)
2249 may introduce new features but retain API/ABI compatibility.
2250 - Patch releases (indicated by incrementing the PATCH number)
2251 are intended for bug fixes and other improvements of existing
2252 features only (like improving performance or adding documentation)
2253 and retain API/ABI compatibility.
2257 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2261 * Remove the 'dist' target and add a tarball building script. The
2262 'dist' target has fallen out of use, and it shouldn't be
2263 necessary to configure just to create a source distribution.
2267 * Recreate the OS390-Unix config target. It no longer relies on a
2268 special script like it did for OpenSSL pre-1.1.0.
2272 * Instead of having the source directories listed in Configure, add
2273 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2278 * Add GMAC to EVP_MAC.
2282 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2286 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2287 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2288 to facilitate the continued use of MACs through raw private keys in
2289 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2293 * Deprecate ECDH_KDF_X9_62().
2297 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2298 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2299 are retained for backwards compatibility.
2303 * AES-XTS mode now enforces that its two keys are different to mitigate
2304 the attacked described in "Efficient Instantiations of Tweakable
2305 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2306 Details of this attack can be obtained from:
2307 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2311 * Rename the object files, i.e. give them other names than in previous
2312 versions. Their names now include the name of the final product, as
2313 well as its type mnemonic (bin, lib, shlib).
2317 * Added new option for 'openssl list', '-objects', which will display the
2318 list of built in objects, i.e. OIDs with names.
2322 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2323 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2328 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2329 improves application performance by removing data copies and providing
2330 applications with zero-copy system calls such as sendfile and splice.
2334 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2338 * `PKCS12_parse` now maintains the order of the parsed certificates
2339 when outputting them via `*ca` (rather than reversing it).
2343 * Deprecated pthread fork support methods.
2347 * Added support for FFDHE key exchange in TLS 1.3.
2351 * Added a new concept for OpenSSL plugability: providers. This
2352 functionality is designed to replace the ENGINE API and ENGINE
2353 implementations, and to be much more dynamic, allowing provider
2354 authors to introduce new algorithms among other things, as long as
2355 there's an API that supports the algorithm type.
2357 With this concept comes a new core API for interaction between
2358 libcrypto and provider implementations. Public libcrypto functions
2359 that want to use providers do so through this core API.
2361 The main documentation for this core API is found in
2362 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2363 refer to other manuals describing the API specific for supported
2364 algorithm types (also called operations).
2371 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2373 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2375 * Avoid loading of a dynamic engine twice.
2379 * Prioritise DANE TLSA issuer certs over peer certs
2383 * Fixed random API for MacOS prior to 10.12
2385 These MacOS versions don't support the CommonCrypto APIs
2389 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2391 * Fixed an SM2 Decryption Buffer Overflow.
2393 In order to decrypt SM2 encrypted data an application is expected to
2394 call the API function EVP_PKEY_decrypt(). Typically an application will
2395 call this function twice. The first time, on entry, the "out" parameter
2396 can be NULL and, on exit, the "outlen" parameter is populated with the
2397 buffer size required to hold the decrypted plaintext. The application
2398 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2399 again, but this time passing a non-NULL value for the "out" parameter.
2401 A bug in the implementation of the SM2 decryption code means that the
2402 calculation of the buffer size required to hold the plaintext returned
2403 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2404 size required by the second call. This can lead to a buffer overflow
2405 when EVP_PKEY_decrypt() is called by the application a second time with
2406 a buffer that is too small.
2408 A malicious attacker who is able present SM2 content for decryption to
2409 an application could cause attacker chosen data to overflow the buffer
2410 by up to a maximum of 62 bytes altering the contents of other data held
2411 after the buffer, possibly changing application behaviour or causing
2412 the application to crash. The location of the buffer is application
2413 dependent but is typically heap allocated.
2418 * Fixed various read buffer overruns processing ASN.1 strings
2420 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2421 structure which contains a buffer holding the string data and a field
2422 holding the buffer length. This contrasts with normal C strings which
2423 are represented as a buffer for the string data which is terminated
2424 with a NUL (0) byte.
2426 Although not a strict requirement, ASN.1 strings that are parsed using
2427 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2428 well as any string whose value has been set with the ASN1_STRING_set()
2429 function will additionally NUL terminate the byte array in the
2430 ASN1_STRING structure.
2432 However, it is possible for applications to directly construct valid
2433 ASN1_STRING structures which do not NUL terminate the byte array by
2434 directly setting the "data" and "length" fields in the ASN1_STRING
2435 array. This can also happen by using the ASN1_STRING_set0() function.
2437 Numerous OpenSSL functions that print ASN.1 data have been found to
2438 assume that the ASN1_STRING byte array will be NUL terminated, even
2439 though this is not guaranteed for strings that have been directly
2440 constructed. Where an application requests an ASN.1 structure to be
2441 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2442 been directly constructed by the application without NUL terminating
2443 the "data" field, then a read buffer overrun can occur.
2445 The same thing can also occur during name constraints processing
2446 of certificates (for example if a certificate has been directly
2447 constructed by the application instead of loading it via the OpenSSL
2448 parsing functions, and the certificate contains non NUL terminated
2449 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2450 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2452 If a malicious actor can cause an application to directly construct an
2453 ASN1_STRING and then process it through one of the affected OpenSSL
2454 functions then this issue could be hit. This might result in a crash
2455 (causing a Denial of Service attack). It could also result in the
2456 disclosure of private memory contents (such as private keys, or
2457 sensitive plaintext).
2462 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2464 * Fixed a problem with verifying a certificate chain when using the
2465 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2466 the certificates present in a certificate chain. It is not set by default.
2468 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2469 the chain that have explicitly encoded elliptic curve parameters was added
2470 as an additional strict check.
2472 An error in the implementation of this check meant that the result of a
2473 previous check to confirm that certificates in the chain are valid CA
2474 certificates was overwritten. This effectively bypasses the check
2475 that non-CA certificates must not be able to issue other certificates.
2477 If a "purpose" has been configured then there is a subsequent opportunity
2478 for checks that the certificate is a valid CA. All of the named "purpose"
2479 values implemented in libcrypto perform this check. Therefore, where
2480 a purpose is set the certificate chain will still be rejected even when the
2481 strict flag has been used. A purpose is set by default in libssl client and
2482 server certificate verification routines, but it can be overridden or
2483 removed by an application.
2485 In order to be affected, an application must explicitly set the
2486 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2487 for the certificate verification or, in the case of TLS client or server
2488 applications, override the default purpose.
2493 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2494 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2495 renegotiation ClientHello omits the signature_algorithms extension (where it
2496 was present in the initial ClientHello), but includes a
2497 signature_algorithms_cert extension then a NULL pointer dereference will
2498 result, leading to a crash and a denial of service attack.
2500 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2501 (which is the default configuration). OpenSSL TLS clients are not impacted by
2505 *Peter Kästle and Samuel Sapalski*
2507 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2509 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2510 create a unique hash value based on the issuer and serial number data
2511 contained within an X509 certificate. However, it was failing to correctly
2512 handle any errors that may occur while parsing the issuer field (which might
2513 occur if the issuer field is maliciously constructed). This may subsequently
2514 result in a NULL pointer deref and a crash leading to a potential denial of
2520 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2521 padding mode to correctly check for rollback attacks. This is considered a
2522 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2527 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2528 functions. Previously they could overflow the output length argument in some
2529 cases where the input length is close to the maximum permissible length for
2530 an integer on the platform. In such cases the return value from the function
2531 call would be 1 (indicating success), but the output length value would be
2532 negative. This could cause applications to behave incorrectly or crash.
2537 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2538 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2539 could be exploited in a side channel attack to recover the password. Since
2540 the attack is local host only this is outside of the current OpenSSL
2541 threat model and therefore no CVE is assigned.
2543 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2548 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2550 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2551 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2552 If an attacker can control both items being compared then this could lead
2553 to a possible denial of service attack. OpenSSL itself uses the
2554 GENERAL_NAME_cmp function for two purposes:
2555 1) Comparing CRL distribution point names between an available CRL and a
2556 CRL distribution point embedded in an X509 certificate
2557 2) When verifying that a timestamp response token signer matches the
2558 timestamp authority name (exposed via the API functions
2559 TS_RESP_verify_response and TS_RESP_verify_token)
2564 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2566 * Certificates with explicit curve parameters are now disallowed in
2567 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2571 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2572 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2573 conversely, silently ignore DTLS protocol version bounds when configuring
2574 TLS-based contexts. The commands can be repeated to set bounds of both
2575 types. The same applies with the corresponding "min_protocol" and
2576 "max_protocol" command-line switches, in case some application uses both TLS
2579 SSL_CTX instances that are created for a fixed protocol version (e.g.
2580 TLSv1_server_method()) also silently ignore version bounds. Previously
2581 attempts to apply bounds to these protocol versions would result in an
2582 error. Now only the "version-flexible" SSL_CTX instances are subject to
2583 limits in configuration files in command-line options.
2587 * Handshake now fails if Extended Master Secret extension is dropped
2592 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2594 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2596 * Fixed segmentation fault in SSL_check_chain()
2597 Server or client applications that call the SSL_check_chain() function
2598 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2599 dereference as a result of incorrect handling of the
2600 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2601 or unrecognised signature algorithm is received from the peer. This could
2602 be exploited by a malicious peer in a Denial of Service attack.
2607 * Added AES consttime code for no-asm configurations
2608 an optional constant time support for AES was added
2609 when building openssl for no-asm.
2610 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2611 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2612 At this time this feature is by default disabled.
2613 It will be enabled by default in 3.0.
2617 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2619 * Revert the change of EOF detection while reading in libssl to avoid
2620 regressions in applications depending on the current way of reporting
2621 the EOF. As the existing method is not fully accurate the change to
2622 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2623 branch and will be present in the 3.0 release.
2627 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2628 when primes for RSA keys are computed.
2629 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2630 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2631 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2632 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2633 This avoids possible fingerprinting of newly generated RSA modules.
2637 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2639 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2640 while reading in libssl then we would report an error back to the
2641 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2642 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2643 therefore give a hint as to what went wrong.
2647 * Check that ed25519 and ed448 are allowed by the security level. Previously
2648 signature algorithms not using an MD were not being checked that they were
2649 allowed by the security level.
2653 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2654 was not quite right. The behaviour was not consistent between resumption
2655 and normal handshakes, and also not quite consistent with historical
2656 behaviour. The behaviour in various scenarios has been clarified and
2657 it has been updated to make it match historical behaviour as closely as
2662 * *[VMS only]* The header files that the VMS compilers include automatically,
2663 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2664 that the C++ compiler doesn't understand. This is a shortcoming in the
2665 compiler, but can be worked around with `__cplusplus` guards.
2667 C++ applications that use OpenSSL libraries must be compiled using the
2668 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2669 functions. Otherwise, only functions with symbols of less than 31
2670 characters can be used, as the linker will not be able to successfully
2671 resolve symbols with longer names.
2675 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2676 The presence of this system service is determined at run-time.
2680 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2685 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2687 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2688 number generator (RNG). This was intended to include protection in the
2689 event of a fork() system call in order to ensure that the parent and child
2690 processes did not share the same RNG state. However, this protection was not
2691 being used in the default case.
2693 A partial mitigation for this issue is that the output from a high
2694 precision timer is mixed into the RNG state so the likelihood of a parent
2695 and child process sharing state is significantly reduced.
2697 If an application already calls OPENSSL_init_crypto() explicitly using
2698 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2701 *Matthias St. Pierre*
2703 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2704 used even when parsing explicit parameters, when loading a encoded key
2705 or calling `EC_GROUP_new_from_ecpkparameters()`/
2706 `EC_GROUP_new_from_ecparameters()`.
2707 This prevents bypass of security hardening and performance gains,
2708 especially for curves with specialized EC_METHODs.
2709 By default, if a key encoded with explicit parameters is loaded and later
2710 encoded, the output is still encoded with explicit parameters, even if
2711 internally a "named" EC_GROUP is used for computation.
2715 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2716 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2717 NULL. After this change, only the cofactor parameter can be NULL. It also
2718 does some minimal sanity checks on the passed order.
2723 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2724 An attack is simple, if the first CMS_recipientInfo is valid but the
2725 second CMS_recipientInfo is chosen ciphertext. If the second
2726 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2727 encryption key will be replaced by garbage, and the message cannot be
2728 decoded, but if the RSA decryption fails, the correct encryption key is
2729 used and the recipient will not notice the attack.
2730 As a work around for this potential attack the length of the decrypted
2731 key must be equal to the cipher default key length, in case the
2732 certificate is not given and all recipientInfo are tried out.
2733 The old behaviour can be re-enabled in the CMS code by setting the
2734 CMS_DEBUG_DECRYPT flag.
2739 * Early start up entropy quality from the DEVRANDOM seed source has been
2740 improved for older Linux systems. The RAND subsystem will wait for
2741 /dev/random to be producing output before seeding from /dev/urandom.
2742 The seeded state is stored for future library initialisations using
2743 a system global shared memory segment. The shared memory identifier
2744 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2745 the desired value. The default identifier is 114.
2749 * Correct the extended master secret constant on EBCDIC systems. Without this
2750 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2751 negotiate EMS will fail. Unfortunately this also means that TLS connections
2752 between EBCDIC systems with this fix, and EBCDIC systems without this
2753 fix will fail if they negotiate EMS.
2757 * Use Windows installation paths in the mingw builds
2759 Mingw isn't a POSIX environment per se, which means that Windows
2760 paths should be used for installation.
2765 * Changed DH_check to accept parameters with order q and 2q subgroups.
2766 With order 2q subgroups the bit 0 of the private key is not secret
2767 but DH_generate_key works around that by clearing bit 0 of the
2768 private key for those. This avoids leaking bit 0 of the private key.
2772 * Significantly reduce secure memory usage by the randomness pools.
2776 * Revert the DEVRANDOM_WAIT feature for Linux systems
2778 The DEVRANDOM_WAIT feature added a select() call to wait for the
2779 /dev/random device to become readable before reading from the
2780 /dev/urandom device.
2782 It turned out that this change had negative side effects on
2783 performance which were not acceptable. After some discussion it
2784 was decided to revert this feature and leave it up to the OS
2785 resp. the platform maintainer to ensure a proper initialization
2786 during early boot time.
2788 *Matthias St. Pierre*
2790 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2792 * Add build tests for C++. These are generated files that only do one
2793 thing, to include one public OpenSSL head file each. This tests that
2794 the public header files can be usefully included in a C++ application.
2796 This test isn't enabled by default. It can be enabled with the option
2797 'enable-buildtest-c++'.
2801 * Enable SHA3 pre-hashing for ECDSA and DSA.
2805 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2806 This changes the size when using the `genpkey` command when no size is given.
2807 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2808 generation commands to use 2048 bits by default.
2812 * Reorganize the manual pages to consistently have RETURN VALUES,
2813 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2814 util/fix-doc-nits accordingly.
2816 *Paul Yang, Joshua Lock*
2818 * Add the missing accessor EVP_PKEY_get0_engine()
2822 * Have commands like `s_client` and `s_server` output the signature scheme
2823 along with other cipher suite parameters when debugging.
2827 * Make OPENSSL_config() error agnostic again.
2831 * Do the error handling in RSA decryption constant time.
2835 * Prevent over long nonces in ChaCha20-Poly1305.
2837 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2838 for every encryption operation. RFC 7539 specifies that the nonce value
2839 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2840 and front pads the nonce with 0 bytes if it is less than 12
2841 bytes. However it also incorrectly allows a nonce to be set of up to 16
2842 bytes. In this case only the last 12 bytes are significant and any
2843 additional leading bytes are ignored.
2845 It is a requirement of using this cipher that nonce values are
2846 unique. Messages encrypted using a reused nonce value are susceptible to
2847 serious confidentiality and integrity attacks. If an application changes
2848 the default nonce length to be longer than 12 bytes and then makes a
2849 change to the leading bytes of the nonce expecting the new value to be a
2850 new unique nonce then such an application could inadvertently encrypt
2851 messages with a reused nonce.
2853 Additionally the ignored bytes in a long nonce are not covered by the
2854 integrity guarantee of this cipher. Any application that relies on the
2855 integrity of these ignored leading bytes of a long nonce may be further
2856 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2857 is safe because no such use sets such a long nonce value. However user
2858 applications that use this cipher directly and set a non-default nonce
2859 length to be longer than 12 bytes may be vulnerable.
2861 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2867 * Add DEVRANDOM_WAIT feature for Linux systems
2869 On older Linux systems where the getrandom() system call is not available,
2870 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2871 Contrary to getrandom(), the /dev/urandom device will not block during
2872 early boot when the kernel CSPRNG has not been seeded yet.
2874 To mitigate this known weakness, use select() to wait for /dev/random to
2875 become readable before reading from /dev/urandom.
2877 * Ensure that SM2 only uses SM3 as digest algorithm
2881 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2883 * Change the info callback signals for the start and end of a post-handshake
2884 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2885 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2886 confused by this and assume that a TLSv1.2 renegotiation has started. This
2887 can break KeyUpdate handling. Instead we no longer signal the start and end
2888 of a post handshake message exchange (although the messages themselves are
2889 still signalled). This could break some applications that were expecting
2890 the old signals. However without this KeyUpdate is not usable for many
2895 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2897 * Timing vulnerability in DSA signature generation
2899 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2900 timing side channel attack. An attacker could use variations in the signing
2901 algorithm to recover the private key.
2903 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2908 * Timing vulnerability in ECDSA signature generation
2910 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2911 timing side channel attack. An attacker could use variations in the signing
2912 algorithm to recover the private key.
2914 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2919 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2920 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2921 of two gigabytes and the error handling improved.
2923 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2924 categorized as a normal bug, not a security issue, because the DRBG reseeds
2925 automatically and is fully functional even without additional randomness
2926 provided by the application.
2928 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2930 * Add a new ClientHello callback. Provides a callback interface that gives
2931 the application the ability to adjust the nascent SSL object at the
2932 earliest stage of ClientHello processing, immediately after extensions have
2933 been collected but before they have been processed. In particular, this
2934 callback can adjust the supported TLS versions in response to the contents
2939 * Add SM2 base algorithm support.
2943 * s390x assembly pack: add (improved) hardware-support for the following
2944 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2945 aes-cfb/cfb8, aes-ecb.
2949 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2950 parameter is no longer accepted, as it leads to a corrupt table. NULL
2951 pem_str is reserved for alias entries only.
2955 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2956 step for prime curves. The new implementation is based on formulae from
2957 differential addition-and-doubling in homogeneous projective coordinates
2958 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2959 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2960 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2961 to work in projective coordinates.
2963 *Billy Bob Brumley, Nicola Tuveri*
2965 * Change generating and checking of primes so that the error rate of not
2966 being prime depends on the intended use based on the size of the input.
2967 For larger primes this will result in more rounds of Miller-Rabin.
2968 The maximal error rate for primes with more than 1080 bits is lowered
2971 *Kurt Roeckx, Annie Yousar*
2973 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2977 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2978 moving between systems, and to avoid confusion when a Windows build is
2979 done with mingw vs with MSVC. For POSIX installs, there's still a
2980 symlink or copy named 'tsget' to avoid that confusion as well.
2984 * Revert blinding in ECDSA sign and instead make problematic addition
2985 length-invariant. Switch even to fixed-length Montgomery multiplication.
2989 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2990 step for binary curves. The new implementation is based on formulae from
2991 differential addition-and-doubling in mixed Lopez-Dahab projective
2992 coordinates, modified to independently blind the operands.
2994 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2996 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2997 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2998 EC_METHODs to implement their own specialized "ladder step", to take
2999 advantage of more favorable coordinate systems or more efficient
3000 differential addition-and-doubling algorithms.
3002 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3004 * Modified the random device based seed sources to keep the relevant
3005 file descriptors open rather than reopening them on each access.
3006 This allows such sources to operate in a chroot() jail without
3007 the associated device nodes being available. This behaviour can be
3008 controlled using RAND_keep_random_devices_open().
3012 * Numerous side-channel attack mitigations have been applied. This may have
3013 performance impacts for some algorithms for the benefit of improved
3014 security. Specific changes are noted in this change log by their respective
3019 * AIX shared library support overhaul. Switch to AIX "natural" way of
3020 handling shared libraries, which means collecting shared objects of
3021 different versions and bitnesses in one common archive. This allows to
3022 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3023 doesn't affect the way 3rd party applications are linked, only how
3024 multi-version installation is managed.
3028 * Make ec_group_do_inverse_ord() more robust and available to other
3029 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3030 mitigations are applied to the fallback BN_mod_inverse().
3031 When using this function rather than BN_mod_inverse() directly, new
3032 EC cryptosystem implementations are then safer-by-default.
3036 * Add coordinate blinding for EC_POINT and implement projective
3037 coordinate blinding for generic prime curves as a countermeasure to
3038 chosen point SCA attacks.
3040 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3042 * Add blinding to ECDSA and DSA signatures to protect against side channel
3043 attacks discovered by Keegan Ryan (NCC Group).
3047 * Enforce checking in the `pkeyutl` command to ensure that the input
3048 length does not exceed the maximum supported digest length when performing
3049 a sign, verify or verifyrecover operation.
3053 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3054 I/O in combination with something like select() or poll() will hang. This
3055 can be turned off again using SSL_CTX_clear_mode().
3056 Many applications do not properly handle non-application data records, and
3057 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3058 around the problems in those applications, but can also break some.
3059 It's recommended to read the manpages about SSL_read(), SSL_write(),
3060 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3061 SSL_CTX_set_read_ahead() again.
3065 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3066 now allow empty (zero character) pass phrases.
3070 * Apply blinding to binary field modular inversion and remove patent
3071 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3075 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3076 binary and prime elliptic curves.
3080 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3081 constant time fixed point multiplication.
3085 * Revise elliptic curve scalar multiplication with timing attack
3086 defenses: ec_wNAF_mul redirects to a constant time implementation
3087 when computing fixed point and variable point multiplication (which
3088 in OpenSSL are mostly used with secret scalars in keygen, sign,
3089 ECDH derive operations).
3090 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3093 * Updated CONTRIBUTING
3097 * Updated DRBG / RAND to request nonce and additional low entropy
3098 randomness from the system.
3100 *Matthias St. Pierre*
3102 * Updated 'openssl rehash' to use OpenSSL consistent default.
3106 * Moved the load of the ssl_conf module to libcrypto, which helps
3107 loading engines that libssl uses before libssl is initialised.
3111 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3115 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3117 *Ingo Schwarze, Rich Salz*
3119 * Added output of accepting IP address and port for 'openssl s_server'
3123 * Added a new API for TLSv1.3 ciphersuites:
3124 SSL_CTX_set_ciphersuites()
3125 SSL_set_ciphersuites()
3129 * Memory allocation failures consistently add an error to the error
3134 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3135 in libcrypto when run as setuid/setgid.
3139 * Load any config file by default when libssl is used.
3143 * Added new public header file <openssl/rand_drbg.h> and documentation
3144 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3146 *Matthias St. Pierre*
3148 * QNX support removed (cannot find contributors to get their approval
3149 for the license change).
3153 * TLSv1.3 replay protection for early data has been implemented. See the
3154 SSL_read_early_data() man page for further details.
3158 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3159 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3160 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3161 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3162 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3163 configuration has been separated out. See the ciphers man page or the
3164 SSL_CTX_set_ciphersuites() man page for more information.
3168 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3169 in responder mode now supports the new "-multi" option, which
3170 spawns the specified number of child processes to handle OCSP
3171 requests. The "-timeout" option now also limits the OCSP
3172 responder's patience to wait to receive the full client request
3173 on a newly accepted connection. Child processes are respawned
3174 as needed, and the CA index file is automatically reloaded
3175 when changed. This makes it possible to run the "ocsp" responder
3176 as a long-running service, making the OpenSSL CA somewhat more
3177 feature-complete. In this mode, most diagnostic messages logged
3178 after entering the event loop are logged via syslog(3) rather than
3183 * Added support for X448 and Ed448. Heavily based on original work by
3188 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3189 objects loaded. This adds the functions OSSL_STORE_expect() and
3190 OSSL_STORE_find() as well as needed tools to construct searches and
3191 get the search data out of them.
3195 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3196 version of OpenSSL should review their configuration settings to ensure
3197 that they are still appropriate for TLSv1.3. For further information see:
3198 <https://wiki.openssl.org/index.php/TLS1.3>
3202 * Grand redesign of the OpenSSL random generator
3204 The default RAND method now utilizes an AES-CTR DRBG according to
3205 NIST standard SP 800-90Ar1. The new random generator is essentially
3206 a port of the default random generator from the OpenSSL FIPS 2.0
3207 object module. It is a hybrid deterministic random bit generator
3208 using an AES-CTR bit stream and which seeds and reseeds itself
3209 automatically using trusted system entropy sources.
3211 Some of its new features are:
3212 - Support for multiple DRBG instances with seed chaining.
3213 - The default RAND method makes use of a DRBG.
3214 - There is a public and private DRBG instance.
3215 - The DRBG instances are fork-safe.
3216 - Keep all global DRBG instances on the secure heap if it is enabled.
3217 - The public and private DRBG instance are per thread for lock free
3220 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3222 * Changed Configure so it only says what it does and doesn't dump
3223 so much data. Instead, ./configdata.pm should be used as a script
3224 to display all sorts of configuration data.
3228 * Added processing of "make variables" to Configure.
3232 * Added SHA512/224 and SHA512/256 algorithm support.
3236 * The last traces of Netware support, first removed in 1.1.0, have
3241 * Get rid of Makefile.shared, and in the process, make the processing
3242 of certain files (rc.obj, or the .def/.map/.opt files produced from
3243 the ordinal files) more visible and hopefully easier to trace and
3244 debug (or make silent).
3248 * Make it possible to have environment variable assignments as
3249 arguments to config / Configure.
3253 * Add multi-prime RSA (RFC 8017) support.
3257 * Add SM3 implemented according to GB/T 32905-2016
3258 *Jack Lloyd <jack.lloyd@ribose.com>,*
3259 *Ronald Tse <ronald.tse@ribose.com>,*
3260 *Erick Borsboom <erick.borsboom@ribose.com>*
3262 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3263 as documented in RFC6066.
3264 Based on a patch from Tomasz Moń
3266 *Filipe Raimundo da Silva*
3268 * Add SM4 implemented according to GB/T 32907-2016.
3269 *Jack Lloyd <jack.lloyd@ribose.com>,*
3270 *Ronald Tse <ronald.tse@ribose.com>,*
3271 *Erick Borsboom <erick.borsboom@ribose.com>*
3273 * Reimplement -newreq-nodes and ERR_error_string_n; the
3274 original author does not agree with the license change.
3278 * Add ARIA AEAD TLS support.
3282 * Some macro definitions to support VS6 have been removed. Visual
3283 Studio 6 has not worked since 1.1.0
3287 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3288 without clearing the errors.
3292 * Add "atfork" functions. If building on a system that without
3293 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3294 requirements. The RAND facility now uses/requires this.
3302 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3303 not possible to disable entirely. However, it's still possible to
3304 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3307 To disable, configure with 'no-ui-console'. 'no-ui' is still
3308 possible to use as an alias. Check at compile time with the
3309 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3310 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3314 * Add a STORE module, which implements a uniform and URI based reader of
3315 stores that can contain keys, certificates, CRLs and numerous other
3316 objects. The main API is loosely based on a few stdio functions,
3317 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3318 OSSL_STORE_error and OSSL_STORE_close.
3319 The implementation uses backends called "loaders" to implement arbitrary
3320 URI schemes. There is one built in "loader" for the 'file' scheme.
3324 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3325 then adjusted to work on FreeBSD 8.4 as well.
3326 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3327 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3331 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3332 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3333 error code calls like this:
3335 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3337 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3338 that can be encoded in C. For the foreseeable future, this will only
3341 *Richard Levitte and Tim Hudson*
3343 * Removed BSD cryptodev engine.
3347 * Add a build target 'build_all_generated', to build all generated files
3348 and only that. This can be used to prepare everything that requires
3349 things like perl for a system that lacks perl and then move everything
3350 to that system and do the rest of the build there.
3354 * In the UI interface, make it possible to duplicate the user data. This
3355 can be used by engines that need to retain the data for a longer time
3356 than just the call where this user data is passed.
3360 * Ignore the '-named_curve auto' value for compatibility of applications
3363 *Tomáš Mráz <tmraz@fedoraproject.org>*
3365 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3366 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3367 alerts across multiple records (some of which could be empty). In practice
3368 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3369 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3370 support this at all. Supporting it adds significant complexity to the
3371 record layer, and its removal is unlikely to cause interoperability
3376 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3377 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3378 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3383 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3384 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3386 *Richard Levitte, Andy Polyakov*
3388 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3393 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3394 platform rather than 'mingw'.
3398 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3399 success if they are asked to add an object which already exists
3400 in the store. This change cascades to other functions which load
3401 certificates and CRLs.
3405 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3406 facilitate stack unwinding even from assembly subroutines.
3410 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3411 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3415 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3416 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3417 which is the minimum version we support.
3421 * Certificate time validation (X509_cmp_time) enforces stricter
3422 compliance with RFC 5280. Fractional seconds and timezone offsets
3423 are no longer allowed.
3427 * Add support for ARIA
3431 * s_client will now send the Server Name Indication (SNI) extension by
3432 default unless the new "-noservername" option is used. The server name is
3433 based on the host provided to the "-connect" option unless overridden by
3434 using "-servername".
3438 * Add support for SipHash
3442 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3443 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3444 prevent issues where no progress is being made and the peer continually
3445 sends unrecognised record types, using up resources processing them.
3449 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3450 using the algorithm defined in
3451 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3455 * Heartbeat support has been removed; the ABI is changed for now.
3457 *Richard Levitte, Rich Salz*
3459 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3463 * The RSA "null" method, which was partially supported to avoid patent
3464 issues, has been replaced to always returns NULL.
3471 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3473 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3474 used even when parsing explicit parameters, when loading a encoded key
3475 or calling `EC_GROUP_new_from_ecpkparameters()`/
3476 `EC_GROUP_new_from_ecparameters()`.
3477 This prevents bypass of security hardening and performance gains,
3478 especially for curves with specialized EC_METHODs.
3479 By default, if a key encoded with explicit parameters is loaded and later
3480 encoded, the output is still encoded with explicit parameters, even if
3481 internally a "named" EC_GROUP is used for computation.
3485 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3486 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3487 NULL. After this change, only the cofactor parameter can be NULL. It also
3488 does some minimal sanity checks on the passed order.
3493 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3494 An attack is simple, if the first CMS_recipientInfo is valid but the
3495 second CMS_recipientInfo is chosen ciphertext. If the second
3496 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3497 encryption key will be replaced by garbage, and the message cannot be
3498 decoded, but if the RSA decryption fails, the correct encryption key is
3499 used and the recipient will not notice the attack.
3500 As a work around for this potential attack the length of the decrypted
3501 key must be equal to the cipher default key length, in case the
3502 certificate is not given and all recipientInfo are tried out.
3503 The old behaviour can be re-enabled in the CMS code by setting the
3504 CMS_DEBUG_DECRYPT flag.
3509 * Use Windows installation paths in the mingw builds
3511 Mingw isn't a POSIX environment per se, which means that Windows
3512 paths should be used for installation.
3517 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3519 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3520 This changes the size when using the `genpkey` command when no size is given.
3521 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3522 generation commands to use 2048 bits by default.
3526 * Prevent over long nonces in ChaCha20-Poly1305.
3528 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3529 for every encryption operation. RFC 7539 specifies that the nonce value
3530 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3531 and front pads the nonce with 0 bytes if it is less than 12
3532 bytes. However it also incorrectly allows a nonce to be set of up to 16
3533 bytes. In this case only the last 12 bytes are significant and any
3534 additional leading bytes are ignored.
3536 It is a requirement of using this cipher that nonce values are
3537 unique. Messages encrypted using a reused nonce value are susceptible to
3538 serious confidentiality and integrity attacks. If an application changes
3539 the default nonce length to be longer than 12 bytes and then makes a
3540 change to the leading bytes of the nonce expecting the new value to be a
3541 new unique nonce then such an application could inadvertently encrypt
3542 messages with a reused nonce.
3544 Additionally the ignored bytes in a long nonce are not covered by the
3545 integrity guarantee of this cipher. Any application that relies on the
3546 integrity of these ignored leading bytes of a long nonce may be further
3547 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3548 is safe because no such use sets such a long nonce value. However user
3549 applications that use this cipher directly and set a non-default nonce
3550 length to be longer than 12 bytes may be vulnerable.
3552 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3558 * Added SCA hardening for modular field inversion in EC_GROUP through
3559 a new dedicated field_inv() pointer in EC_METHOD.
3560 This also addresses a leakage affecting conversions from projective
3561 to affine coordinates.
3563 *Billy Bob Brumley, Nicola Tuveri*
3565 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3566 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3570 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3574 * Remove the 'dist' target and add a tarball building script. The
3575 'dist' target has fallen out of use, and it shouldn't be
3576 necessary to configure just to create a source distribution.
3580 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3582 * Timing vulnerability in DSA signature generation
3584 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3585 timing side channel attack. An attacker could use variations in the signing
3586 algorithm to recover the private key.
3588 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3593 * Timing vulnerability in ECDSA signature generation
3595 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3596 timing side channel attack. An attacker could use variations in the signing
3597 algorithm to recover the private key.
3599 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3604 * Add coordinate blinding for EC_POINT and implement projective
3605 coordinate blinding for generic prime curves as a countermeasure to
3606 chosen point SCA attacks.
3608 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3610 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3612 * Client DoS due to large DH parameter
3614 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3615 malicious server can send a very large prime value to the client. This will
3616 cause the client to spend an unreasonably long period of time generating a
3617 key for this prime resulting in a hang until the client has finished. This
3618 could be exploited in a Denial Of Service attack.
3620 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3625 * Cache timing vulnerability in RSA Key Generation
3627 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3628 a cache timing side channel attack. An attacker with sufficient access to
3629 mount cache timing attacks during the RSA key generation process could
3630 recover the private key.
3632 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3633 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3638 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3639 parameter is no longer accepted, as it leads to a corrupt table. NULL
3640 pem_str is reserved for alias entries only.
3644 * Revert blinding in ECDSA sign and instead make problematic addition
3645 length-invariant. Switch even to fixed-length Montgomery multiplication.
3649 * Change generating and checking of primes so that the error rate of not
3650 being prime depends on the intended use based on the size of the input.
3651 For larger primes this will result in more rounds of Miller-Rabin.
3652 The maximal error rate for primes with more than 1080 bits is lowered
3655 *Kurt Roeckx, Annie Yousar*
3657 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3661 * Add blinding to ECDSA and DSA signatures to protect against side channel
3662 attacks discovered by Keegan Ryan (NCC Group).
3666 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3667 now allow empty (zero character) pass phrases.
3671 * Certificate time validation (X509_cmp_time) enforces stricter
3672 compliance with RFC 5280. Fractional seconds and timezone offsets
3673 are no longer allowed.
3677 * Fixed a text canonicalisation bug in CMS
3679 Where a CMS detached signature is used with text content the text goes
3680 through a canonicalisation process first prior to signing or verifying a
3681 signature. This process strips trailing space at the end of lines, converts
3682 line terminators to CRLF and removes additional trailing line terminators
3683 at the end of a file. A bug in the canonicalisation process meant that
3684 some characters, such as form-feed, were incorrectly treated as whitespace
3685 and removed. This is contrary to the specification (RFC5485). This fix
3686 could mean that detached text data signed with an earlier version of
3687 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3688 signed with a fixed OpenSSL may fail to verify with an earlier version of
3689 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3690 and use the "-binary" flag (for the "cms" command line application) or set
3691 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3695 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3697 * Constructed ASN.1 types with a recursive definition could exceed the stack
3699 Constructed ASN.1 types with a recursive definition (such as can be found
3700 in PKCS7) could eventually exceed the stack given malicious input with
3701 excessive recursion. This could result in a Denial Of Service attack. There
3702 are no such structures used within SSL/TLS that come from untrusted sources
3703 so this is considered safe.
3705 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3711 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3713 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3714 effectively reduced to only comparing the least significant bit of each
3715 byte. This allows an attacker to forge messages that would be considered as
3716 authenticated in an amount of tries lower than that guaranteed by the
3717 security claims of the scheme. The module can only be compiled by the
3718 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3720 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3726 * Add a build target 'build_all_generated', to build all generated files
3727 and only that. This can be used to prepare everything that requires
3728 things like perl for a system that lacks perl and then move everything
3729 to that system and do the rest of the build there.
3733 * Backport SSL_OP_NO_RENGOTIATION
3735 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3736 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3737 changes this is no longer possible in 1.1.0. Therefore, the new
3738 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3739 1.1.0 to provide equivalent functionality.
3741 Note that if an application built against 1.1.0h headers (or above) is run
3742 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3743 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3747 * Removed the OS390-Unix config target. It relied on a script that doesn't
3752 * rsaz_1024_mul_avx2 overflow bug on x86_64
3754 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3755 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3756 Analysis suggests that attacks against RSA and DSA as a result of this
3757 defect would be very difficult to perform and are not believed likely.
3758 Attacks against DH1024 are considered just feasible, because most of the
3759 work necessary to deduce information about a private key may be performed
3760 offline. The amount of resources required for such an attack would be
3761 significant. However, for an attack on TLS to be meaningful, the server
3762 would have to share the DH1024 private key among multiple clients, which is
3763 no longer an option since CVE-2016-0701.
3765 This only affects processors that support the AVX2 but not ADX extensions
3766 like Intel Haswell (4th generation).
3768 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3769 was originally found via the OSS-Fuzz project.
3774 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3776 * bn_sqrx8x_internal carry bug on x86_64
3778 There is a carry propagating bug in the x86_64 Montgomery squaring
3779 procedure. No EC algorithms are affected. Analysis suggests that attacks
3780 against RSA and DSA as a result of this defect would be very difficult to
3781 perform and are not believed likely. Attacks against DH are considered just
3782 feasible (although very difficult) because most of the work necessary to
3783 deduce information about a private key may be performed offline. The amount
3784 of resources required for such an attack would be very significant and
3785 likely only accessible to a limited number of attackers. An attacker would
3786 additionally need online access to an unpatched system using the target
3787 private key in a scenario with persistent DH parameters and a private
3788 key that is shared between multiple clients.
3790 This only affects processors that support the BMI1, BMI2 and ADX extensions
3791 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3793 This issue was reported to OpenSSL by the OSS-Fuzz project.
3798 * Malformed X.509 IPAddressFamily could cause OOB read
3800 If an X.509 certificate has a malformed IPAddressFamily extension,
3801 OpenSSL could do a one-byte buffer overread. The most likely result
3802 would be an erroneous display of the certificate in text format.
3804 This issue was reported to OpenSSL by the OSS-Fuzz project.
3809 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3811 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3812 platform rather than 'mingw'.
3816 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3817 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3818 which is the minimum version we support.
3822 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3824 * Encrypt-Then-Mac renegotiation crash
3826 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3827 negotiated where it was not in the original handshake (or vice-versa) then
3828 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3829 and servers are affected.
3831 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3836 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3838 * Truncated packet could crash via OOB read
3840 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3841 cipher is being used, then a truncated packet can cause that host to
3842 perform an out-of-bounds read, usually resulting in a crash.
3844 This issue was reported to OpenSSL by Robert Święcki of Google.
3849 * Bad (EC)DHE parameters cause a client crash
3851 If a malicious server supplies bad parameters for a DHE or ECDHE key
3852 exchange then this can result in the client attempting to dereference a
3853 NULL pointer leading to a client crash. This could be exploited in a Denial
3856 This issue was reported to OpenSSL by Guido Vranken.
3861 * BN_mod_exp may produce incorrect results on x86_64
3863 There is a carry propagating bug in the x86_64 Montgomery squaring
3864 procedure. No EC algorithms are affected. Analysis suggests that attacks
3865 against RSA and DSA as a result of this defect would be very difficult to
3866 perform and are not believed likely. Attacks against DH are considered just
3867 feasible (although very difficult) because most of the work necessary to
3868 deduce information about a private key may be performed offline. The amount
3869 of resources required for such an attack would be very significant and
3870 likely only accessible to a limited number of attackers. An attacker would
3871 additionally need online access to an unpatched system using the target
3872 private key in a scenario with persistent DH parameters and a private
3873 key that is shared between multiple clients. For example this can occur by
3874 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3875 similar to CVE-2015-3193 but must be treated as a separate problem.
3877 This issue was reported to OpenSSL by the OSS-Fuzz project.
3882 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3884 * ChaCha20/Poly1305 heap-buffer-overflow
3886 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3887 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3888 crash. This issue is not considered to be exploitable beyond a DoS.
3890 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3895 * CMS Null dereference
3897 Applications parsing invalid CMS structures can crash with a NULL pointer
3898 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3899 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3900 structure callback if an attempt is made to free certain invalid encodings.
3901 Only CHOICE structures using a callback which do not handle NULL value are
3904 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3909 * Montgomery multiplication may produce incorrect results
3911 There is a carry propagating bug in the Broadwell-specific Montgomery
3912 multiplication procedure that handles input lengths divisible by, but
3913 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3914 and DH private keys are impossible. This is because the subroutine in
3915 question is not used in operations with the private key itself and an input
3916 of the attacker's direct choice. Otherwise the bug can manifest itself as
3917 transient authentication and key negotiation failures or reproducible
3918 erroneous outcome of public-key operations with specially crafted input.
3919 Among EC algorithms only Brainpool P-512 curves are affected and one
3920 presumably can attack ECDH key negotiation. Impact was not analyzed in
3921 detail, because pre-requisites for attack are considered unlikely. Namely
3922 multiple clients have to choose the curve in question and the server has to
3923 share the private key among them, neither of which is default behaviour.
3924 Even then only clients that chose the curve will be affected.
3926 This issue was publicly reported as transient failures and was not
3927 initially recognized as a security issue. Thanks to Richard Morgan for
3928 providing reproducible case.
3933 * Removed automatic addition of RPATH in shared libraries and executables,
3934 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3938 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3940 * Fix Use After Free for large message sizes
3942 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3943 message larger than approx 16k is received then the underlying buffer to
3944 store the incoming message is reallocated and moved. Unfortunately a
3945 dangling pointer to the old location is left which results in an attempt to
3946 write to the previously freed location. This is likely to result in a
3947 crash, however it could potentially lead to execution of arbitrary code.
3949 This issue only affects OpenSSL 1.1.0a.
3951 This issue was reported to OpenSSL by Robert Święcki.
3956 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3958 * OCSP Status Request extension unbounded memory growth
3960 A malicious client can send an excessively large OCSP Status Request
3961 extension. If that client continually requests renegotiation, sending a
3962 large OCSP Status Request extension each time, then there will be unbounded
3963 memory growth on the server. This will eventually lead to a Denial Of
3964 Service attack through memory exhaustion. Servers with a default
3965 configuration are vulnerable even if they do not support OCSP. Builds using
3966 the "no-ocsp" build time option are not affected.
3968 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3973 * SSL_peek() hang on empty record
3975 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3976 sends an empty record. This could be exploited by a malicious peer in a
3977 Denial Of Service attack.
3979 This issue was reported to OpenSSL by Alex Gaynor.
3984 * Excessive allocation of memory in tls_get_message_header() and
3985 dtls1_preprocess_fragment()
3987 A (D)TLS message includes 3 bytes for its length in the header for the
3988 message. This would allow for messages up to 16Mb in length. Messages of
3989 this length are excessive and OpenSSL includes a check to ensure that a
3990 peer is sending reasonably sized messages in order to avoid too much memory
3991 being consumed to service a connection. A flaw in the logic of version
3992 1.1.0 means that memory for the message is allocated too early, prior to
3993 the excessive message length check. Due to way memory is allocated in
3994 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3995 to service a connection. This could lead to a Denial of Service through
3996 memory exhaustion. However, the excessive message length check still takes
3997 place, and this would cause the connection to immediately fail. Assuming
3998 that the application calls SSL_free() on the failed connection in a timely
3999 manner then the 21Mb of allocated memory will then be immediately freed
4000 again. Therefore, the excessive memory allocation will be transitory in
4001 nature. This then means that there is only a security impact if:
4003 1) The application does not call SSL_free() in a timely manner in the event
4004 that the connection fails
4006 2) The application is working in a constrained environment where there is
4007 very little free memory
4009 3) The attacker initiates multiple connection attempts such that there are
4010 multiple connections in a state where memory has been allocated for the
4011 connection; SSL_free() has not yet been called; and there is insufficient
4012 memory to service the multiple requests.
4014 Except in the instance of (1) above any Denial Of Service is likely to be
4015 transitory because as soon as the connection fails the memory is
4016 subsequently freed again in the SSL_free() call. However there is an
4017 increased risk during this period of application crashes due to the lack of
4018 memory - which would then mean a more serious Denial of Service.
4020 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4021 (CVE-2016-6307 and CVE-2016-6308)
4025 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4026 had to be removed. Primary reason is that vendor assembler can't
4027 assemble our modules with -KPIC flag. As result it, assembly
4028 support, was not even available as option. But its lack means
4029 lack of side-channel resistant code, which is incompatible with
4030 security by todays standards. Fortunately gcc is readily available
4031 prepackaged option, which we firmly point at...
4035 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4037 * Windows command-line tool supports UTF-8 opt-in option for arguments
4038 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4039 (to any value) allows Windows user to access PKCS#12 file generated
4040 with Windows CryptoAPI and protected with non-ASCII password, as well
4041 as files generated under UTF-8 locale on Linux also protected with
4046 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4047 have been disabled by default and removed from DEFAULT, just like RC4.
4048 See the RC4 item below to re-enable both.
4052 * The method for finding the storage location for the Windows RAND seed file
4053 has changed. First we check %RANDFILE%. If that is not set then we check
4054 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4055 all else fails we fall back to C:\.
4059 * The EVP_EncryptUpdate() function has had its return type changed from void
4060 to int. A return of 0 indicates and error while a return of 1 indicates
4065 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4066 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4067 off the constant time implementation for RSA, DSA and DH have been made
4068 no-ops and deprecated.
4072 * Windows RAND implementation was simplified to only get entropy by
4073 calling CryptGenRandom(). Various other RAND-related tickets
4076 *Joseph Wylie Yandle, Rich Salz*
4078 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4079 and `OPENSSL_LH_`, respectively. The old names are available
4080 with API compatibility. They new names are now completely documented.
4084 * Unify TYPE_up_ref(obj) methods signature.
4085 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4086 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4087 int (instead of void) like all others TYPE_up_ref() methods.
4088 So now these methods also check the return value of CRYPTO_atomic_add(),
4089 and the validity of object reference counter.
4091 *fdasilvayy@gmail.com*
4093 * With Windows Visual Studio builds, the .pdb files are installed
4094 alongside the installed libraries and executables. For a static
4095 library installation, ossl_static.pdb is the associate compiler
4096 generated .pdb file to be used when linking programs.
4100 * Remove openssl.spec. Packaging files belong with the packagers.
4104 * Automatic Darwin/OSX configuration has had a refresh, it will now
4105 recognise x86_64 architectures automatically. You can still decide
4106 to build for a different bitness with the environment variable
4107 KERNEL_BITS (can be 32 or 64), for example:
4109 KERNEL_BITS=32 ./config
4113 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4114 256 bit AES and HMAC with SHA256.
4118 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4122 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4126 * To enable users to have their own config files and build file templates,
4127 Configure looks in the directory indicated by the environment variable
4128 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4129 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4130 name and is used as is.
4134 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4135 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4136 X509_CERT_FILE_CTX was removed.
4140 * "shared" builds are now the default. To create only static libraries use
4141 the "no-shared" Configure option.
4145 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4146 All of these option have not worked for some while and are fundamental
4151 * Make various cleanup routines no-ops and mark them as deprecated. Most
4152 global cleanup functions are no longer required because they are handled
4153 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4154 Explicitly de-initing can cause problems (e.g. where a library that uses
4155 OpenSSL de-inits, but an application is still using it). The affected
4156 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4157 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4158 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4159 COMP_zlib_cleanup().
4163 * --strict-warnings no longer enables runtime debugging options
4164 such as REF_DEBUG. Instead, debug options are automatically
4165 enabled with '--debug' builds.
4167 *Andy Polyakov, Emilia Käsper*
4169 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4170 have been moved out of the public header files. New functions for managing
4171 these have been added.
4175 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4176 objects have been moved out of the public header files. New
4177 functions for managing these have been added.
4181 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4182 have been moved out of the public header files. New functions for managing
4183 these have been added.
4187 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4188 moved out of the public header files. New functions for managing these
4193 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4197 * Removed the mk1mf build scripts.
4201 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4202 it is always safe to #include a header now.
4206 * Removed the aged BC-32 config and all its supporting scripts
4210 * Removed support for Ultrix, Netware, and OS/2.
4214 * Add support for HKDF.
4216 *Alessandro Ghedini*
4218 * Add support for blake2b and blake2s
4222 * Added support for "pipelining". Ciphers that have the
4223 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4224 encryptions/decryptions simultaneously. There are currently no built-in
4225 ciphers with this property but the expectation is that engines will be able
4226 to offer it to significantly improve throughput. Support has been extended
4227 into libssl so that multiple records for a single connection can be
4228 processed in one go (for >=TLS 1.1).
4232 * Added the AFALG engine. This is an async capable engine which is able to
4233 offload work to the Linux kernel. In this initial version it only supports
4234 AES128-CBC. The kernel must be version 4.1.0 or greater.
4238 * OpenSSL now uses a new threading API. It is no longer necessary to
4239 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4240 are two supported threading models: pthreads and windows threads. It is
4241 also possible to configure OpenSSL at compile time for "no-threads". The
4242 old threading API should no longer be used. The functions have been
4243 replaced with "no-op" compatibility macros.
4245 *Alessandro Ghedini, Matt Caswell*
4247 * Modify behavior of ALPN to invoke callback after SNI/servername
4248 callback, such that updates to the SSL_CTX affect ALPN.
4252 * Add SSL_CIPHER queries for authentication and key-exchange.
4256 * Changes to the DEFAULT cipherlist:
4257 - Prefer (EC)DHE handshakes over plain RSA.
4258 - Prefer AEAD ciphers over legacy ciphers.
4259 - Prefer ECDSA over RSA when both certificates are available.
4260 - Prefer TLSv1.2 ciphers/PRF.
4261 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4266 * Change the ECC default curve list to be this, in order: x25519,
4267 secp256r1, secp521r1, secp384r1.
4271 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4272 disabled by default. They can be re-enabled using the
4273 enable-weak-ssl-ciphers option to Configure.
4277 * If the server has ALPN configured, but supports no protocols that the
4278 client advertises, send a fatal "no_application_protocol" alert.
4279 This behaviour is SHALL in RFC 7301, though it isn't universally
4280 implemented by other servers.
4284 * Add X25519 support.
4285 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4286 for public and private key encoding using the format documented in
4287 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4288 key generation and key derivation.
4290 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4295 * Deprecate SRP_VBASE_get_by_user.
4296 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4297 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4298 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4299 seed, even if the seed is configured.
4301 Users should use SRP_VBASE_get1_by_user instead. Note that in
4302 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4303 also that even though configuring the SRP seed attempts to hide
4304 invalid usernames by continuing the handshake with fake
4305 credentials, this behaviour is not constant time and no strong
4306 guarantees are made that the handshake is indistinguishable from
4307 that of a valid user.
4311 * Configuration change; it's now possible to build dynamic engines
4312 without having to build shared libraries and vice versa. This
4313 only applies to the engines in `engines/`, those in `crypto/engine/`
4314 will always be built into libcrypto (i.e. "static").
4316 Building dynamic engines is enabled by default; to disable, use
4317 the configuration option "disable-dynamic-engine".
4319 The only requirements for building dynamic engines are the
4320 presence of the DSO module and building with position independent
4321 code, so they will also automatically be disabled if configuring
4322 with "disable-dso" or "disable-pic".
4324 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4325 are also taken away from openssl/opensslconf.h, as they are
4330 * Configuration change; if there is a known flag to compile
4331 position independent code, it will always be applied on the
4332 libcrypto and libssl object files, and never on the application
4333 object files. This means other libraries that use routines from
4334 libcrypto / libssl can be made into shared libraries regardless
4335 of how OpenSSL was configured.
4337 If this isn't desirable, the configuration options "disable-pic"
4338 or "no-pic" can be used to disable the use of PIC. This will
4339 also disable building shared libraries and dynamic engines.
4343 * Removed JPAKE code. It was experimental and has no wide use.
4347 * The INSTALL_PREFIX Makefile variable has been renamed to
4348 DESTDIR. That makes for less confusion on what this variable
4349 is for. Also, the configuration option --install_prefix is
4354 * Heartbeat for TLS has been removed and is disabled by default
4355 for DTLS; configure with enable-heartbeats. Code that uses the
4356 old #define's might need to be updated.
4358 *Emilia Käsper, Rich Salz*
4360 * Rename REF_CHECK to REF_DEBUG.
4364 * New "unified" build system
4366 The "unified" build system is aimed to be a common system for all
4367 platforms we support. With it comes new support for VMS.
4369 This system builds supports building in a different directory tree
4370 than the source tree. It produces one Makefile (for unix family
4371 or lookalikes), or one descrip.mms (for VMS).
4373 The source of information to make the Makefile / descrip.mms is
4374 small files called 'build.info', holding the necessary
4375 information for each directory with source to compile, and a
4376 template in Configurations, like unix-Makefile.tmpl or
4379 With this change, the library names were also renamed on Windows
4380 and on VMS. They now have names that are closer to the standard
4381 on Unix, and include the major version number, and in certain
4382 cases, the architecture they are built for. See "Notes on shared
4383 libraries" in INSTALL.
4385 We rely heavily on the perl module Text::Template.
4389 * Added support for auto-initialisation and de-initialisation of the library.
4390 OpenSSL no longer requires explicit init or deinit routines to be called,
4391 except in certain circumstances. See the OPENSSL_init_crypto() and
4392 OPENSSL_init_ssl() man pages for further information.
4396 * The arguments to the DTLSv1_listen function have changed. Specifically the
4397 "peer" argument is now expected to be a BIO_ADDR object.
4399 * Rewrite of BIO networking library. The BIO library lacked consistent
4400 support of IPv6, and adding it required some more extensive
4401 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4402 which hold all types of addresses and chains of address information.
4403 It also introduces a new API, with functions like BIO_socket,
4404 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4405 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4406 have been adapted accordingly.
4410 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4415 * CRIME protection: disable compression by default, even if OpenSSL is
4416 compiled with zlib enabled. Applications can still enable compression
4417 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4418 using the SSL_CONF library to configure compression.
4422 * The signature of the session callback configured with
4423 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4424 was explicitly marked as `const unsigned char*` instead of
4429 * Always DPURIFY. Remove the use of uninitialized memory in the
4430 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4434 * Removed many obsolete configuration items, including
4435 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4436 MD2_CHAR, MD2_INT, MD2_LONG
4438 IDEA_SHORT, IDEA_LONG
4439 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4441 *Rich Salz, with advice from Andy Polyakov*
4443 * Many BN internals have been moved to an internal header file.
4445 *Rich Salz with help from Andy Polyakov*
4447 * Configuration and writing out the results from it has changed.
4448 Files such as Makefile include/openssl/opensslconf.h and are now
4449 produced through general templates, such as Makefile.in and
4450 crypto/opensslconf.h.in and some help from the perl module
4453 Also, the center of configuration information is no longer
4454 Makefile. Instead, Configure produces a perl module in
4455 configdata.pm which holds most of the config data (in the hash
4456 table %config), the target data that comes from the target
4457 configuration in one of the `Configurations/*.conf` files (in
4462 * To clarify their intended purposes, the Configure options
4463 --prefix and --openssldir change their semantics, and become more
4464 straightforward and less interdependent.
4466 --prefix shall be used exclusively to give the location INSTALLTOP
4467 where programs, scripts, libraries, include files and manuals are
4468 going to be installed. The default is now /usr/local.
4470 --openssldir shall be used exclusively to give the default
4471 location OPENSSLDIR where certificates, private keys, CRLs are
4472 managed. This is also where the default openssl.cnf gets
4474 If the directory given with this option is a relative path, the
4475 values of both the --prefix value and the --openssldir value will
4476 be combined to become OPENSSLDIR.
4477 The default for --openssldir is INSTALLTOP/ssl.
4479 Anyone who uses --openssldir to specify where OpenSSL is to be
4480 installed MUST change to use --prefix instead.
4484 * The GOST engine was out of date and therefore it has been removed. An up
4485 to date GOST engine is now being maintained in an external repository.
4486 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4487 support for GOST ciphersuites (these are only activated if a GOST engine
4492 * EGD is no longer supported by default; use enable-egd when
4495 *Ben Kaduk and Rich Salz*
4497 * The distribution now has Makefile.in files, which are used to
4498 create Makefile's when Configure is run. *Configure must be run
4499 before trying to build now.*
4503 * The return value for SSL_CIPHER_description() for error conditions
4508 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4510 Obtaining and performing DNSSEC validation of TLSA records is
4511 the application's responsibility. The application provides
4512 the TLSA records of its choice to OpenSSL, and these are then
4513 used to authenticate the peer.
4515 The TLSA records need not even come from DNS. They can, for
4516 example, be used to implement local end-entity certificate or
4517 trust-anchor "pinning", where the "pin" data takes the form
4518 of TLSA records, which can augment or replace verification
4519 based on the usual WebPKI public certification authorities.
4523 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4524 continues to support deprecated interfaces in default builds.
4525 However, applications are strongly advised to compile their
4526 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4527 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4528 or the 1.1.0 releases.
4530 In environments in which all applications have been ported to
4531 not use any deprecated interfaces OpenSSL's Configure script
4532 should be used with the --api=1.1.0 option to entirely remove
4533 support for the deprecated features from the library and
4534 unconditionally disable them in the installed headers.
4535 Essentially the same effect can be achieved with the "no-deprecated"
4536 argument to Configure, except that this will always restrict
4537 the build to just the latest API, rather than a fixed API
4540 As applications are ported to future revisions of the API,
4541 they should update their compile-time OPENSSL_API_COMPAT define
4542 accordingly, but in most cases should be able to continue to
4543 compile with later releases.
4545 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4546 0x10000000L and 0x00908000L, respectively. However those
4547 versions did not support the OPENSSL_API_COMPAT feature, and
4548 so applications are not typically tested for explicit support
4549 of just the undeprecated features of either release.
4553 * Add support for setting the minimum and maximum supported protocol.
4554 It can bet set via the SSL_set_min_proto_version() and
4555 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4556 MaxProtocol. It's recommended to use the new APIs to disable
4557 protocols instead of disabling individual protocols using
4558 SSL_set_options() or SSL_CONF's Protocol. This change also
4559 removes support for disabling TLS 1.2 in the OpenSSL TLS
4560 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4564 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4568 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4569 and integrates ECDSA and ECDH functionality into EC. Implementations can
4570 now redirect key generation and no longer need to convert to or from
4573 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4574 include the ec.h header file instead.
4578 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4579 ciphers who are no longer supported and drops support the ephemeral RSA key
4580 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4584 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4585 opaque. For HMAC_CTX, the following constructors and destructors
4588 HMAC_CTX *HMAC_CTX_new(void);
4589 void HMAC_CTX_free(HMAC_CTX *ctx);
4591 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4592 destroy such methods has been added. See EVP_MD_meth_new(3) and
4593 EVP_CIPHER_meth_new(3) for documentation.
4596 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4597 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4598 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4599 an already created structure.
4600 2) For consistency with the majority of our object creators and
4601 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4602 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4603 for deprecated builds.
4607 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4608 cryptographic operations to be performed asynchronously as long as an
4609 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4610 further details. Libssl has also had this capability integrated with the
4611 introduction of the new mode SSL_MODE_ASYNC and associated error
4612 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4613 pages. This work was developed in partnership with Intel Corp.
4617 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4618 always enabled now. If you want to disable the support you should
4619 exclude it using the list of supported ciphers. This also means that the
4620 "-no_ecdhe" option has been removed from s_server.
4624 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4625 SSL_{CTX_}set1_curves() which can set a list.
4629 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4630 curve you want to support using SSL_{CTX_}set1_curves().
4634 * State machine rewrite. The state machine code has been significantly
4635 refactored in order to remove much duplication of code and solve issues
4636 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4637 further details). This change does have some associated API changes.
4638 Notably the SSL_state() function has been removed and replaced by
4639 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4640 SSL_set_state() has been removed altogether. The previous handshake states
4641 defined in ssl.h and ssl3.h have also been removed.
4645 * All instances of the string "ssleay" in the public API were replaced
4646 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4647 Some error codes related to internal RSA_eay API's were renamed.
4651 * The demo files in crypto/threads were moved to demo/threads.
4655 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4658 *Matt Caswell, Rich Salz*
4660 * New ASN.1 embed macro.
4662 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4663 structure is not allocated: it is part of the parent. That is instead of
4671 This reduces memory fragmentation and make it impossible to accidentally
4672 set a mandatory field to NULL.
4674 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4675 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4676 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4681 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4685 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4686 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4687 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4688 DES and RC4 ciphersuites.
4692 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4693 This changes the decoding behaviour for some invalid messages,
4694 though the change is mostly in the more lenient direction, and
4695 legacy behaviour is preserved as much as possible.
4699 * Fix no-stdio build.
4700 *David Woodhouse <David.Woodhouse@intel.com> and also*
4701 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4703 * New testing framework
4704 The testing framework has been largely rewritten and is now using
4705 perl and the perl modules Test::Harness and an extended variant of
4706 Test::More called OpenSSL::Test to do its work. All test scripts in
4707 test/ have been rewritten into test recipes, and all direct calls to
4708 executables in test/Makefile have become individual recipes using the
4709 simplified testing OpenSSL::Test::Simple.
4711 For documentation on our testing modules, do:
4713 perldoc test/testlib/OpenSSL/Test/Simple.pm
4714 perldoc test/testlib/OpenSSL/Test.pm
4718 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4719 are used; the latter aborts on memory leaks (usually checked on exit).
4720 Some undocumented "set malloc, etc., hooks" functions were removed
4721 and others were changed. All are now documented.
4725 * In DSA_generate_parameters_ex, if the provided seed is too short,
4728 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4730 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4731 from RFC4279, RFC4785, RFC5487, RFC5489.
4733 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4734 original RSA_PSK patch.
4738 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4739 era flag was never set throughout the codebase (only read). Also removed
4740 SSL3_FLAGS_POP_BUFFER which was only used if
4741 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4745 * Changed the default name options in the "ca", "crl", "req" and "x509"
4746 to be "oneline" instead of "compat".
4750 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4751 not aware of clients that still exhibit this bug, and the workaround
4752 hasn't been working properly for a while.
4756 * The return type of BIO_number_read() and BIO_number_written() as well as
4757 the corresponding num_read and num_write members in the BIO structure has
4758 changed from unsigned long to uint64_t. On platforms where an unsigned
4759 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4764 * Given the pervasive nature of TLS extensions it is inadvisable to run
4765 OpenSSL without support for them. It also means that maintaining
4766 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4767 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4771 * Removed support for the two export grade static DH ciphersuites
4772 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4773 were newly added (along with a number of other static DH ciphersuites) to
4774 1.0.2. However the two export ones have *never* worked since they were
4775 introduced. It seems strange in any case to be adding new export
4776 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4780 * Version negotiation has been rewritten. In particular SSLv23_method(),
4781 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4782 and turned into macros which simply call the new preferred function names
4783 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4784 should use the new names instead. Also as part of this change the ssl23.h
4785 header file has been removed.
4789 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4790 code and the associated standard is no longer considered fit-for-purpose.
4794 * RT2547 was closed. When generating a private key, try to make the
4795 output file readable only by the owner. This behavior change might
4796 be noticeable when interacting with other software.
4798 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4803 * Added HTTP GET support to the ocsp command.
4807 * Changed default digest for the dgst and enc commands from MD5 to
4812 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4816 * Added support for TLS extended master secret from
4817 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4818 initial patch which was a great help during development.
4822 * All libssl internal structures have been removed from the public header
4823 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4824 now redundant). Users should not attempt to access internal structures
4825 directly. Instead they should use the provided API functions.
4829 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4830 Access to deprecated functions can be re-enabled by running config with
4831 "enable-deprecated". In addition applications wishing to use deprecated
4832 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4833 will, by default, disable some transitive includes that previously existed
4834 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4838 * Added support for OCB mode. OpenSSL has been granted a patent license
4839 compatible with the OpenSSL license for use of OCB. Details are available
4840 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4841 for OCB can be removed by calling config with no-ocb.
4845 * SSLv2 support has been removed. It still supports receiving an SSLv2
4846 compatible client hello.
4850 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4851 done while fixing the error code for the key-too-small case.
4853 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4855 * CA.sh has been removed; use CA.pl instead.
4859 * Removed old DES API.
4863 * Remove various unsupported platforms:
4869 Sinix/ReliantUNIX RM400
4874 16-bit platforms such as WIN16
4878 * Clean up OPENSSL_NO_xxx #define's
4879 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4880 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4881 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4882 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4883 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4884 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4885 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4886 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4887 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4888 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4892 * Cleaned up dead code
4893 Remove all but one '#ifdef undef' which is to be looked at.
4897 * Clean up calling of xxx_free routines.
4898 Just like free(), fix most of the xxx_free routines to accept
4899 NULL. Remove the non-null checks from callers. Save much code.
4903 * Add secure heap for storage of private keys (when possible).
4904 Add BIO_s_secmem(), CBIGNUM, etc.
4905 Contributed by Akamai Technologies under our Corporate CLA.
4909 * Experimental support for a new, fast, unbiased prime candidate generator,
4910 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4912 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4914 * New output format NSS in the sess_id command line tool. This allows
4915 exporting the session id and the master key in NSS keylog format.
4917 *Martin Kaiser <martin@kaiser.cx>*
4919 * Harmonize version and its documentation. -f flag is used to display
4922 *mancha <mancha1@zoho.com>*
4924 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4925 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4927 *mancha <mancha1@zoho.com>*
4929 * Fix some double frees. These are not thought to be exploitable.
4931 *mancha <mancha1@zoho.com>*
4933 * A missing bounds check in the handling of the TLS heartbeat extension
4934 can be used to reveal up to 64k of memory to a connected client or
4937 Thanks for Neel Mehta of Google Security for discovering this bug and to
4938 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4939 preparing the fix ([CVE-2014-0160])
4941 *Adam Langley, Bodo Moeller*
4943 * Fix for the attack described in the paper "Recovering OpenSSL
4944 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4945 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4946 <http://eprint.iacr.org/2014/140>
4948 Thanks to Yuval Yarom and Naomi Benger for discovering this
4949 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4951 *Yuval Yarom and Naomi Benger*
4953 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4954 this fixes a limitation in previous versions of OpenSSL.
4958 * Experimental encrypt-then-mac support.
4960 Experimental support for encrypt then mac from
4961 draft-gutmann-tls-encrypt-then-mac-02.txt
4963 To enable it set the appropriate extension number (0x42 for the test
4964 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4966 For non-compliant peers (i.e. just about everything) this should have no
4969 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4973 * Add EVP support for key wrapping algorithms, to avoid problems with
4974 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4975 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4976 algorithms and include tests cases.
4980 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4985 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4986 MGF1 digest and OAEP label.
4990 * Make openssl verify return errors.
4992 *Chris Palmer <palmer@google.com> and Ben Laurie*
4994 * New function ASN1_TIME_diff to calculate the difference between two
4995 ASN1_TIME structures or one structure and the current time.
4999 * Update fips_test_suite to support multiple command line options. New
5000 test to induce all self test errors in sequence and check expected
5005 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5006 sign or verify all in one operation.
5010 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5011 test programs and fips_test_suite. Includes functionality to parse
5012 the minimal script output of fipsalgest.pl directly.
5016 * Add authorisation parameter to FIPS_module_mode_set().
5020 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5024 * Use separate DRBG fields for internal and external flags. New function
5025 FIPS_drbg_health_check() to perform on demand health checking. Add
5026 generation tests to fips_test_suite with reduced health check interval to
5027 demonstrate periodic health checking. Add "nodh" option to
5028 fips_test_suite to skip very slow DH test.
5032 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5037 * More extensive health check for DRBG checking many more failure modes.
5038 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5039 combination: call this in fips_test_suite.
5043 * Add support for canonical generation of DSA parameter 'g'. See
5046 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5047 POST to handle HMAC cases.
5051 * Add functions FIPS_module_version() and FIPS_module_version_text()
5052 to return numerical and string versions of the FIPS module number.
5056 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5057 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5058 outside the validated module in the FIPS capable OpenSSL.
5062 * Minor change to DRBG entropy callback semantics. In some cases
5063 there is no multiple of the block length between min_len and
5064 max_len. Allow the callback to return more than max_len bytes
5065 of entropy but discard any extra: it is the callback's responsibility
5066 to ensure that the extra data discarded does not impact the
5067 requested amount of entropy.
5071 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5072 information in FIPS186-3, SP800-57 and SP800-131A.
5076 * CCM support via EVP. Interface is very similar to GCM case except we
5077 must supply all data in one chunk (i.e. no update, final) and the
5078 message length must be supplied if AAD is used. Add algorithm test
5083 * Initial version of POST overhaul. Add POST callback to allow the status
5084 of POST to be monitored and/or failures induced. Modify fips_test_suite
5085 to use callback. Always run all selftests even if one fails.
5089 * XTS support including algorithm test driver in the fips_gcmtest program.
5090 Note: this does increase the maximum key length from 32 to 64 bytes but
5091 there should be no binary compatibility issues as existing applications
5092 will never use XTS mode.
5096 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5097 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5098 performs algorithm blocking for unapproved PRNG types. Also do not
5099 set PRNG type in FIPS_mode_set(): leave this to the application.
5100 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5101 the standard OpenSSL PRNG: set additional data to a date time vector.
5105 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5106 This shouldn't present any incompatibility problems because applications
5107 shouldn't be using these directly and any that are will need to rethink
5108 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5112 * Extensive self tests and health checking required by SP800-90 DRBG.
5113 Remove strength parameter from FIPS_drbg_instantiate and always
5114 instantiate at maximum supported strength.
5118 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5122 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5126 * New function DH_compute_key_padded() to compute a DH key and pad with
5127 leading zeroes if needed: this complies with SP800-56A et al.
5131 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5132 anything, incomplete, subject to change and largely untested at present.
5136 * Modify fipscanisteronly build option to only build the necessary object
5137 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5141 * Add experimental option FIPSSYMS to give all symbols in
5142 fipscanister.o and FIPS or fips prefix. This will avoid
5143 conflicts with future versions of OpenSSL. Add perl script
5144 util/fipsas.pl to preprocess assembly language source files
5145 and rename any affected symbols.
5149 * Add selftest checks and algorithm block of non-fips algorithms in
5150 FIPS mode. Remove DES2 from selftests.
5154 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5155 return internal method without any ENGINE dependencies. Add new
5156 tiny fips sign and verify functions.
5160 * New build option no-ec2m to disable characteristic 2 code.
5164 * New build option "fipscanisteronly". This only builds fipscanister.o
5165 and (currently) associated fips utilities. Uses the file Makefile.fips
5166 instead of Makefile.org as the prototype.
5170 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5171 Update fips_gcmtest to use IV generator.
5175 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5176 setting output buffer to NULL. The `*Final` function must be
5177 called although it will not retrieve any additional data. The tag
5178 can be set or retrieved with a ctrl. The IV length is by default 12
5179 bytes (96 bits) but can be set to an alternative value. If the IV
5180 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5185 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5186 underlying do_cipher function handles all cipher semantics itself
5187 including padding and finalisation. This is useful if (for example)
5188 an ENGINE cipher handles block padding itself. The behaviour of
5189 do_cipher is subtly changed if this flag is set: the return value
5190 is the number of characters written to the output buffer (zero is
5191 no longer an error code) or a negative error code. Also if the
5192 input buffer is NULL and length 0 finalisation should be performed.
5196 * If a candidate issuer certificate is already part of the constructed
5197 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5201 * Improve forward-security support: add functions
5203 void SSL_CTX_set_not_resumable_session_callback(
5204 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5205 void SSL_set_not_resumable_session_callback(
5206 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5208 for use by SSL/TLS servers; the callback function will be called whenever a
5209 new session is created, and gets to decide whether the session may be
5210 cached to make it resumable (return 0) or not (return 1). (As by the
5211 SSL/TLS protocol specifications, the session_id sent by the server will be
5212 empty to indicate that the session is not resumable; also, the server will
5213 not generate RFC 4507 (RFC 5077) session tickets.)
5215 A simple reasonable callback implementation is to return is_forward_secure.
5216 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5217 by the SSL/TLS server library, indicating whether it can provide forward
5220 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5222 * New -verify_name option in command line utilities to set verification
5227 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5228 Add CMAC pkey methods.
5232 * Experimental renegotiation in s_server -www mode. If the client
5233 browses /reneg connection is renegotiated. If /renegcert it is
5234 renegotiated requesting a certificate.
5238 * Add an "external" session cache for debugging purposes to s_server. This
5239 should help trace issues which normally are only apparent in deployed
5240 multi-process servers.
5244 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5245 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5246 BIO_set_cipher() and some obscure PEM functions were changed so they
5247 can now return an error. The RAND changes required a change to the
5248 RAND_METHOD structure.
5252 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5253 a gcc attribute to warn if the result of a function is ignored. This
5254 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5255 whose return value is often ignored.
5259 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5260 These allow SCTs (signed certificate timestamps) to be requested and
5261 validated when establishing a connection.
5263 *Rob Percival <robpercival@google.com>*
5268 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5270 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5271 used even when parsing explicit parameters, when loading a encoded key
5272 or calling `EC_GROUP_new_from_ecpkparameters()`/
5273 `EC_GROUP_new_from_ecparameters()`.
5274 This prevents bypass of security hardening and performance gains,
5275 especially for curves with specialized EC_METHODs.
5276 By default, if a key encoded with explicit parameters is loaded and later
5277 encoded, the output is still encoded with explicit parameters, even if
5278 internally a "named" EC_GROUP is used for computation.
5282 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5283 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5284 NULL. After this change, only the cofactor parameter can be NULL. It also
5285 does some minimal sanity checks on the passed order.
5290 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5291 An attack is simple, if the first CMS_recipientInfo is valid but the
5292 second CMS_recipientInfo is chosen ciphertext. If the second
5293 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5294 encryption key will be replaced by garbage, and the message cannot be
5295 decoded, but if the RSA decryption fails, the correct encryption key is
5296 used and the recipient will not notice the attack.
5297 As a work around for this potential attack the length of the decrypted
5298 key must be equal to the cipher default key length, in case the
5299 certificate is not given and all recipientInfo are tried out.
5300 The old behaviour can be re-enabled in the CMS code by setting the
5301 CMS_DEBUG_DECRYPT flag.
5306 * Document issue with installation paths in diverse Windows builds
5308 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5309 binaries and run-time config file.
5314 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5316 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5317 This changes the size when using the `genpkey` command when no size is given.
5318 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5319 generation commands to use 2048 bits by default.
5323 * Add FIPS support for Android Arm 64-bit
5325 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5326 Module in Version 2.0.10. For some reason, the corresponding target
5327 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5328 built with FIPS support on Android Arm 64-bit. This omission has been
5331 *Matthias St. Pierre*
5333 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5335 * 0-byte record padding oracle
5337 If an application encounters a fatal protocol error and then calls
5338 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5339 then OpenSSL can respond differently to the calling application if a 0 byte
5340 record is received with invalid padding compared to if a 0 byte record is
5341 received with an invalid MAC. If the application then behaves differently
5342 based on that in a way that is detectable to the remote peer, then this
5343 amounts to a padding oracle that could be used to decrypt data.
5345 In order for this to be exploitable "non-stitched" ciphersuites must be in
5346 use. Stitched ciphersuites are optimised implementations of certain
5347 commonly used ciphersuites. Also the application must call SSL_shutdown()
5348 twice even if a protocol error has occurred (applications should not do
5349 this but some do anyway).
5351 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5352 Aviram, with additional investigation by Steven Collison and Andrew
5353 Hourselt. It was reported to OpenSSL on 10th December 2018.
5358 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5362 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5364 * Microarchitecture timing vulnerability in ECC scalar multiplication
5366 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5367 shown to be vulnerable to a microarchitecture timing side channel attack.
5368 An attacker with sufficient access to mount local timing attacks during
5369 ECDSA signature generation could recover the private key.
5371 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5372 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5378 * Timing vulnerability in DSA signature generation
5380 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5381 timing side channel attack. An attacker could use variations in the signing
5382 algorithm to recover the private key.
5384 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5389 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5390 Module, accidentally introduced while backporting security fixes from the
5391 development branch and hindering the use of ECC in FIPS mode.
5395 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5397 * Client DoS due to large DH parameter
5399 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5400 malicious server can send a very large prime value to the client. This will
5401 cause the client to spend an unreasonably long period of time generating a
5402 key for this prime resulting in a hang until the client has finished. This
5403 could be exploited in a Denial Of Service attack.
5405 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5410 * Cache timing vulnerability in RSA Key Generation
5412 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5413 a cache timing side channel attack. An attacker with sufficient access to
5414 mount cache timing attacks during the RSA key generation process could
5415 recover the private key.
5417 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5418 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5423 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5424 parameter is no longer accepted, as it leads to a corrupt table. NULL
5425 pem_str is reserved for alias entries only.
5429 * Revert blinding in ECDSA sign and instead make problematic addition
5430 length-invariant. Switch even to fixed-length Montgomery multiplication.
5434 * Change generating and checking of primes so that the error rate of not
5435 being prime depends on the intended use based on the size of the input.
5436 For larger primes this will result in more rounds of Miller-Rabin.
5437 The maximal error rate for primes with more than 1080 bits is lowered
5440 *Kurt Roeckx, Annie Yousar*
5442 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5446 * Add blinding to ECDSA and DSA signatures to protect against side channel
5447 attacks discovered by Keegan Ryan (NCC Group).
5451 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5452 now allow empty (zero character) pass phrases.
5456 * Certificate time validation (X509_cmp_time) enforces stricter
5457 compliance with RFC 5280. Fractional seconds and timezone offsets
5458 are no longer allowed.
5462 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5464 * Constructed ASN.1 types with a recursive definition could exceed the stack
5466 Constructed ASN.1 types with a recursive definition (such as can be found
5467 in PKCS7) could eventually exceed the stack given malicious input with
5468 excessive recursion. This could result in a Denial Of Service attack. There
5469 are no such structures used within SSL/TLS that come from untrusted sources
5470 so this is considered safe.
5472 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5478 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5480 * Read/write after SSL object in error state
5482 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5483 mechanism. The intent was that if a fatal error occurred during a handshake
5484 then OpenSSL would move into the error state and would immediately fail if
5485 you attempted to continue the handshake. This works as designed for the
5486 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5487 SSL_connect()), however due to a bug it does not work correctly if
5488 SSL_read() or SSL_write() is called directly. In that scenario, if the
5489 handshake fails then a fatal error will be returned in the initial function
5490 call. If SSL_read()/SSL_write() is subsequently called by the application
5491 for the same SSL object then it will succeed and the data is passed without
5492 being decrypted/encrypted directly from the SSL/TLS record layer.
5494 In order to exploit this issue an application bug would have to be present
5495 that resulted in a call to SSL_read()/SSL_write() being issued after having
5496 already received a fatal error.
5498 This issue was reported to OpenSSL by David Benjamin (Google).
5503 * rsaz_1024_mul_avx2 overflow bug on x86_64
5505 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5506 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5507 Analysis suggests that attacks against RSA and DSA as a result of this
5508 defect would be very difficult to perform and are not believed likely.
5509 Attacks against DH1024 are considered just feasible, because most of the
5510 work necessary to deduce information about a private key may be performed
5511 offline. The amount of resources required for such an attack would be
5512 significant. However, for an attack on TLS to be meaningful, the server
5513 would have to share the DH1024 private key among multiple clients, which is
5514 no longer an option since CVE-2016-0701.
5516 This only affects processors that support the AVX2 but not ADX extensions
5517 like Intel Haswell (4th generation).
5519 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5520 was originally found via the OSS-Fuzz project.
5525 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5527 * bn_sqrx8x_internal carry bug on x86_64
5529 There is a carry propagating bug in the x86_64 Montgomery squaring
5530 procedure. No EC algorithms are affected. Analysis suggests that attacks
5531 against RSA and DSA as a result of this defect would be very difficult to
5532 perform and are not believed likely. Attacks against DH are considered just
5533 feasible (although very difficult) because most of the work necessary to
5534 deduce information about a private key may be performed offline. The amount
5535 of resources required for such an attack would be very significant and
5536 likely only accessible to a limited number of attackers. An attacker would
5537 additionally need online access to an unpatched system using the target
5538 private key in a scenario with persistent DH parameters and a private
5539 key that is shared between multiple clients.
5541 This only affects processors that support the BMI1, BMI2 and ADX extensions
5542 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5544 This issue was reported to OpenSSL by the OSS-Fuzz project.
5549 * Malformed X.509 IPAddressFamily could cause OOB read
5551 If an X.509 certificate has a malformed IPAddressFamily extension,
5552 OpenSSL could do a one-byte buffer overread. The most likely result
5553 would be an erroneous display of the certificate in text format.
5555 This issue was reported to OpenSSL by the OSS-Fuzz project.
5559 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5561 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5562 platform rather than 'mingw'.
5566 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5568 * Truncated packet could crash via OOB read
5570 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5571 cipher is being used, then a truncated packet can cause that host to
5572 perform an out-of-bounds read, usually resulting in a crash.
5574 This issue was reported to OpenSSL by Robert Święcki of Google.
5579 * BN_mod_exp may produce incorrect results on x86_64
5581 There is a carry propagating bug in the x86_64 Montgomery squaring
5582 procedure. No EC algorithms are affected. Analysis suggests that attacks
5583 against RSA and DSA as a result of this defect would be very difficult to
5584 perform and are not believed likely. Attacks against DH are considered just
5585 feasible (although very difficult) because most of the work necessary to
5586 deduce information about a private key may be performed offline. The amount
5587 of resources required for such an attack would be very significant and
5588 likely only accessible to a limited number of attackers. An attacker would
5589 additionally need online access to an unpatched system using the target
5590 private key in a scenario with persistent DH parameters and a private
5591 key that is shared between multiple clients. For example this can occur by
5592 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5593 similar to CVE-2015-3193 but must be treated as a separate problem.
5595 This issue was reported to OpenSSL by the OSS-Fuzz project.
5600 * Montgomery multiplication may produce incorrect results
5602 There is a carry propagating bug in the Broadwell-specific Montgomery
5603 multiplication procedure that handles input lengths divisible by, but
5604 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5605 and DH private keys are impossible. This is because the subroutine in
5606 question is not used in operations with the private key itself and an input
5607 of the attacker's direct choice. Otherwise the bug can manifest itself as
5608 transient authentication and key negotiation failures or reproducible
5609 erroneous outcome of public-key operations with specially crafted input.
5610 Among EC algorithms only Brainpool P-512 curves are affected and one
5611 presumably can attack ECDH key negotiation. Impact was not analyzed in
5612 detail, because pre-requisites for attack are considered unlikely. Namely
5613 multiple clients have to choose the curve in question and the server has to
5614 share the private key among them, neither of which is default behaviour.
5615 Even then only clients that chose the curve will be affected.
5617 This issue was publicly reported as transient failures and was not
5618 initially recognized as a security issue. Thanks to Richard Morgan for
5619 providing reproducible case.
5624 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5625 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5626 prevent issues where no progress is being made and the peer continually
5627 sends unrecognised record types, using up resources processing them.
5631 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5633 * Missing CRL sanity check
5635 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5636 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5637 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5639 This issue only affects the OpenSSL 1.0.2i
5644 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5646 * OCSP Status Request extension unbounded memory growth
5648 A malicious client can send an excessively large OCSP Status Request
5649 extension. If that client continually requests renegotiation, sending a
5650 large OCSP Status Request extension each time, then there will be unbounded
5651 memory growth on the server. This will eventually lead to a Denial Of
5652 Service attack through memory exhaustion. Servers with a default
5653 configuration are vulnerable even if they do not support OCSP. Builds using
5654 the "no-ocsp" build time option are not affected.
5656 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5661 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5664 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5670 * OOB write in MDC2_Update()
5672 An overflow can occur in MDC2_Update() either if called directly or
5673 through the EVP_DigestUpdate() function using MDC2. If an attacker
5674 is able to supply very large amounts of input data after a previous
5675 call to EVP_EncryptUpdate() with a partial block then a length check
5676 can overflow resulting in a heap corruption.
5678 The amount of data needed is comparable to SIZE_MAX which is impractical
5681 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5686 * Malformed SHA512 ticket DoS
5688 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5689 DoS attack where a malformed ticket will result in an OOB read which will
5692 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5693 a custom server callback and ticket lookup mechanism.
5695 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5700 * OOB write in BN_bn2dec()
5702 The function BN_bn2dec() does not check the return value of BN_div_word().
5703 This can cause an OOB write if an application uses this function with an
5704 overly large BIGNUM. This could be a problem if an overly large certificate
5705 or CRL is printed out from an untrusted source. TLS is not affected because
5706 record limits will reject an oversized certificate before it is parsed.
5708 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5713 * OOB read in TS_OBJ_print_bio()
5715 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5716 the total length the OID text representation would use and not the amount
5717 of data written. This will result in OOB reads when large OIDs are
5720 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5725 * Pointer arithmetic undefined behaviour
5727 Avoid some undefined pointer arithmetic
5729 A common idiom in the codebase is to check limits in the following manner:
5732 Where "p" points to some malloc'd data of SIZE bytes and
5735 "len" here could be from some externally supplied data (e.g. from a TLS
5738 The rules of C pointer arithmetic are such that "p + len" is only well
5739 defined where len <= SIZE. Therefore the above idiom is actually
5740 undefined behaviour.
5742 For example this could cause problems if some malloc implementation
5743 provides an address for "p" such that "p + len" actually overflows for
5744 values of len that are too big and therefore p + len < limit.
5746 This issue was reported to OpenSSL by Guido Vranken
5751 * Constant time flag not preserved in DSA signing
5753 Operations in the DSA signing algorithm should run in constant time in
5754 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5755 implementation means that a non-constant time codepath is followed for
5756 certain operations. This has been demonstrated through a cache-timing
5757 attack to be sufficient for an attacker to recover the private DSA key.
5759 This issue was reported by César Pereida (Aalto University), Billy Brumley
5760 (Tampere University of Technology), and Yuval Yarom (The University of
5761 Adelaide and NICTA).
5766 * DTLS buffered message DoS
5768 In a DTLS connection where handshake messages are delivered out-of-order
5769 those messages that OpenSSL is not yet ready to process will be buffered
5770 for later use. Under certain circumstances, a flaw in the logic means that
5771 those messages do not get removed from the buffer even though the handshake
5772 has been completed. An attacker could force up to approx. 15 messages to
5773 remain in the buffer when they are no longer required. These messages will
5774 be cleared when the DTLS connection is closed. The default maximum size for
5775 a message is 100k. Therefore, the attacker could force an additional 1500k
5776 to be consumed per connection. By opening many simultaneous connections an
5777 attacker could cause a DoS attack through memory exhaustion.
5779 This issue was reported to OpenSSL by Quan Luo.
5784 * DTLS replay protection DoS
5786 A flaw in the DTLS replay attack protection mechanism means that records
5787 that arrive for future epochs update the replay protection "window" before
5788 the MAC for the record has been validated. This could be exploited by an
5789 attacker by sending a record for the next epoch (which does not have to
5790 decrypt or have a valid MAC), with a very large sequence number. This means
5791 that all subsequent legitimate packets are dropped causing a denial of
5792 service for a specific DTLS connection.
5794 This issue was reported to OpenSSL by the OCAP audit team.
5799 * Certificate message OOB reads
5801 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5802 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5803 theoretical DoS risk but this has not been observed in practice on common
5806 The messages affected are client certificate, client certificate request
5807 and server certificate. As a result the attack can only be performed
5808 against a client or a server which enables client authentication.
5810 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5815 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5817 * Prevent padding oracle in AES-NI CBC MAC check
5819 A MITM attacker can use a padding oracle attack to decrypt traffic
5820 when the connection uses an AES CBC cipher and the server support
5823 This issue was introduced as part of the fix for Lucky 13 padding
5824 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5825 constant time by making sure that always the same bytes are read and
5826 compared against either the MAC or padding bytes. But it no longer
5827 checked that there was enough data to have both the MAC and padding
5830 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5834 * Fix EVP_EncodeUpdate overflow
5836 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5837 Base64 encoding of binary data. If an attacker is able to supply very large
5838 amounts of input data then a length check can overflow resulting in a heap
5841 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5842 the `PEM_write_bio*` family of functions. These are mainly used within the
5843 OpenSSL command line applications, so any application which processes data
5844 from an untrusted source and outputs it as a PEM file should be considered
5845 vulnerable to this issue. User applications that call these APIs directly
5846 with large amounts of untrusted data may also be vulnerable.
5848 This issue was reported by Guido Vranken.
5853 * Fix EVP_EncryptUpdate overflow
5855 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5856 is able to supply very large amounts of input data after a previous call to
5857 EVP_EncryptUpdate() with a partial block then a length check can overflow
5858 resulting in a heap corruption. Following an analysis of all OpenSSL
5859 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5860 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5861 the first called function after an EVP_EncryptInit(), and therefore that
5862 specific call must be safe. The second form is where the length passed to
5863 EVP_EncryptUpdate() can be seen from the code to be some small value and
5864 therefore there is no possibility of an overflow. Since all instances are
5865 one of these two forms, it is believed that there can be no overflows in
5866 internal code due to this problem. It should be noted that
5867 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5868 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5869 of these calls have also been analysed too and it is believed there are no
5870 instances in internal usage where an overflow could occur.
5872 This issue was reported by Guido Vranken.
5877 * Prevent ASN.1 BIO excessive memory allocation
5879 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5880 a short invalid encoding can cause allocation of large amounts of memory
5881 potentially consuming excessive resources or exhausting memory.
5883 Any application parsing untrusted data through d2i BIO functions is
5884 affected. The memory based functions such as d2i_X509() are *not* affected.
5885 Since the memory based functions are used by the TLS library, TLS
5886 applications are not affected.
5888 This issue was reported by Brian Carpenter.
5895 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5896 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5897 in arbitrary stack data being returned in the buffer.
5899 This issue was reported by Guido Vranken.
5904 * Modify behavior of ALPN to invoke callback after SNI/servername
5905 callback, such that updates to the SSL_CTX affect ALPN.
5909 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5914 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5915 methods are enabled and ssl2 is disabled the methods return NULL.
5919 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5921 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5922 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5923 provide any "EXPORT" or "LOW" strength ciphers.
5927 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5928 is by default disabled at build-time. Builds that are not configured with
5929 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5930 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5931 will need to explicitly call either of:
5933 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5935 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5937 as appropriate. Even if either of those is used, or the application
5938 explicitly uses the version-specific SSLv2_method() or its client and
5939 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5940 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5941 ciphers, and SSLv2 56-bit DES are no longer available.
5946 * Fix a double-free in DSA code
5948 A double free bug was discovered when OpenSSL parses malformed DSA private
5949 keys and could lead to a DoS attack or memory corruption for applications
5950 that receive DSA private keys from untrusted sources. This scenario is
5953 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5959 * Disable SRP fake user seed to address a server memory leak.
5961 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5963 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5964 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5965 was changed to ignore the "fake user" SRP seed, even if the seed
5968 Users should use SRP_VBASE_get1_by_user instead. Note that in
5969 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5970 also that even though configuring the SRP seed attempts to hide
5971 invalid usernames by continuing the handshake with fake
5972 credentials, this behaviour is not constant time and no strong
5973 guarantees are made that the handshake is indistinguishable from
5974 that of a valid user.
5979 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5981 In the BN_hex2bn function the number of hex digits is calculated using an
5982 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5983 large values of `i` this can result in `bn_expand` not allocating any
5984 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5985 field as NULL leading to a subsequent NULL ptr deref. For very large values
5986 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5987 In this case memory is allocated to the internal BIGNUM data field, but it
5988 is insufficiently sized leading to heap corruption. A similar issue exists
5989 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5990 is ever called by user applications with very large untrusted hex/dec data.
5991 This is anticipated to be a rare occurrence.
5993 All OpenSSL internal usage of these functions use data that is not expected
5994 to be untrusted, e.g. config file data or application command line
5995 arguments. If user developed applications generate config file data based
5996 on untrusted data then it is possible that this could also lead to security
5997 consequences. This is also anticipated to be rare.
5999 This issue was reported to OpenSSL by Guido Vranken.
6004 * Fix memory issues in `BIO_*printf` functions
6006 The internal `fmtstr` function used in processing a "%s" format string in
6007 the `BIO_*printf` functions could overflow while calculating the length of a
6008 string and cause an OOB read when printing very long strings.
6010 Additionally the internal `doapr_outch` function can attempt to write to an
6011 OOB memory location (at an offset from the NULL pointer) in the event of a
6012 memory allocation failure. In 1.0.2 and below this could be caused where
6013 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6014 could be in processing a very long "%s" format string. Memory leaks can
6017 The first issue may mask the second issue dependent on compiler behaviour.
6018 These problems could enable attacks where large amounts of untrusted data
6019 is passed to the `BIO_*printf` functions. If applications use these functions
6020 in this way then they could be vulnerable. OpenSSL itself uses these
6021 functions when printing out human-readable dumps of ASN.1 data. Therefore
6022 applications that print this data could be vulnerable if the data is from
6023 untrusted sources. OpenSSL command line applications could also be
6024 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6025 as command line arguments.
6027 Libssl is not considered directly vulnerable. Additionally certificates etc
6028 received via remote connections via libssl are also unlikely to be able to
6029 trigger these issues because of message size limits enforced within libssl.
6031 This issue was reported to OpenSSL Guido Vranken.
6036 * Side channel attack on modular exponentiation
6038 A side-channel attack was found which makes use of cache-bank conflicts on
6039 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6040 of RSA keys. The ability to exploit this issue is limited as it relies on
6041 an attacker who has control of code in a thread running on the same
6042 hyper-threaded core as the victim thread which is performing decryptions.
6044 This issue was reported to OpenSSL by Yuval Yarom, The University of
6045 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6046 Nadia Heninger, University of Pennsylvania with more information at
6047 <http://cachebleed.info>.
6052 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6053 if no keysize is specified with default_bits. This fixes an
6054 omission in an earlier change that changed all RSA/DSA key generation
6055 commands to use 2048 bits by default.
6059 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6061 * DH small subgroups
6063 Historically OpenSSL only ever generated DH parameters based on "safe"
6064 primes. More recently (in version 1.0.2) support was provided for
6065 generating X9.42 style parameter files such as those required for RFC 5114
6066 support. The primes used in such files may not be "safe". Where an
6067 application is using DH configured with parameters based on primes that are
6068 not "safe" then an attacker could use this fact to find a peer's private
6069 DH exponent. This attack requires that the attacker complete multiple
6070 handshakes in which the peer uses the same private DH exponent. For example
6071 this could be used to discover a TLS server's private DH exponent if it's
6072 reusing the private DH exponent or it's using a static DH ciphersuite.
6074 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6075 TLS. It is not on by default. If the option is not set then the server
6076 reuses the same private DH exponent for the life of the server process and
6077 would be vulnerable to this attack. It is believed that many popular
6078 applications do set this option and would therefore not be at risk.
6080 The fix for this issue adds an additional check where a "q" parameter is
6081 available (as is the case in X9.42 based parameters). This detects the
6082 only known attack, and is the only possible defense for static DH
6083 ciphersuites. This could have some performance impact.
6085 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6086 default and cannot be disabled. This could have some performance impact.
6088 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6093 * SSLv2 doesn't block disabled ciphers
6095 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6096 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6097 been disabled, provided that the SSLv2 protocol was not also disabled via
6100 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6101 and Sebastian Schinzel.
6106 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6108 * BN_mod_exp may produce incorrect results on x86_64
6110 There is a carry propagating bug in the x86_64 Montgomery squaring
6111 procedure. No EC algorithms are affected. Analysis suggests that attacks
6112 against RSA and DSA as a result of this defect would be very difficult to
6113 perform and are not believed likely. Attacks against DH are considered just
6114 feasible (although very difficult) because most of the work necessary to
6115 deduce information about a private key may be performed offline. The amount
6116 of resources required for such an attack would be very significant and
6117 likely only accessible to a limited number of attackers. An attacker would
6118 additionally need online access to an unpatched system using the target
6119 private key in a scenario with persistent DH parameters and a private
6120 key that is shared between multiple clients. For example this can occur by
6121 default in OpenSSL DHE based SSL/TLS ciphersuites.
6123 This issue was reported to OpenSSL by Hanno Böck.
6128 * Certificate verify crash with missing PSS parameter
6130 The signature verification routines will crash with a NULL pointer
6131 dereference if presented with an ASN.1 signature using the RSA PSS
6132 algorithm and absent mask generation function parameter. Since these
6133 routines are used to verify certificate signature algorithms this can be
6134 used to crash any certificate verification operation and exploited in a
6135 DoS attack. Any application which performs certificate verification is
6136 vulnerable including OpenSSL clients and servers which enable client
6139 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6144 * X509_ATTRIBUTE memory leak
6146 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6147 memory. This structure is used by the PKCS#7 and CMS routines so any
6148 application which reads PKCS#7 or CMS data from untrusted sources is
6149 affected. SSL/TLS is not affected.
6151 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6157 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6158 This changes the decoding behaviour for some invalid messages,
6159 though the change is mostly in the more lenient direction, and
6160 legacy behaviour is preserved as much as possible.
6164 * In DSA_generate_parameters_ex, if the provided seed is too short,
6167 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6169 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6171 * Alternate chains certificate forgery
6173 During certificate verification, OpenSSL will attempt to find an
6174 alternative certificate chain if the first attempt to build such a chain
6175 fails. An error in the implementation of this logic can mean that an
6176 attacker could cause certain checks on untrusted certificates to be
6177 bypassed, such as the CA flag, enabling them to use a valid leaf
6178 certificate to act as a CA and "issue" an invalid certificate.
6180 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6185 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6187 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6188 incompatibility in the handling of HMAC. The previous ABI has now been
6193 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6195 * Malformed ECParameters causes infinite loop
6197 When processing an ECParameters structure OpenSSL enters an infinite loop
6198 if the curve specified is over a specially malformed binary polynomial
6201 This can be used to perform denial of service against any
6202 system which processes public keys, certificate requests or
6203 certificates. This includes TLS clients and TLS servers with
6204 client authentication enabled.
6206 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6211 * Exploitable out-of-bounds read in X509_cmp_time
6213 X509_cmp_time does not properly check the length of the ASN1_TIME
6214 string and can read a few bytes out of bounds. In addition,
6215 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6218 An attacker can use this to craft malformed certificates and CRLs of
6219 various sizes and potentially cause a segmentation fault, resulting in
6220 a DoS on applications that verify certificates or CRLs. TLS clients
6221 that verify CRLs are affected. TLS clients and servers with client
6222 authentication enabled may be affected if they use custom verification
6225 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6226 independently by Hanno Böck.
6231 * PKCS7 crash with missing EnvelopedContent
6233 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6234 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6235 with missing content and trigger a NULL pointer dereference on parsing.
6237 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6238 structures from untrusted sources are affected. OpenSSL clients and
6239 servers are not affected.
6241 This issue was reported to OpenSSL by Michal Zalewski (Google).
6246 * CMS verify infinite loop with unknown hash function
6248 When verifying a signedData message the CMS code can enter an infinite loop
6249 if presented with an unknown hash function OID. This can be used to perform
6250 denial of service against any system which verifies signedData messages using
6252 This issue was reported to OpenSSL by Johannes Bauer.
6257 * Race condition handling NewSessionTicket
6259 If a NewSessionTicket is received by a multi-threaded client when attempting to
6260 reuse a previous ticket then a race condition can occur potentially leading to
6261 a double free of the ticket data.
6266 * Only support 256-bit or stronger elliptic curves with the
6267 'ecdh_auto' setting (server) or by default (client). Of supported
6268 curves, prefer P-256 (both).
6272 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6274 * ClientHello sigalgs DoS fix
6276 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6277 invalid signature algorithms extension a NULL pointer dereference will
6278 occur. This can be exploited in a DoS attack against the server.
6280 This issue was was reported to OpenSSL by David Ramos of Stanford
6284 *Stephen Henson and Matt Caswell*
6286 * Multiblock corrupted pointer fix
6288 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6289 feature only applies on 64 bit x86 architecture platforms that support AES
6290 NI instructions. A defect in the implementation of "multiblock" can cause
6291 OpenSSL's internal write buffer to become incorrectly set to NULL when
6292 using non-blocking IO. Typically, when the user application is using a
6293 socket BIO for writing, this will only result in a failed connection.
6294 However if some other BIO is used then it is likely that a segmentation
6295 fault will be triggered, thus enabling a potential DoS attack.
6297 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6302 * Segmentation fault in DTLSv1_listen fix
6304 The DTLSv1_listen function is intended to be stateless and processes the
6305 initial ClientHello from many peers. It is common for user code to loop
6306 over the call to DTLSv1_listen until a valid ClientHello is received with
6307 an associated cookie. A defect in the implementation of DTLSv1_listen means
6308 that state is preserved in the SSL object from one invocation to the next
6309 that can lead to a segmentation fault. Errors processing the initial
6310 ClientHello can trigger this scenario. An example of such an error could be
6311 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6314 This issue was reported to OpenSSL by Per Allansson.
6319 * Segmentation fault in ASN1_TYPE_cmp fix
6321 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6322 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6323 certificate signature algorithm consistency this can be used to crash any
6324 certificate verification operation and exploited in a DoS attack. Any
6325 application which performs certificate verification is vulnerable including
6326 OpenSSL clients and servers which enable client authentication.
6331 * Segmentation fault for invalid PSS parameters fix
6333 The signature verification routines will crash with a NULL pointer
6334 dereference if presented with an ASN.1 signature using the RSA PSS
6335 algorithm and invalid parameters. Since these routines are used to verify
6336 certificate signature algorithms this can be used to crash any
6337 certificate verification operation and exploited in a DoS attack. Any
6338 application which performs certificate verification is vulnerable including
6339 OpenSSL clients and servers which enable client authentication.
6341 This issue was was reported to OpenSSL by Brian Carpenter.
6346 * ASN.1 structure reuse memory corruption fix
6348 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6349 memory corruption via an invalid write. Such reuse is and has been
6350 strongly discouraged and is believed to be rare.
6352 Applications that parse structures containing CHOICE or ANY DEFINED BY
6353 components may be affected. Certificate parsing (d2i_X509 and related
6354 functions) are however not affected. OpenSSL clients and servers are
6360 * PKCS7 NULL pointer dereferences fix
6362 The PKCS#7 parsing code does not handle missing outer ContentInfo
6363 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6364 missing content and trigger a NULL pointer dereference on parsing.
6366 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6367 otherwise parse PKCS#7 structures from untrusted sources are
6368 affected. OpenSSL clients and servers are not affected.
6370 This issue was reported to OpenSSL by Michal Zalewski (Google).
6375 * DoS via reachable assert in SSLv2 servers fix
6377 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6378 servers that both support SSLv2 and enable export cipher suites by sending
6379 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6381 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6382 (OpenSSL development team).
6387 * Empty CKE with client auth and DHE fix
6389 If client auth is used then a server can seg fault in the event of a DHE
6390 ciphersuite being selected and a zero length ClientKeyExchange message
6391 being sent by the client. This could be exploited in a DoS attack.
6396 * Handshake with unseeded PRNG fix
6398 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6399 with an unseeded PRNG. The conditions are:
6400 - The client is on a platform where the PRNG has not been seeded
6401 automatically, and the user has not seeded manually
6402 - A protocol specific client method version has been used (i.e. not
6403 SSL_client_methodv23)
6404 - A ciphersuite is used that does not require additional random data from
6405 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6407 If the handshake succeeds then the client random that has been used will
6408 have been generated from a PRNG with insufficient entropy and therefore the
6409 output may be predictable.
6411 For example using the following command with an unseeded openssl will
6412 succeed on an unpatched platform:
6414 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6419 * Use After Free following d2i_ECPrivatekey error fix
6421 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6422 could cause a use after free condition. This, in turn, could cause a double
6423 free in several private key parsing functions (such as d2i_PrivateKey
6424 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6425 for applications that receive EC private keys from untrusted
6426 sources. This scenario is considered rare.
6428 This issue was discovered by the BoringSSL project and fixed in their
6434 * X509_to_X509_REQ NULL pointer deref fix
6436 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6437 the certificate key is invalid. This function is rarely used in practice.
6439 This issue was discovered by Brian Carpenter.
6444 * Removed the export ciphers from the DEFAULT ciphers
6448 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6450 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6451 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6452 So far those who have to target multiple platforms would compromise
6453 and argue that binary targeting say ARMv5 would still execute on
6454 ARMv8. "Universal" build resolves this compromise by providing
6455 near-optimal performance even on newer platforms.
6459 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6460 (other platforms pending).
6462 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6464 * Add support for the SignedCertificateTimestampList certificate and
6465 OCSP response extensions from RFC6962.
6469 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6470 for corner cases. (Certain input points at infinity could lead to
6471 bogus results, with non-infinity inputs mapped to infinity too.)
6475 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6476 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6477 common cases are optimized and there still is room for further
6478 improvements. Vector Permutation AES for Altivec is also added.
6482 * Add support for little-endian ppc64 Linux target.
6484 *Marcelo Cerri (IBM)*
6486 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6487 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6488 are optimized and there still is room for further improvements.
6489 Both 32- and 64-bit modes are supported.
6491 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6493 * Improved ARMv7 NEON support.
6497 * Support for SPARC Architecture 2011 crypto extensions, first
6498 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6499 SHA256/512, MD5, GHASH and modular exponentiation.
6501 *Andy Polyakov, David Miller*
6503 * Accelerated modular exponentiation for Intel processors, a.k.a.
6506 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6508 * Support for new and upcoming Intel processors, including AVX2,
6509 BMI and SHA ISA extensions. This includes additional "stitched"
6510 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6513 This work was sponsored by Intel Corp.
6517 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6518 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6519 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6523 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6524 this fixes a limitation in previous versions of OpenSSL.
6528 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6529 MGF1 digest and OAEP label.
6533 * Add EVP support for key wrapping algorithms, to avoid problems with
6534 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6535 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6536 algorithms and include tests cases.
6540 * Add functions to allocate and set the fields of an ECDSA_METHOD
6543 *Douglas E. Engert, Steve Henson*
6545 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6546 difference in days and seconds between two tm or ASN1_TIME structures.
6550 * Add -rev test option to s_server to just reverse order of characters
6551 received by client and send back to server. Also prints an abbreviated
6552 summary of the connection parameters.
6556 * New option -brief for s_client and s_server to print out a brief summary
6557 of connection parameters.
6561 * Add callbacks for arbitrary TLS extensions.
6563 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6565 * New option -crl_download in several openssl utilities to download CRLs
6566 from CRLDP extension in certificates.
6570 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6574 * New function X509_CRL_diff to generate a delta CRL from the difference
6575 of two full CRLs. Add support to "crl" utility.
6579 * New functions to set lookup_crls function and to retrieve
6580 X509_STORE from X509_STORE_CTX.
6584 * Print out deprecated issuer and subject unique ID fields in
6589 * Extend OCSP I/O functions so they can be used for simple general purpose
6590 HTTP as well as OCSP. New wrapper function which can be used to download
6591 CRLs using the OCSP API.
6595 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6599 * `SSL_CONF*` functions. These provide a common framework for application
6600 configuration using configuration files or command lines.
6604 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6605 message callback and prints the results. Needs compile time option
6606 "enable-ssl-trace". New options to s_client and s_server to enable
6611 * New ctrl and macro to retrieve supported points extensions.
6612 Print out extension in s_server and s_client.
6616 * New functions to retrieve certificate signature and signature
6621 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6626 * New Suite B modes for TLS code. These use and enforce the requirements
6627 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6628 only use Suite B curves. The Suite B modes can be set by using the
6629 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6633 * New chain verification flags for Suite B levels of security. Check
6634 algorithms are acceptable when flags are set in X509_verify_cert.
6638 * Make tls1_check_chain return a set of flags indicating checks passed
6639 by a certificate chain. Add additional tests to handle client
6640 certificates: checks for matching certificate type and issuer name
6645 * If an attempt is made to use a signature algorithm not in the peer
6646 preference list abort the handshake. If client has no suitable
6647 signature algorithms in response to a certificate request do not
6648 use the certificate.
6652 * If server EC tmp key is not in client preference list abort handshake.
6656 * Add support for certificate stores in CERT structure. This makes it
6657 possible to have different stores per SSL structure or one store in
6658 the parent SSL_CTX. Include distinct stores for certificate chain
6659 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6660 to build and store a certificate chain in CERT structure: returning
6661 an error if the chain cannot be built: this will allow applications
6662 to test if a chain is correctly configured.
6664 Note: if the CERT based stores are not set then the parent SSL_CTX
6665 store is used to retain compatibility with existing behaviour.
6669 * New function ssl_set_client_disabled to set a ciphersuite disabled
6670 mask based on the current session, check mask when sending client
6671 hello and checking the requested ciphersuite.
6675 * New ctrls to retrieve and set certificate types in a certificate
6676 request message. Print out received values in s_client. If certificate
6677 types is not set with custom values set sensible values based on
6678 supported signature algorithms.
6682 * Support for distinct client and server supported signature algorithms.
6686 * Add certificate callback. If set this is called whenever a certificate
6687 is required by client or server. An application can decide which
6688 certificate chain to present based on arbitrary criteria: for example
6689 supported signature algorithms. Add very simple example to s_server.
6690 This fixes many of the problems and restrictions of the existing client
6691 certificate callback: for example you can now clear an existing
6692 certificate and specify the whole chain.
6696 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6697 the certificate can be used for (if anything). Set valid_flags field
6698 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6699 to have similar checks in it.
6701 Add new "cert_flags" field to CERT structure and include a "strict mode".
6702 This enforces some TLS certificate requirements (such as only permitting
6703 certificate signature algorithms contained in the supported algorithms
6704 extension) which some implementations ignore: this option should be used
6705 with caution as it could cause interoperability issues.
6709 * Update and tidy signature algorithm extension processing. Work out
6710 shared signature algorithms based on preferences and peer algorithms
6711 and print them out in s_client and s_server. Abort handshake if no
6712 shared signature algorithms.
6716 * Add new functions to allow customised supported signature algorithms
6717 for SSL and SSL_CTX structures. Add options to s_client and s_server
6722 * New function SSL_certs_clear() to delete all references to certificates
6723 from an SSL structure. Before this once a certificate had been added
6724 it couldn't be removed.
6728 * Integrate hostname, email address and IP address checking with certificate
6729 verification. New verify options supporting checking in openssl utility.
6733 * Fixes and wildcard matching support to hostname and email checking
6734 functions. Add manual page.
6736 *Florian Weimer (Red Hat Product Security Team)*
6738 * New functions to check a hostname email or IP address against a
6739 certificate. Add options x509 utility to print results of checks against
6744 * Fix OCSP checking.
6746 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6748 * Initial experimental support for explicitly trusted non-root CAs.
6749 OpenSSL still tries to build a complete chain to a root but if an
6750 intermediate CA has a trust setting included that is used. The first
6751 setting is used: whether to trust (e.g., -addtrust option to the x509
6756 * Add -trusted_first option which attempts to find certificates in the
6757 trusted store even if an untrusted chain is also supplied.
6761 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6762 platform support for Linux and Android.
6766 * Support for linux-x32, ILP32 environment in x86_64 framework.
6770 * Experimental multi-implementation support for FIPS capable OpenSSL.
6771 When in FIPS mode the approved implementations are used as normal,
6772 when not in FIPS mode the internal unapproved versions are used instead.
6773 This means that the FIPS capable OpenSSL isn't forced to use the
6774 (often lower performance) FIPS implementations outside FIPS mode.
6778 * Transparently support X9.42 DH parameters when calling
6779 PEM_read_bio_DHparameters. This means existing applications can handle
6780 the new parameter format automatically.
6784 * Initial experimental support for X9.42 DH parameter format: mainly
6785 to support use of 'q' parameter for RFC5114 parameters.
6789 * Add DH parameters from RFC5114 including test data to dhtest.
6793 * Support for automatic EC temporary key parameter selection. If enabled
6794 the most preferred EC parameters are automatically used instead of
6795 hardcoded fixed parameters. Now a server just has to call:
6796 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6797 support ECDH and use the most appropriate parameters.
6801 * Enhance and tidy EC curve and point format TLS extension code. Use
6802 static structures instead of allocation if default values are used.
6803 New ctrls to set curves we wish to support and to retrieve shared curves.
6804 Print out shared curves in s_server. New options to s_server and s_client
6805 to set list of supported curves.
6809 * New ctrls to retrieve supported signature algorithms and
6810 supported curve values as an array of NIDs. Extend openssl utility
6811 to print out received values.
6815 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6816 between NIDs and the more common NIST names such as "P-256". Enhance
6817 ecparam utility and ECC method to recognise the NIST names for curves.
6821 * Enhance SSL/TLS certificate chain handling to support different
6822 chains for each certificate instead of one chain in the parent SSL_CTX.
6826 * Support for fixed DH ciphersuite client authentication: where both
6827 server and client use DH certificates with common parameters.
6831 * Support for fixed DH ciphersuites: those requiring DH server
6836 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6838 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6839 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6840 X509_CINF_get_signature were reverted post internal team review.
6845 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6847 * OCSP Status Request extension unbounded memory growth
6849 A malicious client can send an excessively large OCSP Status Request
6850 extension. If that client continually requests renegotiation, sending a
6851 large OCSP Status Request extension each time, then there will be unbounded
6852 memory growth on the server. This will eventually lead to a Denial Of
6853 Service attack through memory exhaustion. Servers with a default
6854 configuration are vulnerable even if they do not support OCSP. Builds using
6855 the "no-ocsp" build time option are not affected.
6857 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6862 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6865 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6871 * OOB write in MDC2_Update()
6873 An overflow can occur in MDC2_Update() either if called directly or
6874 through the EVP_DigestUpdate() function using MDC2. If an attacker
6875 is able to supply very large amounts of input data after a previous
6876 call to EVP_EncryptUpdate() with a partial block then a length check
6877 can overflow resulting in a heap corruption.
6879 The amount of data needed is comparable to SIZE_MAX which is impractical
6882 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6887 * Malformed SHA512 ticket DoS
6889 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6890 DoS attack where a malformed ticket will result in an OOB read which will
6893 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6894 a custom server callback and ticket lookup mechanism.
6896 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6901 * OOB write in BN_bn2dec()
6903 The function BN_bn2dec() does not check the return value of BN_div_word().
6904 This can cause an OOB write if an application uses this function with an
6905 overly large BIGNUM. This could be a problem if an overly large certificate
6906 or CRL is printed out from an untrusted source. TLS is not affected because
6907 record limits will reject an oversized certificate before it is parsed.
6909 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6914 * OOB read in TS_OBJ_print_bio()
6916 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6917 the total length the OID text representation would use and not the amount
6918 of data written. This will result in OOB reads when large OIDs are
6921 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6926 * Pointer arithmetic undefined behaviour
6928 Avoid some undefined pointer arithmetic
6930 A common idiom in the codebase is to check limits in the following manner:
6933 Where "p" points to some malloc'd data of SIZE bytes and
6936 "len" here could be from some externally supplied data (e.g. from a TLS
6939 The rules of C pointer arithmetic are such that "p + len" is only well
6940 defined where len <= SIZE. Therefore, the above idiom is actually
6941 undefined behaviour.
6943 For example this could cause problems if some malloc implementation
6944 provides an address for "p" such that "p + len" actually overflows for
6945 values of len that are too big and therefore p + len < limit.
6947 This issue was reported to OpenSSL by Guido Vranken
6952 * Constant time flag not preserved in DSA signing
6954 Operations in the DSA signing algorithm should run in constant time in
6955 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6956 implementation means that a non-constant time codepath is followed for
6957 certain operations. This has been demonstrated through a cache-timing
6958 attack to be sufficient for an attacker to recover the private DSA key.
6960 This issue was reported by César Pereida (Aalto University), Billy Brumley
6961 (Tampere University of Technology), and Yuval Yarom (The University of
6962 Adelaide and NICTA).
6967 * DTLS buffered message DoS
6969 In a DTLS connection where handshake messages are delivered out-of-order
6970 those messages that OpenSSL is not yet ready to process will be buffered
6971 for later use. Under certain circumstances, a flaw in the logic means that
6972 those messages do not get removed from the buffer even though the handshake
6973 has been completed. An attacker could force up to approx. 15 messages to
6974 remain in the buffer when they are no longer required. These messages will
6975 be cleared when the DTLS connection is closed. The default maximum size for
6976 a message is 100k. Therefore, the attacker could force an additional 1500k
6977 to be consumed per connection. By opening many simultaneous connections an
6978 attacker could cause a DoS attack through memory exhaustion.
6980 This issue was reported to OpenSSL by Quan Luo.
6985 * DTLS replay protection DoS
6987 A flaw in the DTLS replay attack protection mechanism means that records
6988 that arrive for future epochs update the replay protection "window" before
6989 the MAC for the record has been validated. This could be exploited by an
6990 attacker by sending a record for the next epoch (which does not have to
6991 decrypt or have a valid MAC), with a very large sequence number. This means
6992 that all subsequent legitimate packets are dropped causing a denial of
6993 service for a specific DTLS connection.
6995 This issue was reported to OpenSSL by the OCAP audit team.
7000 * Certificate message OOB reads
7002 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7003 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7004 theoretical DoS risk but this has not been observed in practice on common
7007 The messages affected are client certificate, client certificate request
7008 and server certificate. As a result the attack can only be performed
7009 against a client or a server which enables client authentication.
7011 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7016 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7018 * Prevent padding oracle in AES-NI CBC MAC check
7020 A MITM attacker can use a padding oracle attack to decrypt traffic
7021 when the connection uses an AES CBC cipher and the server support
7024 This issue was introduced as part of the fix for Lucky 13 padding
7025 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7026 constant time by making sure that always the same bytes are read and
7027 compared against either the MAC or padding bytes. But it no longer
7028 checked that there was enough data to have both the MAC and padding
7031 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7036 * Fix EVP_EncodeUpdate overflow
7038 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7039 Base64 encoding of binary data. If an attacker is able to supply very large
7040 amounts of input data then a length check can overflow resulting in a heap
7043 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7044 the `PEM_write_bio*` family of functions. These are mainly used within the
7045 OpenSSL command line applications, so any application which processes data
7046 from an untrusted source and outputs it as a PEM file should be considered
7047 vulnerable to this issue. User applications that call these APIs directly
7048 with large amounts of untrusted data may also be vulnerable.
7050 This issue was reported by Guido Vranken.
7055 * Fix EVP_EncryptUpdate overflow
7057 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7058 is able to supply very large amounts of input data after a previous call to
7059 EVP_EncryptUpdate() with a partial block then a length check can overflow
7060 resulting in a heap corruption. Following an analysis of all OpenSSL
7061 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7062 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7063 the first called function after an EVP_EncryptInit(), and therefore that
7064 specific call must be safe. The second form is where the length passed to
7065 EVP_EncryptUpdate() can be seen from the code to be some small value and
7066 therefore there is no possibility of an overflow. Since all instances are
7067 one of these two forms, it is believed that there can be no overflows in
7068 internal code due to this problem. It should be noted that
7069 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7070 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7071 of these calls have also been analysed too and it is believed there are no
7072 instances in internal usage where an overflow could occur.
7074 This issue was reported by Guido Vranken.
7079 * Prevent ASN.1 BIO excessive memory allocation
7081 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7082 a short invalid encoding can casuse allocation of large amounts of memory
7083 potentially consuming excessive resources or exhausting memory.
7085 Any application parsing untrusted data through d2i BIO functions is
7086 affected. The memory based functions such as d2i_X509() are *not* affected.
7087 Since the memory based functions are used by the TLS library, TLS
7088 applications are not affected.
7090 This issue was reported by Brian Carpenter.
7097 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7098 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7099 in arbitrary stack data being returned in the buffer.
7101 This issue was reported by Guido Vranken.
7106 * Modify behavior of ALPN to invoke callback after SNI/servername
7107 callback, such that updates to the SSL_CTX affect ALPN.
7111 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7116 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7117 methods are enabled and ssl2 is disabled the methods return NULL.
7121 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7123 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7124 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7125 provide any "EXPORT" or "LOW" strength ciphers.
7129 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7130 is by default disabled at build-time. Builds that are not configured with
7131 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7132 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7133 will need to explicitly call either of:
7135 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7137 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7139 as appropriate. Even if either of those is used, or the application
7140 explicitly uses the version-specific SSLv2_method() or its client and
7141 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7142 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7143 ciphers, and SSLv2 56-bit DES are no longer available.
7148 * Fix a double-free in DSA code
7150 A double free bug was discovered when OpenSSL parses malformed DSA private
7151 keys and could lead to a DoS attack or memory corruption for applications
7152 that receive DSA private keys from untrusted sources. This scenario is
7155 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7161 * Disable SRP fake user seed to address a server memory leak.
7163 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7165 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7166 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7167 was changed to ignore the "fake user" SRP seed, even if the seed
7170 Users should use SRP_VBASE_get1_by_user instead. Note that in
7171 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7172 also that even though configuring the SRP seed attempts to hide
7173 invalid usernames by continuing the handshake with fake
7174 credentials, this behaviour is not constant time and no strong
7175 guarantees are made that the handshake is indistinguishable from
7176 that of a valid user.
7181 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7183 In the BN_hex2bn function the number of hex digits is calculated using an
7184 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7185 large values of `i` this can result in `bn_expand` not allocating any
7186 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7187 field as NULL leading to a subsequent NULL ptr deref. For very large values
7188 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7189 In this case memory is allocated to the internal BIGNUM data field, but it
7190 is insufficiently sized leading to heap corruption. A similar issue exists
7191 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7192 is ever called by user applications with very large untrusted hex/dec data.
7193 This is anticipated to be a rare occurrence.
7195 All OpenSSL internal usage of these functions use data that is not expected
7196 to be untrusted, e.g. config file data or application command line
7197 arguments. If user developed applications generate config file data based
7198 on untrusted data then it is possible that this could also lead to security
7199 consequences. This is also anticipated to be rare.
7201 This issue was reported to OpenSSL by Guido Vranken.
7206 * Fix memory issues in `BIO_*printf` functions
7208 The internal `fmtstr` function used in processing a "%s" format string in
7209 the `BIO_*printf` functions could overflow while calculating the length of a
7210 string and cause an OOB read when printing very long strings.
7212 Additionally the internal `doapr_outch` function can attempt to write to an
7213 OOB memory location (at an offset from the NULL pointer) in the event of a
7214 memory allocation failure. In 1.0.2 and below this could be caused where
7215 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7216 could be in processing a very long "%s" format string. Memory leaks can
7219 The first issue may mask the second issue dependent on compiler behaviour.
7220 These problems could enable attacks where large amounts of untrusted data
7221 is passed to the `BIO_*printf` functions. If applications use these functions
7222 in this way then they could be vulnerable. OpenSSL itself uses these
7223 functions when printing out human-readable dumps of ASN.1 data. Therefore
7224 applications that print this data could be vulnerable if the data is from
7225 untrusted sources. OpenSSL command line applications could also be
7226 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7227 as command line arguments.
7229 Libssl is not considered directly vulnerable. Additionally certificates etc
7230 received via remote connections via libssl are also unlikely to be able to
7231 trigger these issues because of message size limits enforced within libssl.
7233 This issue was reported to OpenSSL Guido Vranken.
7238 * Side channel attack on modular exponentiation
7240 A side-channel attack was found which makes use of cache-bank conflicts on
7241 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7242 of RSA keys. The ability to exploit this issue is limited as it relies on
7243 an attacker who has control of code in a thread running on the same
7244 hyper-threaded core as the victim thread which is performing decryptions.
7246 This issue was reported to OpenSSL by Yuval Yarom, The University of
7247 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7248 Nadia Heninger, University of Pennsylvania with more information at
7249 <http://cachebleed.info>.
7254 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7255 if no keysize is specified with default_bits. This fixes an
7256 omission in an earlier change that changed all RSA/DSA key generation
7257 commands to use 2048 bits by default.
7261 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7263 * Protection for DH small subgroup attacks
7265 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7266 switched on by default and cannot be disabled. This could have some
7271 * SSLv2 doesn't block disabled ciphers
7273 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7274 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7275 been disabled, provided that the SSLv2 protocol was not also disabled via
7278 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7279 and Sebastian Schinzel.
7284 * Reject DH handshakes with parameters shorter than 1024 bits.
7288 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7290 * Certificate verify crash with missing PSS parameter
7292 The signature verification routines will crash with a NULL pointer
7293 dereference if presented with an ASN.1 signature using the RSA PSS
7294 algorithm and absent mask generation function parameter. Since these
7295 routines are used to verify certificate signature algorithms this can be
7296 used to crash any certificate verification operation and exploited in a
7297 DoS attack. Any application which performs certificate verification is
7298 vulnerable including OpenSSL clients and servers which enable client
7301 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7306 * X509_ATTRIBUTE memory leak
7308 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7309 memory. This structure is used by the PKCS#7 and CMS routines so any
7310 application which reads PKCS#7 or CMS data from untrusted sources is
7311 affected. SSL/TLS is not affected.
7313 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7319 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7320 This changes the decoding behaviour for some invalid messages,
7321 though the change is mostly in the more lenient direction, and
7322 legacy behaviour is preserved as much as possible.
7326 * In DSA_generate_parameters_ex, if the provided seed is too short,
7327 use a random seed, as already documented.
7329 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7331 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7333 * Alternate chains certificate forgery
7335 During certificate verfification, OpenSSL will attempt to find an
7336 alternative certificate chain if the first attempt to build such a chain
7337 fails. An error in the implementation of this logic can mean that an
7338 attacker could cause certain checks on untrusted certificates to be
7339 bypassed, such as the CA flag, enabling them to use a valid leaf
7340 certificate to act as a CA and "issue" an invalid certificate.
7342 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7348 * Race condition handling PSK identify hint
7350 If PSK identity hints are received by a multi-threaded client then
7351 the values are wrongly updated in the parent SSL_CTX structure. This can
7352 result in a race condition potentially leading to a double free of the
7358 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7360 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7361 incompatibility in the handling of HMAC. The previous ABI has now been
7364 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7366 * Malformed ECParameters causes infinite loop
7368 When processing an ECParameters structure OpenSSL enters an infinite loop
7369 if the curve specified is over a specially malformed binary polynomial
7372 This can be used to perform denial of service against any
7373 system which processes public keys, certificate requests or
7374 certificates. This includes TLS clients and TLS servers with
7375 client authentication enabled.
7377 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7382 * Exploitable out-of-bounds read in X509_cmp_time
7384 X509_cmp_time does not properly check the length of the ASN1_TIME
7385 string and can read a few bytes out of bounds. In addition,
7386 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7389 An attacker can use this to craft malformed certificates and CRLs of
7390 various sizes and potentially cause a segmentation fault, resulting in
7391 a DoS on applications that verify certificates or CRLs. TLS clients
7392 that verify CRLs are affected. TLS clients and servers with client
7393 authentication enabled may be affected if they use custom verification
7396 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7397 independently by Hanno Böck.
7402 * PKCS7 crash with missing EnvelopedContent
7404 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7405 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7406 with missing content and trigger a NULL pointer dereference on parsing.
7408 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7409 structures from untrusted sources are affected. OpenSSL clients and
7410 servers are not affected.
7412 This issue was reported to OpenSSL by Michal Zalewski (Google).
7417 * CMS verify infinite loop with unknown hash function
7419 When verifying a signedData message the CMS code can enter an infinite loop
7420 if presented with an unknown hash function OID. This can be used to perform
7421 denial of service against any system which verifies signedData messages using
7423 This issue was reported to OpenSSL by Johannes Bauer.
7428 * Race condition handling NewSessionTicket
7430 If a NewSessionTicket is received by a multi-threaded client when attempting to
7431 reuse a previous ticket then a race condition can occur potentially leading to
7432 a double free of the ticket data.
7437 * Reject DH handshakes with parameters shorter than 768 bits.
7439 *Kurt Roeckx and Emilia Kasper*
7441 * dhparam: generate 2048-bit parameters by default.
7443 *Kurt Roeckx and Emilia Kasper*
7445 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7447 * Segmentation fault in ASN1_TYPE_cmp fix
7449 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7450 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7451 certificate signature algorithm consistency this can be used to crash any
7452 certificate verification operation and exploited in a DoS attack. Any
7453 application which performs certificate verification is vulnerable including
7454 OpenSSL clients and servers which enable client authentication.
7459 * ASN.1 structure reuse memory corruption fix
7461 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7462 memory corruption via an invalid write. Such reuse is and has been
7463 strongly discouraged and is believed to be rare.
7465 Applications that parse structures containing CHOICE or ANY DEFINED BY
7466 components may be affected. Certificate parsing (d2i_X509 and related
7467 functions) are however not affected. OpenSSL clients and servers are
7473 * PKCS7 NULL pointer dereferences fix
7475 The PKCS#7 parsing code does not handle missing outer ContentInfo
7476 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7477 missing content and trigger a NULL pointer dereference on parsing.
7479 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7480 otherwise parse PKCS#7 structures from untrusted sources are
7481 affected. OpenSSL clients and servers are not affected.
7483 This issue was reported to OpenSSL by Michal Zalewski (Google).
7488 * DoS via reachable assert in SSLv2 servers fix
7490 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7491 servers that both support SSLv2 and enable export cipher suites by sending
7492 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7494 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7495 (OpenSSL development team).
7500 * Use After Free following d2i_ECPrivatekey error fix
7502 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7503 could cause a use after free condition. This, in turn, could cause a double
7504 free in several private key parsing functions (such as d2i_PrivateKey
7505 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7506 for applications that receive EC private keys from untrusted
7507 sources. This scenario is considered rare.
7509 This issue was discovered by the BoringSSL project and fixed in their
7515 * X509_to_X509_REQ NULL pointer deref fix
7517 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7518 the certificate key is invalid. This function is rarely used in practice.
7520 This issue was discovered by Brian Carpenter.
7525 * Removed the export ciphers from the DEFAULT ciphers
7529 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7531 * Build fixes for the Windows and OpenVMS platforms
7533 *Matt Caswell and Richard Levitte*
7535 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7537 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7538 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7539 dereference. This could lead to a Denial Of Service attack. Thanks to
7540 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7545 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7546 dtls1_buffer_record function under certain conditions. In particular this
7547 could occur if an attacker sent repeated DTLS records with the same
7548 sequence number but for the next epoch. The memory leak could be exploited
7549 by an attacker in a Denial of Service attack through memory exhaustion.
7550 Thanks to Chris Mueller for reporting this issue.
7555 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7556 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7557 method would be set to NULL which could later result in a NULL pointer
7558 dereference. Thanks to Frank Schmirler for reporting this issue.
7563 * Abort handshake if server key exchange message is omitted for ephemeral
7566 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7567 reporting this issue.
7572 * Remove non-export ephemeral RSA code on client and server. This code
7573 violated the TLS standard by allowing the use of temporary RSA keys in
7574 non-export ciphersuites and could be used by a server to effectively
7575 downgrade the RSA key length used to a value smaller than the server
7576 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7577 INRIA or reporting this issue.
7582 * Fixed issue where DH client certificates are accepted without verification.
7583 An OpenSSL server will accept a DH certificate for client authentication
7584 without the certificate verify message. This effectively allows a client to
7585 authenticate without the use of a private key. This only affects servers
7586 which trust a client certificate authority which issues certificates
7587 containing DH keys: these are extremely rare and hardly ever encountered.
7588 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7594 * Ensure that the session ID context of an SSL is updated when its
7595 SSL_CTX is updated via SSL_set_SSL_CTX.
7597 The session ID context is typically set from the parent SSL_CTX,
7598 and can vary with the CTX.
7602 * Fix various certificate fingerprint issues.
7604 By using non-DER or invalid encodings outside the signed portion of a
7605 certificate the fingerprint can be changed without breaking the signature.
7606 Although no details of the signed portion of the certificate can be changed
7607 this can cause problems with some applications: e.g. those using the
7608 certificate fingerprint for blacklists.
7610 1. Reject signatures with non zero unused bits.
7612 If the BIT STRING containing the signature has non zero unused bits reject
7613 the signature. All current signature algorithms require zero unused bits.
7615 2. Check certificate algorithm consistency.
7617 Check the AlgorithmIdentifier inside TBS matches the one in the
7618 certificate signature. NB: this will result in signature failure
7619 errors for some broken certificates.
7621 Thanks to Konrad Kraszewski from Google for reporting this issue.
7623 3. Check DSA/ECDSA signatures use DER.
7625 Re-encode DSA/ECDSA signatures and compare with the original received
7626 signature. Return an error if there is a mismatch.
7628 This will reject various cases including garbage after signature
7629 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7630 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7631 (negative or with leading zeroes).
7633 Further analysis was conducted and fixes were developed by Stephen Henson
7634 of the OpenSSL core team.
7640 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7641 results on some platforms, including x86_64. This bug occurs at random
7642 with a very low probability, and is not known to be exploitable in any
7643 way, though its exact impact is difficult to determine. Thanks to Pieter
7644 Wuille (Blockstream) who reported this issue and also suggested an initial
7645 fix. Further analysis was conducted by the OpenSSL development team and
7646 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7647 the OpenSSL core team.
7652 * Do not resume sessions on the server if the negotiated protocol
7653 version does not match the session's version. Resuming with a different
7654 version, while not strictly forbidden by the RFC, is of questionable
7655 sanity and breaks all known clients.
7657 *David Benjamin, Emilia Käsper*
7659 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7660 early CCS messages during renegotiation. (Note that because
7661 renegotiation is encrypted, this early CCS was not exploitable.)
7665 * Tighten client-side session ticket handling during renegotiation:
7666 ensure that the client only accepts a session ticket if the server sends
7667 the extension anew in the ServerHello. Previously, a TLS client would
7668 reuse the old extension state and thus accept a session ticket if one was
7669 announced in the initial ServerHello.
7671 Similarly, ensure that the client requires a session ticket if one
7672 was advertised in the ServerHello. Previously, a TLS client would
7673 ignore a missing NewSessionTicket message.
7677 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7681 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7682 sends a carefully crafted handshake message, to cause OpenSSL to fail
7683 to free up to 64k of memory causing a memory leak. This could be
7684 exploited in a Denial Of Service attack. This issue affects OpenSSL
7685 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7686 whether SRTP is used or configured. Implementations of OpenSSL that
7687 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7689 The fix was developed by the OpenSSL team.
7694 * Session Ticket Memory Leak.
7696 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7697 integrity of that ticket is first verified. In the event of a session
7698 ticket integrity check failing, OpenSSL will fail to free memory
7699 causing a memory leak. By sending a large number of invalid session
7700 tickets an attacker could exploit this issue in a Denial Of Service
7706 * Build option no-ssl3 is incomplete.
7708 When OpenSSL is configured with "no-ssl3" as a build option, servers
7709 could accept and complete an SSL 3.0 handshake, and clients could be
7710 configured to send them.
7713 *Akamai and the OpenSSL team*
7715 * Add support for TLS_FALLBACK_SCSV.
7716 Client applications doing fallback retries should call
7717 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7720 *Adam Langley, Bodo Moeller*
7722 * Add additional DigestInfo checks.
7724 Re-encode DigestInto in DER and check against the original when
7725 verifying RSA signature: this will reject any improperly encoded
7726 DigestInfo structures.
7728 Note: this is a precautionary measure and no attacks are currently known.
7732 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7734 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7735 SRP code can be overrun an internal buffer. Add sanity check that
7736 g, A, B < N to SRP code.
7738 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7739 Group for discovering this issue.
7744 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7745 TLS 1.0 instead of higher protocol versions when the ClientHello message
7746 is badly fragmented. This allows a man-in-the-middle attacker to force a
7747 downgrade to TLS 1.0 even if both the server and the client support a
7748 higher protocol version, by modifying the client's TLS records.
7750 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7751 researching this issue.
7756 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7757 to a denial of service attack. A malicious server can crash the client
7758 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7759 ciphersuite and sending carefully crafted handshake messages.
7761 Thanks to Felix Gröbert (Google) for discovering and researching this
7767 * By sending carefully crafted DTLS packets an attacker could cause openssl
7768 to leak memory. This can be exploited through a Denial of Service attack.
7769 Thanks to Adam Langley for discovering and researching this issue.
7774 * An attacker can force openssl to consume large amounts of memory whilst
7775 processing DTLS handshake messages. This can be exploited through a
7776 Denial of Service attack.
7777 Thanks to Adam Langley for discovering and researching this issue.
7782 * An attacker can force an error condition which causes openssl to crash
7783 whilst processing DTLS packets due to memory being freed twice. This
7784 can be exploited through a Denial of Service attack.
7785 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7791 * If a multithreaded client connects to a malicious server using a resumed
7792 session and the server sends an ec point format extension it could write
7793 up to 255 bytes to freed memory.
7795 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7801 * A malicious server can crash an OpenSSL client with a null pointer
7802 dereference (read) by specifying an SRP ciphersuite even though it was not
7803 properly negotiated with the client. This can be exploited through a
7804 Denial of Service attack.
7806 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7807 discovering and researching this issue.
7812 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7813 X509_name_oneline, X509_name_print_ex et al. to leak some information
7814 from the stack. Applications may be affected if they echo pretty printing
7815 output to the attacker.
7817 Thanks to Ivan Fratric (Google) for discovering this issue.
7820 *Emilia Käsper, and Steve Henson*
7822 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7823 for corner cases. (Certain input points at infinity could lead to
7824 bogus results, with non-infinity inputs mapped to infinity too.)
7828 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7830 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7831 handshake can force the use of weak keying material in OpenSSL
7832 SSL/TLS clients and servers.
7834 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7835 researching this issue. ([CVE-2014-0224])
7837 *KIKUCHI Masashi, Steve Henson*
7839 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7840 OpenSSL DTLS client the code can be made to recurse eventually crashing
7843 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7846 *Imre Rad, Steve Henson*
7848 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7849 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7850 client or server. This is potentially exploitable to run arbitrary
7851 code on a vulnerable client or server.
7853 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7855 *Jüri Aedla, Steve Henson*
7857 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7858 are subject to a denial of service attack.
7860 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7861 this issue. ([CVE-2014-3470])
7863 *Felix Gröbert, Ivan Fratric, Steve Henson*
7865 * Harmonize version and its documentation. -f flag is used to display
7868 *mancha <mancha1@zoho.com>*
7870 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7871 in i2d_ECPrivateKey.
7873 *mancha <mancha1@zoho.com>*
7875 * Fix some double frees. These are not thought to be exploitable.
7877 *mancha <mancha1@zoho.com>*
7879 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7881 * A missing bounds check in the handling of the TLS heartbeat extension
7882 can be used to reveal up to 64k of memory to a connected client or
7885 Thanks for Neel Mehta of Google Security for discovering this bug and to
7886 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7887 preparing the fix ([CVE-2014-0160])
7889 *Adam Langley, Bodo Moeller*
7891 * Fix for the attack described in the paper "Recovering OpenSSL
7892 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7893 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7894 <http://eprint.iacr.org/2014/140>
7896 Thanks to Yuval Yarom and Naomi Benger for discovering this
7897 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7899 *Yuval Yarom and Naomi Benger*
7901 * TLS pad extension: draft-agl-tls-padding-03
7903 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7904 TLS client Hello record length value would otherwise be > 255 and
7905 less that 512 pad with a dummy extension containing zeroes so it
7906 is at least 512 bytes long.
7908 *Adam Langley, Steve Henson*
7910 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7912 * Fix for TLS record tampering bug. A carefully crafted invalid
7913 handshake could crash OpenSSL with a NULL pointer exception.
7914 Thanks to Anton Johansson for reporting this issues.
7917 * Keep original DTLS digest and encryption contexts in retransmission
7918 structures so we can use the previous session parameters if they need
7919 to be resent. ([CVE-2013-6450])
7923 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7924 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7925 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7926 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7927 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7928 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7930 *Rob Stradling, Adam Langley*
7932 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7934 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7935 supporting platforms or when small records were transferred.
7937 *Andy Polyakov, Steve Henson*
7939 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7941 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7943 This addresses the flaw in CBC record processing discovered by
7944 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7945 at: <http://www.isg.rhul.ac.uk/tls/>
7947 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7948 Security Group at Royal Holloway, University of London
7949 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7950 Emilia Käsper for the initial patch.
7953 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7955 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7956 ciphersuites which can be exploited in a denial of service attack.
7957 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7958 and detecting this bug and to Wolfgang Ettlinger
7959 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7964 * Return an error when checking OCSP signatures when key is NULL.
7965 This fixes a DoS attack. ([CVE-2013-0166])
7969 * Make openssl verify return errors.
7971 *Chris Palmer <palmer@google.com> and Ben Laurie*
7973 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7974 the right response is stapled. Also change SSL_get_certificate()
7975 so it returns the certificate actually sent.
7976 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7978 *Rob Stradling <rob.stradling@comodo.com>*
7980 * Fix possible deadlock when decoding public keys.
7984 * Don't use TLS 1.0 record version number in initial client hello
7989 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7991 * Sanity check record length before skipping explicit IV in TLS
7992 1.2, 1.1 and DTLS to fix DoS attack.
7994 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7995 fuzzing as a service testing platform.
8000 * Initialise tkeylen properly when encrypting CMS messages.
8001 Thanks to Solar Designer of Openwall for reporting this issue.
8005 * In FIPS mode don't try to use composite ciphers as they are not
8010 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8012 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8013 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8014 mean any application compiled against OpenSSL 1.0.0 headers setting
8015 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8016 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8017 0x10000000L Any application which was previously compiled against
8018 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8019 will need to be recompiled as a result. Letting be results in
8020 inability to disable specifically TLS 1.1 and in client context,
8021 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8025 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8026 disable just protocol X, but all protocols above X *if* there are
8027 protocols *below* X still enabled. In more practical terms it means
8028 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8029 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8030 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8035 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8037 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8038 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8039 in CRYPTO_realloc_clean.
8041 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8042 issue and to Adam Langley <agl@chromium.org> for fixing it.
8045 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8047 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8051 * Workarounds for some broken servers that "hang" if a client hello
8052 record length exceeds 255 bytes.
8054 1. Do not use record version number > TLS 1.0 in initial client
8055 hello: some (but not all) hanging servers will now work.
8056 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8057 the number of ciphers sent in the client hello. This should be
8058 set to an even number, such as 50, for example by passing:
8059 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8060 Most broken servers should now work.
8061 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8062 TLS 1.2 client support entirely.
8066 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8070 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8072 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8073 STRING form instead of a DigestInfo.
8077 * The format used for MDC2 RSA signatures is inconsistent between EVP
8078 and the RSA_sign/RSA_verify functions. This was made more apparent when
8079 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8080 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8081 the correct format in RSA_verify so both forms transparently work.
8085 * Some servers which support TLS 1.0 can choke if we initially indicate
8086 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8087 encrypted premaster secret. As a workaround use the maximum permitted
8088 client version in client hello, this should keep such servers happy
8089 and still work with previous versions of OpenSSL.
8093 * Add support for TLS/DTLS heartbeats.
8095 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8097 * Add support for SCTP.
8099 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8101 * Improved PRNG seeding for VOS.
8103 *Paul Green <Paul.Green@stratus.com>*
8105 * Extensive assembler packs updates, most notably:
8107 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8108 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8109 - x86_64: bit-sliced AES implementation;
8110 - ARM: NEON support, contemporary platforms optimizations;
8111 - s390x: z196 support;
8112 - `*`: GHASH and GF(2^m) multiplication implementations;
8116 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8117 (removal of unnecessary code)
8119 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8121 * Add TLS key material exporter from RFC 5705.
8125 * Add DTLS-SRTP negotiation from RFC 5764.
8129 * Add Next Protocol Negotiation,
8130 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8131 disabled with a no-npn flag to config or Configure. Code donated
8134 *Adam Langley <agl@google.com> and Ben Laurie*
8136 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8137 NIST-P256, NIST-P521, with constant-time single point multiplication on
8138 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8139 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8140 Code made available under Apache License version 2.0.
8142 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8143 line to include this in your build of OpenSSL, and run "make depend" (or
8144 "make update"). This enables the following EC_METHODs:
8146 EC_GFp_nistp224_method()
8147 EC_GFp_nistp256_method()
8148 EC_GFp_nistp521_method()
8150 EC_GROUP_new_by_curve_name() will automatically use these (while
8151 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8154 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8156 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8157 all platforms. Move ssize_t definition from e_os.h to the public
8158 header file e_os2.h as it now appears in public header file cms.h
8162 * New -sigopt option to the ca, req and x509 utilities. Additional
8163 signature parameters can be passed using this option and in
8168 * Add RSA PSS signing function. This will generate and set the
8169 appropriate AlgorithmIdentifiers for PSS based on those in the
8170 corresponding EVP_MD_CTX structure. No application support yet.
8174 * Support for companion algorithm specific ASN1 signing routines.
8175 New function ASN1_item_sign_ctx() signs a pre-initialised
8176 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8177 the appropriate parameters.
8181 * Add new algorithm specific ASN1 verification initialisation function
8182 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8183 handling will be the same no matter what EVP_PKEY_METHOD is used.
8184 Add a PSS handler to support verification of PSS signatures: checked
8185 against a number of sample certificates.
8189 * Add signature printing for PSS. Add PSS OIDs.
8191 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8193 * Add algorithm specific signature printing. An individual ASN1 method
8194 can now print out signatures instead of the standard hex dump.
8196 More complex signatures (e.g. PSS) can print out more meaningful
8197 information. Include DSA version that prints out the signature
8202 * Password based recipient info support for CMS library: implementing
8207 * Split password based encryption into PBES2 and PBKDF2 functions. This
8208 neatly separates the code into cipher and PBE sections and is required
8209 for some algorithms that split PBES2 into separate pieces (such as
8210 password based CMS).
8214 * Session-handling fixes:
8215 - Fix handling of connections that are resuming with a session ID,
8216 but also support Session Tickets.
8217 - Fix a bug that suppressed issuing of a new ticket if the client
8218 presented a ticket with an expired session.
8219 - Try to set the ticket lifetime hint to something reasonable.
8220 - Make tickets shorter by excluding irrelevant information.
8221 - On the client side, don't ignore renewed tickets.
8223 *Adam Langley, Bodo Moeller (Google)*
8225 * Fix PSK session representation.
8229 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8231 This work was sponsored by Intel.
8235 * Add GCM support to TLS library. Some custom code is needed to split
8236 the IV between the fixed (from PRF) and explicit (from TLS record)
8237 portions. This adds all GCM ciphersuites supported by RFC5288 and
8238 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8239 add a special AESGCM string for GCM only.
8243 * Expand range of ctrls for AES GCM. Permit setting invocation
8244 field on decrypt and retrieval of invocation field only on encrypt.
8248 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8249 As required by RFC5289 these ciphersuites cannot be used if for
8250 versions of TLS earlier than 1.2.
8254 * For FIPS capable OpenSSL interpret a NULL default public key method
8255 as unset and return the appropriate default but do *not* set the default.
8256 This means we can return the appropriate method in applications that
8257 switch between FIPS and non-FIPS modes.
8261 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8262 ENGINE is used then we cannot handle that in the FIPS module so we
8263 keep original code iff non-FIPS operations are allowed.
8267 * Add -attime option to openssl utilities.
8269 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8271 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8275 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8276 FIPS EC methods unconditionally for now.
8280 * New build option no-ec2m to disable characteristic 2 code.
8284 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8285 all cases can be covered as some introduce binary incompatibilities.
8289 * Redirect RSA operations to FIPS module including keygen,
8290 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8294 * Add similar low-level API blocking to ciphers.
8298 * low-level digest APIs are not approved in FIPS mode: any attempt
8299 to use these will cause a fatal error. Applications that *really* want
8300 to use them can use the `private_*` version instead.
8304 * Redirect cipher operations to FIPS module for FIPS builds.
8308 * Redirect digest operations to FIPS module for FIPS builds.
8312 * Update build system to add "fips" flag which will link in fipscanister.o
8313 for static and shared library builds embedding a signature if needed.
8317 * Output TLS supported curves in preference order instead of numerical
8318 order. This is currently hardcoded for the highest order curves first.
8319 This should be configurable so applications can judge speed vs strength.
8323 * Add TLS v1.2 server support for client authentication.
8327 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8332 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8333 FIPS modules versions.
8337 * Add TLS v1.2 client side support for client authentication. Keep cache
8338 of handshake records longer as we don't know the hash algorithm to use
8339 until after the certificate request message is received.
8343 * Initial TLS v1.2 client support. Add a default signature algorithms
8344 extension including all the algorithms we support. Parse new signature
8345 format in client key exchange. Relax some ECC signing restrictions for
8346 TLS v1.2 as indicated in RFC5246.
8350 * Add server support for TLS v1.2 signature algorithms extension. Switch
8351 to new signature format when needed using client digest preference.
8352 All server ciphersuites should now work correctly in TLS v1.2. No client
8353 support yet and no support for client certificates.
8357 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8358 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8359 ciphersuites. At present only RSA key exchange ciphersuites work with
8360 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8361 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8362 and version checking.
8366 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8367 with this defined it will not be affected by any changes to ssl internal
8368 structures. Add several utility functions to allow openssl application
8369 to work with OPENSSL_NO_SSL_INTERN defined.
8373 * A long standing patch to add support for SRP from EdelWeb (Peter
8374 Sylvester and Christophe Renou) was integrated.
8375 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8376 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8379 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8383 * Permit abbreviated handshakes when renegotiating using the function
8384 SSL_renegotiate_abbreviated().
8386 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8388 * Add call to ENGINE_register_all_complete() to
8389 ENGINE_load_builtin_engines(), so some implementations get used
8390 automatically instead of needing explicit application support.
8394 * Add support for TLS key exporter as described in RFC5705.
8396 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8398 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8399 a few changes are required:
8401 Add SSL_OP_NO_TLSv1_1 flag.
8402 Add TLSv1_1 methods.
8403 Update version checking logic to handle version 1.1.
8404 Add explicit IV handling (ported from DTLS code).
8405 Add command line options to s_client/s_server.
8412 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8414 * X509_ATTRIBUTE memory leak
8416 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8417 memory. This structure is used by the PKCS#7 and CMS routines so any
8418 application which reads PKCS#7 or CMS data from untrusted sources is
8419 affected. SSL/TLS is not affected.
8421 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8427 * Race condition handling PSK identify hint
8429 If PSK identity hints are received by a multi-threaded client then
8430 the values are wrongly updated in the parent SSL_CTX structure. This can
8431 result in a race condition potentially leading to a double free of the
8437 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8439 * Malformed ECParameters causes infinite loop
8441 When processing an ECParameters structure OpenSSL enters an infinite loop
8442 if the curve specified is over a specially malformed binary polynomial
8445 This can be used to perform denial of service against any
8446 system which processes public keys, certificate requests or
8447 certificates. This includes TLS clients and TLS servers with
8448 client authentication enabled.
8450 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8455 * Exploitable out-of-bounds read in X509_cmp_time
8457 X509_cmp_time does not properly check the length of the ASN1_TIME
8458 string and can read a few bytes out of bounds. In addition,
8459 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8462 An attacker can use this to craft malformed certificates and CRLs of
8463 various sizes and potentially cause a segmentation fault, resulting in
8464 a DoS on applications that verify certificates or CRLs. TLS clients
8465 that verify CRLs are affected. TLS clients and servers with client
8466 authentication enabled may be affected if they use custom verification
8469 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8470 independently by Hanno Böck.
8475 * PKCS7 crash with missing EnvelopedContent
8477 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8478 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8479 with missing content and trigger a NULL pointer dereference on parsing.
8481 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8482 structures from untrusted sources are affected. OpenSSL clients and
8483 servers are not affected.
8485 This issue was reported to OpenSSL by Michal Zalewski (Google).
8490 * CMS verify infinite loop with unknown hash function
8492 When verifying a signedData message the CMS code can enter an infinite loop
8493 if presented with an unknown hash function OID. This can be used to perform
8494 denial of service against any system which verifies signedData messages using
8496 This issue was reported to OpenSSL by Johannes Bauer.
8501 * Race condition handling NewSessionTicket
8503 If a NewSessionTicket is received by a multi-threaded client when attempting to
8504 reuse a previous ticket then a race condition can occur potentially leading to
8505 a double free of the ticket data.
8510 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8512 * Segmentation fault in ASN1_TYPE_cmp fix
8514 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8515 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8516 certificate signature algorithm consistency this can be used to crash any
8517 certificate verification operation and exploited in a DoS attack. Any
8518 application which performs certificate verification is vulnerable including
8519 OpenSSL clients and servers which enable client authentication.
8524 * ASN.1 structure reuse memory corruption fix
8526 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8527 memory corruption via an invalid write. Such reuse is and has been
8528 strongly discouraged and is believed to be rare.
8530 Applications that parse structures containing CHOICE or ANY DEFINED BY
8531 components may be affected. Certificate parsing (d2i_X509 and related
8532 functions) are however not affected. OpenSSL clients and servers are
8538 * PKCS7 NULL pointer dereferences fix
8540 The PKCS#7 parsing code does not handle missing outer ContentInfo
8541 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8542 missing content and trigger a NULL pointer dereference on parsing.
8544 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8545 otherwise parse PKCS#7 structures from untrusted sources are
8546 affected. OpenSSL clients and servers are not affected.
8548 This issue was reported to OpenSSL by Michal Zalewski (Google).
8553 * DoS via reachable assert in SSLv2 servers fix
8555 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8556 servers that both support SSLv2 and enable export cipher suites by sending
8557 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8559 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8560 (OpenSSL development team).
8565 * Use After Free following d2i_ECPrivatekey error fix
8567 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8568 could cause a use after free condition. This, in turn, could cause a double
8569 free in several private key parsing functions (such as d2i_PrivateKey
8570 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8571 for applications that receive EC private keys from untrusted
8572 sources. This scenario is considered rare.
8574 This issue was discovered by the BoringSSL project and fixed in their
8580 * X509_to_X509_REQ NULL pointer deref fix
8582 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8583 the certificate key is invalid. This function is rarely used in practice.
8585 This issue was discovered by Brian Carpenter.
8590 * Removed the export ciphers from the DEFAULT ciphers
8594 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8596 * Build fixes for the Windows and OpenVMS platforms
8598 *Matt Caswell and Richard Levitte*
8600 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8602 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8603 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8604 dereference. This could lead to a Denial Of Service attack. Thanks to
8605 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8610 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8611 dtls1_buffer_record function under certain conditions. In particular this
8612 could occur if an attacker sent repeated DTLS records with the same
8613 sequence number but for the next epoch. The memory leak could be exploited
8614 by an attacker in a Denial of Service attack through memory exhaustion.
8615 Thanks to Chris Mueller for reporting this issue.
8620 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8621 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8622 method would be set to NULL which could later result in a NULL pointer
8623 dereference. Thanks to Frank Schmirler for reporting this issue.
8628 * Abort handshake if server key exchange message is omitted for ephemeral
8631 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8632 reporting this issue.
8637 * Remove non-export ephemeral RSA code on client and server. This code
8638 violated the TLS standard by allowing the use of temporary RSA keys in
8639 non-export ciphersuites and could be used by a server to effectively
8640 downgrade the RSA key length used to a value smaller than the server
8641 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8642 INRIA or reporting this issue.
8647 * Fixed issue where DH client certificates are accepted without verification.
8648 An OpenSSL server will accept a DH certificate for client authentication
8649 without the certificate verify message. This effectively allows a client to
8650 authenticate without the use of a private key. This only affects servers
8651 which trust a client certificate authority which issues certificates
8652 containing DH keys: these are extremely rare and hardly ever encountered.
8653 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8659 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8660 results on some platforms, including x86_64. This bug occurs at random
8661 with a very low probability, and is not known to be exploitable in any
8662 way, though its exact impact is difficult to determine. Thanks to Pieter
8663 Wuille (Blockstream) who reported this issue and also suggested an initial
8664 fix. Further analysis was conducted by the OpenSSL development team and
8665 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8666 the OpenSSL core team.
8671 * Fix various certificate fingerprint issues.
8673 By using non-DER or invalid encodings outside the signed portion of a
8674 certificate the fingerprint can be changed without breaking the signature.
8675 Although no details of the signed portion of the certificate can be changed
8676 this can cause problems with some applications: e.g. those using the
8677 certificate fingerprint for blacklists.
8679 1. Reject signatures with non zero unused bits.
8681 If the BIT STRING containing the signature has non zero unused bits reject
8682 the signature. All current signature algorithms require zero unused bits.
8684 2. Check certificate algorithm consistency.
8686 Check the AlgorithmIdentifier inside TBS matches the one in the
8687 certificate signature. NB: this will result in signature failure
8688 errors for some broken certificates.
8690 Thanks to Konrad Kraszewski from Google for reporting this issue.
8692 3. Check DSA/ECDSA signatures use DER.
8694 Reencode DSA/ECDSA signatures and compare with the original received
8695 signature. Return an error if there is a mismatch.
8697 This will reject various cases including garbage after signature
8698 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8699 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8700 (negative or with leading zeroes).
8702 Further analysis was conducted and fixes were developed by Stephen Henson
8703 of the OpenSSL core team.
8709 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8711 * Session Ticket Memory Leak.
8713 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8714 integrity of that ticket is first verified. In the event of a session
8715 ticket integrity check failing, OpenSSL will fail to free memory
8716 causing a memory leak. By sending a large number of invalid session
8717 tickets an attacker could exploit this issue in a Denial Of Service
8723 * Build option no-ssl3 is incomplete.
8725 When OpenSSL is configured with "no-ssl3" as a build option, servers
8726 could accept and complete an SSL 3.0 handshake, and clients could be
8727 configured to send them.
8730 *Akamai and the OpenSSL team*
8732 * Add support for TLS_FALLBACK_SCSV.
8733 Client applications doing fallback retries should call
8734 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8737 *Adam Langley, Bodo Moeller*
8739 * Add additional DigestInfo checks.
8741 Reencode DigestInto in DER and check against the original when
8742 verifying RSA signature: this will reject any improperly encoded
8743 DigestInfo structures.
8745 Note: this is a precautionary measure and no attacks are currently known.
8749 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8751 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8752 to a denial of service attack. A malicious server can crash the client
8753 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8754 ciphersuite and sending carefully crafted handshake messages.
8756 Thanks to Felix Gröbert (Google) for discovering and researching this
8762 * By sending carefully crafted DTLS packets an attacker could cause openssl
8763 to leak memory. This can be exploited through a Denial of Service attack.
8764 Thanks to Adam Langley for discovering and researching this issue.
8769 * An attacker can force openssl to consume large amounts of memory whilst
8770 processing DTLS handshake messages. This can be exploited through a
8771 Denial of Service attack.
8772 Thanks to Adam Langley for discovering and researching this issue.
8777 * An attacker can force an error condition which causes openssl to crash
8778 whilst processing DTLS packets due to memory being freed twice. This
8779 can be exploited through a Denial of Service attack.
8780 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8786 * If a multithreaded client connects to a malicious server using a resumed
8787 session and the server sends an ec point format extension it could write
8788 up to 255 bytes to freed memory.
8790 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8796 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8797 X509_name_oneline, X509_name_print_ex et al. to leak some information
8798 from the stack. Applications may be affected if they echo pretty printing
8799 output to the attacker.
8801 Thanks to Ivan Fratric (Google) for discovering this issue.
8804 *Emilia Käsper, and Steve Henson*
8806 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8807 for corner cases. (Certain input points at infinity could lead to
8808 bogus results, with non-infinity inputs mapped to infinity too.)
8812 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8814 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8815 handshake can force the use of weak keying material in OpenSSL
8816 SSL/TLS clients and servers.
8818 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8819 researching this issue. ([CVE-2014-0224])
8821 *KIKUCHI Masashi, Steve Henson*
8823 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8824 OpenSSL DTLS client the code can be made to recurse eventually crashing
8827 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8830 *Imre Rad, Steve Henson*
8832 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8833 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8834 client or server. This is potentially exploitable to run arbitrary
8835 code on a vulnerable client or server.
8837 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8839 *Jüri Aedla, Steve Henson*
8841 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8842 are subject to a denial of service attack.
8844 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8845 this issue. ([CVE-2014-3470])
8847 *Felix Gröbert, Ivan Fratric, Steve Henson*
8849 * Harmonize version and its documentation. -f flag is used to display
8852 *mancha <mancha1@zoho.com>*
8854 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8855 in i2d_ECPrivateKey.
8857 *mancha <mancha1@zoho.com>*
8859 * Fix some double frees. These are not thought to be exploitable.
8861 *mancha <mancha1@zoho.com>*
8863 * Fix for the attack described in the paper "Recovering OpenSSL
8864 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8865 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8866 <http://eprint.iacr.org/2014/140>
8868 Thanks to Yuval Yarom and Naomi Benger for discovering this
8869 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8871 *Yuval Yarom and Naomi Benger*
8873 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8875 * Keep original DTLS digest and encryption contexts in retransmission
8876 structures so we can use the previous session parameters if they need
8877 to be resent. ([CVE-2013-6450])
8881 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8882 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8883 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8884 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8885 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8886 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8888 *Rob Stradling, Adam Langley*
8890 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8892 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8894 This addresses the flaw in CBC record processing discovered by
8895 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8896 at: <http://www.isg.rhul.ac.uk/tls/>
8898 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8899 Security Group at Royal Holloway, University of London
8900 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8901 Emilia Käsper for the initial patch.
8904 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8906 * Return an error when checking OCSP signatures when key is NULL.
8907 This fixes a DoS attack. ([CVE-2013-0166])
8911 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8912 the right response is stapled. Also change SSL_get_certificate()
8913 so it returns the certificate actually sent.
8914 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8915 (This is a backport)
8917 *Rob Stradling <rob.stradling@comodo.com>*
8919 * Fix possible deadlock when decoding public keys.
8923 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8925 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8928 * Sanity check record length before skipping explicit IV in DTLS
8931 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8932 fuzzing as a service testing platform.
8937 * Initialise tkeylen properly when encrypting CMS messages.
8938 Thanks to Solar Designer of Openwall for reporting this issue.
8942 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8944 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8945 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8946 in CRYPTO_realloc_clean.
8948 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8949 issue and to Adam Langley <agl@chromium.org> for fixing it.
8952 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8954 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8956 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8957 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8958 content decryption and always return the same error. Note: this attack
8959 needs on average 2^20 messages so it only affects automated senders. The
8960 old behaviour can be re-enabled in the CMS code by setting the
8961 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8962 an MMA defence is not necessary.
8963 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8964 this issue. ([CVE-2012-0884])
8968 * Fix CVE-2011-4619: make sure we really are receiving a
8969 client hello before rejecting multiple SGC restarts. Thanks to
8970 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8974 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8976 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8977 Thanks to Antonio Martin, Enterprise Secure Access Research and
8978 Development, Cisco Systems, Inc. for discovering this bug and
8979 preparing a fix. ([CVE-2012-0050])
8983 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8985 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8986 of the Vaudenay padding oracle attack on CBC mode encryption
8987 which enables an efficient plaintext recovery attack against
8988 the OpenSSL implementation of DTLS. Their attack exploits timing
8989 differences arising during decryption processing. A research
8990 paper describing this attack can be found at:
8991 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8992 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8993 Security Group at Royal Holloway, University of London
8994 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8995 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8996 for preparing the fix. ([CVE-2011-4108])
8998 *Robin Seggelmann, Michael Tuexen*
9000 * Clear bytes used for block padding of SSL 3.0 records.
9003 *Adam Langley (Google)*
9005 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9006 Kadianakis <desnacked@gmail.com> for discovering this issue and
9007 Adam Langley for preparing the fix. ([CVE-2011-4619])
9009 *Adam Langley (Google)*
9011 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9013 *Andrey Kulikov <amdeich@gmail.com>*
9015 * Prevent malformed RFC3779 data triggering an assertion failure.
9016 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9017 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9019 *Rob Austein <sra@hactrn.net>*
9021 * Improved PRNG seeding for VOS.
9023 *Paul Green <Paul.Green@stratus.com>*
9025 * Fix ssl_ciph.c set-up race.
9027 *Adam Langley (Google)*
9029 * Fix spurious failures in ecdsatest.c.
9031 *Emilia Käsper (Google)*
9033 * Fix the BIO_f_buffer() implementation (which was mixing different
9034 interpretations of the `..._len` fields).
9036 *Adam Langley (Google)*
9038 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9039 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9040 threads won't reuse the same blinding coefficients.
9042 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9043 lock to call BN_BLINDING_invert_ex, and avoids one use of
9044 BN_BLINDING_update for each BN_BLINDING structure (previously,
9045 the last update always remained unused).
9047 *Emilia Käsper (Google)*
9049 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9051 *Bob Buckholz (Google)*
9053 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9055 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9056 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9058 *Kaspar Brand <ossl@velox.ch>*
9060 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9061 for multi-threaded use of ECDH. ([CVE-2011-3210])
9063 *Adam Langley (Google)*
9065 * Fix x509_name_ex_d2i memory leak on bad inputs.
9069 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9070 signature public key algorithm by using OID xref utilities instead.
9071 Before this you could only use some ECC ciphersuites with SHA1 only.
9075 * Add protection against ECDSA timing attacks as mentioned in the paper
9076 by Billy Bob Brumley and Nicola Tuveri, see:
9077 <http://eprint.iacr.org/2011/232.pdf>
9079 *Billy Bob Brumley and Nicola Tuveri*
9081 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9083 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9085 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9087 * Fix bug in string printing code: if *any* escaping is enabled we must
9088 escape the escape character (backslash) or the resulting string is
9093 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9095 * Disable code workaround for ancient and obsolete Netscape browsers
9096 and servers: an attacker can use it in a ciphersuite downgrade attack.
9097 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9101 * Fixed J-PAKE implementation error, originally discovered by
9102 Sebastien Martini, further info and confirmation from Stefan
9103 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9107 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9109 * Fix extension code to avoid race conditions which can result in a buffer
9110 overrun vulnerability: resumed sessions must not be modified as they can
9111 be shared by multiple threads. CVE-2010-3864
9115 * Fix WIN32 build system to correctly link an ENGINE directory into
9120 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9122 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9125 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9127 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9129 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9130 context. The operation can be customised via the ctrl mechanism in
9131 case ENGINEs want to include additional functionality.
9135 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9139 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9140 output hashes compatible with older versions of OpenSSL.
9142 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9144 * Fix compression algorithm handling: if resuming a session use the
9145 compression algorithm of the resumed session instead of determining
9146 it from client hello again. Don't allow server to change algorithm.
9150 * Add load_crls() function to commands tidying load_certs() too. Add option
9151 to verify utility to allow additional CRLs to be included.
9155 * Update OCSP request code to permit adding custom headers to the request:
9156 some responders need this.
9160 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9163 *Julia Lawall <julia@diku.dk>*
9165 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9166 needlessly dereferenced structures, used obsolete functions and
9167 didn't handle all updated verify codes correctly.
9171 * Disable MD2 in the default configuration.
9175 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9176 indicate the initial BIO being pushed or popped. This makes it possible
9177 to determine whether the BIO is the one explicitly called or as a result
9178 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9179 it handles reference counts correctly and doesn't zero out the I/O bio
9180 when it is not being explicitly popped. WARNING: applications which
9181 included workarounds for the old buggy behaviour will need to be modified
9182 or they could free up already freed BIOs.
9186 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9187 renaming to all platforms (within the 0.9.8 branch, this was
9188 done conditionally on Netware platforms to avoid a name clash).
9190 *Guenter <lists@gknw.net>*
9192 * Add ECDHE and PSK support to DTLS.
9194 *Michael Tuexen <tuexen@fh-muenster.de>*
9196 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9201 * Add "missing" function EVP_MD_flags() (without this the only way to
9202 retrieve a digest flags is by accessing the structure directly. Update
9203 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9204 or cipher is registered as in the "from" argument. Print out all
9205 registered digests in the dgst usage message instead of manually
9206 attempting to work them out.
9210 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9211 this allows the use of compression and extensions. Change default cipher
9212 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9213 by default unless an application cipher string requests it.
9217 * Alter match criteria in PKCS12_parse(). It used to try to use local
9218 key ids to find matching certificates and keys but some PKCS#12 files
9219 don't follow the (somewhat unwritten) rules and this strategy fails.
9220 Now just gather all certificates together and the first private key
9221 then look for the first certificate that matches the key.
9225 * Support use of registered digest and cipher names for dgst and cipher
9226 commands instead of having to add each one as a special case. So now
9233 openssl dgst -sha256 foo
9235 and this works for ENGINE based algorithms too.
9239 * Update Gost ENGINE to support parameter files.
9241 *Victor B. Wagner <vitus@cryptocom.ru>*
9243 * Support GeneralizedTime in ca utility.
9245 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9247 * Enhance the hash format used for certificate directory links. The new
9248 form uses the canonical encoding (meaning equivalent names will work
9249 even if they aren't identical) and uses SHA1 instead of MD5. This form
9250 is incompatible with the older format and as a result c_rehash should
9251 be used to rebuild symbolic links.
9255 * Make PKCS#8 the default write format for private keys, replacing the
9256 traditional format. This form is standardised, more secure and doesn't
9257 include an implicit MD5 dependency.
9261 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9262 committed to OpenSSL should pass this lot as a minimum.
9266 * Add session ticket override functionality for use by EAP-FAST.
9268 *Jouni Malinen <j@w1.fi>*
9270 * Modify HMAC functions to return a value. Since these can be implemented
9271 in an ENGINE errors can occur.
9275 * Type-checked OBJ_bsearch_ex.
9279 * Type-checked OBJ_bsearch. Also some constification necessitated
9280 by type-checking. Still to come: TXT_DB, bsearch(?),
9281 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9286 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9287 seconds to a tm structure directly, instead of going through OS
9288 specific date routines. This avoids any issues with OS routines such
9289 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9290 and X509_time_adj_ex() to cover the extended range. The existing
9291 X509_time_adj() is still usable and will no longer have any date issues.
9295 * Delta CRL support. New use deltas option which will attempt to locate
9296 and search any appropriate delta CRLs available.
9298 This work was sponsored by Google.
9302 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9303 code and add additional score elements. Validate alternate CRL paths
9304 as part of the CRL checking and indicate a new error "CRL path validation
9305 error" in this case. Applications wanting additional details can use
9306 the verify callback and check the new "parent" field. If this is not
9307 NULL CRL path validation is taking place. Existing applications won't
9308 see this because it requires extended CRL support which is off by
9311 This work was sponsored by Google.
9315 * Support for freshest CRL extension.
9317 This work was sponsored by Google.
9321 * Initial indirect CRL support. Currently only supported in the CRLs
9322 passed directly and not via lookup. Process certificate issuer
9323 CRL entry extension and lookup CRL entries by bother issuer name
9324 and serial number. Check and process CRL issuer entry in IDP extension.
9326 This work was sponsored by Google.
9330 * Add support for distinct certificate and CRL paths. The CRL issuer
9331 certificate is validated separately in this case. Only enabled if
9332 an extended CRL support flag is set: this flag will enable additional
9333 CRL functionality in future.
9335 This work was sponsored by Google.
9339 * Add support for policy mappings extension.
9341 This work was sponsored by Google.
9345 * Fixes to pathlength constraint, self issued certificate handling,
9346 policy processing to align with RFC3280 and PKITS tests.
9348 This work was sponsored by Google.
9352 * Support for name constraints certificate extension. DN, email, DNS
9353 and URI types are currently supported.
9355 This work was sponsored by Google.
9359 * To cater for systems that provide a pointer-based thread ID rather
9360 than numeric, deprecate the current numeric thread ID mechanism and
9361 replace it with a structure and associated callback type. This
9362 mechanism allows a numeric "hash" to be extracted from a thread ID in
9363 either case, and on platforms where pointers are larger than 'long',
9364 mixing is done to help ensure the numeric 'hash' is usable even if it
9365 can't be guaranteed unique. The default mechanism is to use "&errno"
9366 as a pointer-based thread ID to distinguish between threads.
9368 Applications that want to provide their own thread IDs should now use
9369 CRYPTO_THREADID_set_callback() to register a callback that will call
9370 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9372 Note that ERR_remove_state() is now deprecated, because it is tied
9373 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9374 to free the current thread's error state should be replaced by
9375 ERR_remove_thread_state(NULL).
9377 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9378 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9379 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9380 application was previously providing a numeric thread callback that
9381 was inappropriate for distinguishing threads, then uniqueness might
9382 have been obtained with &errno that happened immediately in the
9383 intermediate development versions of OpenSSL; this is no longer the
9384 case, the numeric thread callback will now override the automatic use
9387 *Geoff Thorpe, with help from Bodo Moeller*
9389 * Initial support for different CRL issuing certificates. This covers a
9390 simple case where the self issued certificates in the chain exist and
9391 the real CRL issuer is higher in the existing chain.
9393 This work was sponsored by Google.
9397 * Removed effectively defunct crypto/store from the build.
9401 * Revamp of STACK to provide stronger type-checking. Still to come:
9402 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9403 ASN1_STRING, CONF_VALUE.
9407 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9408 RAM on SSL connections. This option can save about 34k per idle SSL.
9412 * Revamp of LHASH to provide stronger type-checking. Still to come:
9413 STACK, TXT_DB, bsearch, qsort.
9417 * Initial support for Cryptographic Message Syntax (aka CMS) based
9418 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9419 support for data, signedData, compressedData, digestedData and
9420 encryptedData, envelopedData types included. Scripts to check against
9421 RFC4134 examples draft and interop and consistency checks of many
9422 content types and variants.
9426 * Add options to enc utility to support use of zlib compression BIO.
9430 * Extend mk1mf to support importing of options and assembly language
9431 files from Configure script, currently only included in VC-WIN32.
9432 The assembly language rules can now optionally generate the source
9433 files from the associated perl scripts.
9437 * Implement remaining functionality needed to support GOST ciphersuites.
9438 Interop testing has been performed using CryptoPro implementations.
9440 *Victor B. Wagner <vitus@cryptocom.ru>*
9442 * s390x assembler pack.
9446 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9451 * Implement Opaque PRF Input TLS extension as specified in
9452 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9453 official specification yet and no extension type assignment by
9454 IANA exists, this extension (for now) will have to be explicitly
9455 enabled when building OpenSSL by providing the extension number
9456 to use. For example, specify an option
9458 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9460 to the "config" or "Configure" script to enable the extension,
9461 assuming extension number 0x9527 (which is a completely arbitrary
9462 and unofficial assignment based on the MD5 hash of the Internet
9463 Draft). Note that by doing so, you potentially lose
9464 interoperability with other TLS implementations since these might
9465 be using the same extension number for other purposes.
9467 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9468 opaque PRF input value to use in the handshake. This will create
9469 an internal copy of the length-'len' string at 'src', and will
9470 return non-zero for success.
9472 To get more control and flexibility, provide a callback function
9475 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9476 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9480 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9483 Callback function 'cb' will be called in handshakes, and is
9484 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9485 Argument 'arg' is for application purposes (the value as given to
9486 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9487 be provided to the callback function). The callback function
9488 has to return non-zero to report success: usually 1 to use opaque
9489 PRF input just if possible, or 2 to enforce use of the opaque PRF
9490 input. In the latter case, the library will abort the handshake
9491 if opaque PRF input is not successfully negotiated.
9493 Arguments 'peerinput' and 'len' given to the callback function
9494 will always be NULL and 0 in the case of a client. A server will
9495 see the client's opaque PRF input through these variables if
9496 available (NULL and 0 otherwise). Note that if the server
9497 provides an opaque PRF input, the length must be the same as the
9498 length of the client's opaque PRF input.
9500 Note that the callback function will only be called when creating
9501 a new session (session resumption can resume whatever was
9502 previously negotiated), and will not be called in SSL 2.0
9503 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9504 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9505 for applications that need to enforce opaque PRF input.
9509 * Update ssl code to support digests other than SHA1+MD5 for handshake
9512 *Victor B. Wagner <vitus@cryptocom.ru>*
9514 * Add RFC4507 support to OpenSSL. This includes the corrections in
9515 RFC4507bis. The encrypted ticket format is an encrypted encoded
9516 SSL_SESSION structure, that way new session features are automatically
9519 If a client application caches session in an SSL_SESSION structure
9520 support is transparent because tickets are now stored in the encoded
9523 The SSL_CTX structure automatically generates keys for ticket
9524 protection in servers so again support should be possible
9525 with no application modification.
9527 If a client or server wishes to disable RFC4507 support then the option
9528 SSL_OP_NO_TICKET can be set.
9530 Add a TLS extension debugging callback to allow the contents of any client
9531 or server extensions to be examined.
9533 This work was sponsored by Google.
9537 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9538 OpenSSL should now compile cleanly on gcc 4.2
9540 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9542 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9543 support including streaming MAC support: this is required for GOST
9544 ciphersuite support.
9546 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9548 * Add option -stream to use PKCS#7 streaming in smime utility. New
9549 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9550 to output in BER and PEM format.
9554 * Experimental support for use of HMAC via EVP_PKEY interface. This
9555 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9556 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9557 ENGINE support for HMAC keys which are unextractable. New -mac and
9558 -macopt options to dgst utility.
9562 * New option -sigopt to dgst utility. Update dgst to use
9563 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9564 alternative signing parameters such as X9.31 or PSS in the dgst
9569 * Change ssl_cipher_apply_rule(), the internal function that does
9570 the work each time a ciphersuite string requests enabling
9571 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9572 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9573 the order of disabled ciphersuites such that those ciphersuites
9574 that most recently went from enabled to disabled not only stay
9575 in order with respect to each other, but also have higher priority
9576 than other disabled ciphersuites the next time ciphersuites are
9579 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9580 the same ciphersuites as with "HIGH" alone, but in a specific
9581 order where the PSK ciphersuites come first (since they are the
9582 most recently disabled ciphersuites when "HIGH" is parsed).
9584 Also, change ssl_create_cipher_list() (using this new
9585 functionality) such that between otherwise identical
9586 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9591 * Change ssl_create_cipher_list() so that it automatically
9592 arranges the ciphersuites in reasonable order before starting
9593 to process the rule string. Thus, the definition for "DEFAULT"
9594 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9595 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9596 This makes it much easier to arrive at a reasonable default order
9597 in applications for which anonymous ciphers are OK (meaning
9598 that you can't actually use DEFAULT).
9600 *Bodo Moeller; suggested by Victor Duchovni*
9602 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9603 processing) into multiple integers instead of setting
9604 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9605 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9606 (These masks as well as the individual bit definitions are hidden
9607 away into the non-exported interface ssl/ssl_locl.h, so this
9608 change to the definition of the SSL_CIPHER structure shouldn't
9609 affect applications.) This give us more bits for each of these
9610 categories, so there is no longer a need to coagulate AES128 and
9611 AES256 into a single algorithm bit, and to coagulate Camellia128
9612 and Camellia256 into a single algorithm bit, which has led to all
9615 Thus, among other things, the kludge introduced in 0.9.7m and
9616 0.9.8e for masking out AES256 independently of AES128 or masking
9617 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9619 With the change, we also introduce new ciphersuite aliases that
9620 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9625 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9626 Use the leftmost N bytes of the signature input if the input is
9627 larger than the prime q (with N being the size in bytes of q).
9631 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9632 it yet and it is largely untested.
9636 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9640 * Initial incomplete changes to avoid need for function casts in OpenSSL
9641 some compilers (gcc 4.2 and later) reject their use. Safestack is
9642 reimplemented. Update ASN1 to avoid use of legacy functions.
9646 * Win32/64 targets are linked with Winsock2.
9650 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9651 to external functions. This can be used to increase CRL handling
9652 efficiency especially when CRLs are very large by (for example) storing
9653 the CRL revoked certificates in a database.
9657 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9658 new CRLs added to a directory can be used. New command line option
9659 -verify_return_error to s_client and s_server. This causes real errors
9660 to be returned by the verify callback instead of carrying on no matter
9661 what. This reflects the way a "real world" verify callback would behave.
9665 * GOST engine, supporting several GOST algorithms and public key formats.
9666 Kindly donated by Cryptocom.
9670 * Partial support for Issuing Distribution Point CRL extension. CRLs
9671 partitioned by DP are handled but no indirect CRL or reason partitioning
9672 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9673 selected via a scoring technique which handles IDP and AKID in CRLs.
9677 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9678 will ultimately be used for all verify operations: this will remove the
9679 X509_STORE dependency on certificate verification and allow alternative
9680 lookup methods. X509_STORE based implementations of these two callbacks.
9684 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9685 Modify get_crl() to find a valid (unexpired) CRL if possible.
9689 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9690 this would be called X509_CRL_cmp() but that name is already used by
9691 a function that just compares CRL issuer names. Cache several CRL
9692 extensions in X509_CRL structure and cache CRLDP in X509.
9696 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9697 this maps equivalent X509_NAME structures into a consistent structure.
9698 Name comparison can then be performed rapidly using memcmp().
9702 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9707 * Allow digests to supply their own micalg string for S/MIME type using
9708 the ctrl EVP_MD_CTRL_MICALG.
9712 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9713 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9714 ctrl. It can then customise the structure before and/or after signing
9719 * New function OBJ_add_sigid() to allow application defined signature OIDs
9720 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9721 to free up any added signature OIDs.
9725 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9726 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9727 digest and cipher tables. New options added to openssl utility:
9728 list-message-digest-algorithms and list-cipher-algorithms.
9732 * Change the array representation of binary polynomials: the list
9733 of degrees of non-zero coefficients is now terminated with -1.
9734 Previously it was terminated with 0, which was also part of the
9735 value; thus, the array representation was not applicable to
9736 polynomials where t^0 has coefficient zero. This change makes
9737 the array representation useful in a more general context.
9741 * Various modifications and fixes to SSL/TLS cipher string
9742 handling. For ECC, the code now distinguishes between fixed ECDH
9743 with RSA certificates on the one hand and with ECDSA certificates
9744 on the other hand, since these are separate ciphersuites. The
9745 unused code for Fortezza ciphersuites has been removed.
9747 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9748 (not "ECDHE"). For consistency with the code for DH
9749 certificates, use of ECDH certificates is now considered ECDH
9750 authentication, not RSA or ECDSA authentication (the latter is
9751 merely the CA's signing algorithm and not actively used in the
9754 The temporary ciphersuite alias "ECCdraft" is no longer
9755 available, and ECC ciphersuites are no longer excluded from "ALL"
9756 and "DEFAULT". The following aliases now exist for RFC 4492
9757 ciphersuites, most of these by analogy with the DH case:
9759 kECDHr - ECDH cert, signed with RSA
9760 kECDHe - ECDH cert, signed with ECDSA
9761 kECDH - ECDH cert (signed with either RSA or ECDSA)
9762 kEECDH - ephemeral ECDH
9763 ECDH - ECDH cert or ephemeral ECDH
9769 AECDH - anonymous ECDH
9770 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9774 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9775 Use correct micalg parameters depending on digest(s) in signed message.
9779 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9780 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9784 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9785 an engine to register a method. Add ENGINE lookups for methods and
9786 functional reference processing.
9790 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9791 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9796 * New -resign option to smime utility. This adds one or more signers
9797 to an existing PKCS#7 signedData structure. Also -md option to use an
9798 alternative message digest algorithm for signing.
9802 * Tidy up PKCS#7 routines and add new functions to make it easier to
9803 create PKCS7 structures containing multiple signers. Update smime
9804 application to support multiple signers.
9808 * New -macalg option to pkcs12 utility to allow setting of an alternative
9813 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9814 Reorganize PBE internals to lookup from a static table using NIDs,
9815 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9816 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9817 PRF which will be automatically used with PBES2.
9821 * Replace the algorithm specific calls to generate keys in "req" with the
9826 * Update PKCS#7 enveloped data routines to use new API. This is now
9827 supported by any public key method supporting the encrypt operation. A
9828 ctrl is added to allow the public key algorithm to examine or modify
9829 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9834 * Add a ctrl to asn1 method to allow a public key algorithm to express
9835 a default digest type to use. In most cases this will be SHA1 but some
9836 algorithms (such as GOST) need to specify an alternative digest. The
9837 return value indicates how strong the preference is 1 means optional and
9838 2 is mandatory (that is it is the only supported type). Modify
9839 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9840 use the default md. Update openssl utilities to use the default digest
9841 type for signing if it is not explicitly indicated.
9845 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9846 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9847 signing method from the key type. This effectively removes the link
9848 between digests and public key types.
9852 * Add an OID cross reference table and utility functions. Its purpose is to
9853 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9854 rsaEncryption. This will allow some of the algorithm specific hackery
9855 needed to use the correct OID to be removed.
9859 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9860 structures for PKCS7_sign(). They are now set up by the relevant public
9865 * Add provisional EC pkey method with support for ECDSA and ECDH.
9869 * Add support for key derivation (agreement) in the API, DH method and
9874 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9875 public and private key formats. As a side effect these add additional
9876 command line functionality not previously available: DSA signatures can be
9877 generated and verified using pkeyutl and DH key support and generation in
9884 *Oliver Tappe <zooey@hirschkaefer.de>*
9886 * New make target "install_html_docs" installs HTML renditions of the
9889 *Oliver Tappe <zooey@hirschkaefer.de>*
9891 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9892 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9893 support key and parameter generation and add initial key generation
9894 functionality for RSA.
9898 * Add functions for main EVP_PKEY_method operations. The undocumented
9899 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9900 `EVP_PKEY_{encrypt,decrypt}_old`.
9904 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9905 key API, doesn't do much yet.
9909 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9910 public key algorithms. New option to openssl utility:
9911 "list-public-key-algorithms" to print out info.
9915 * Implement the Supported Elliptic Curves Extension for
9916 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9920 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9921 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9925 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9926 utilities such as rsa, dsa, dsaparam etc except they process any key
9931 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9932 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9933 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9938 * Initial support for pluggable public key ASN1.
9939 De-spaghettify the public key ASN1 handling. Move public and private
9940 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9941 algorithm specific handling to a single module within the relevant
9942 algorithm directory. Add functions to allow (near) opaque processing
9943 of public and private key structures.
9947 * Implement the Supported Point Formats Extension for
9948 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9952 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9953 for the psk identity [hint] and the psk callback functions to the
9954 SSL_SESSION, SSL and SSL_CTX structure.
9957 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9961 SSL_CTX_use_psk_identity_hint
9962 SSL_get_psk_identity_hint
9963 SSL_get_psk_identity
9964 SSL_use_psk_identity_hint
9966 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9968 * Add RFC 3161 compliant time stamp request creation, response generation
9969 and response verification functionality.
9971 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9973 * Add initial support for TLS extensions, specifically for the server_name
9974 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9975 have new members for a hostname. The SSL data structure has an
9976 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9977 stored in that context to allow for session resumption, even after the
9978 SSL has been switched to a new SSL_CTX in reaction to a client's
9979 server_name extension.
9981 New functions (subject to change):
9983 SSL_get_servername()
9984 SSL_get_servername_type()
9987 New CTRL codes and macros (subject to change):
9989 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9990 - SSL_CTX_set_tlsext_servername_callback()
9991 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9992 - SSL_CTX_set_tlsext_servername_arg()
9993 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9995 openssl s_client has a new '-servername ...' option.
9997 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9998 '-key2 ...', '-servername_fatal' (subject to change). This allows
9999 testing the HostName extension for a specific single hostname ('-cert'
10000 and '-key' remain fallbacks for handshakes without HostName
10001 negotiation). If the unrecognized_name alert has to be sent, this by
10002 default is a warning; it becomes fatal with the '-servername_fatal'
10005 *Peter Sylvester, Remy Allais, Christophe Renou*
10007 * Whirlpool hash implementation is added.
10011 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10012 bn(64,32). Because of instruction set limitations it doesn't have
10013 any negative impact on performance. This was done mostly in order
10014 to make it possible to share assembler modules, such as bn_mul_mont
10015 implementations, between 32- and 64-bit builds without hassle.
10019 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10020 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10025 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10026 dedicated Montgomery multiplication procedure, is introduced.
10027 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10028 "64-bit" performance on certain 32-bit targets.
10032 * New option SSL_OP_NO_COMP to disable use of compression selectively
10033 in SSL structures. New SSL ctrl to set maximum send fragment size.
10034 Save memory by setting the I/O buffer sizes dynamically instead of
10035 using the maximum available value.
10039 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10040 in addition to the text details.
10044 * Very, very preliminary EXPERIMENTAL support for printing of general
10045 ASN1 structures. This currently produces rather ugly output and doesn't
10046 handle several customised structures at all.
10050 * Integrated support for PVK file format and some related formats such
10051 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10052 these in the 'rsa' and 'dsa' utilities.
10056 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10060 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10061 place for the (very old) "NETSCAPE" format certificates which are now
10062 handled using new ASN1 code equivalents.
10066 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10067 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10068 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10072 * Modify CRL distribution points extension code to print out previously
10073 unsupported fields. Enhance extension setting code to allow setting of
10078 * Add print and set support for Issuing Distribution Point CRL extension.
10082 * Change 'Configure' script to enable Camellia by default.
10089 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10091 * When rejecting SSL/TLS records due to an incorrect version number, never
10092 update s->server with a new major version number. As of
10093 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10094 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10095 the previous behavior could result in a read attempt at NULL when
10096 receiving specific incorrect SSL/TLS records once record payload
10097 protection is active. ([CVE-2010-0740])
10099 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10101 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10102 could be crashed if the relevant tables were not present (e.g. chrooted).
10104 *Tomas Hoger <thoger@redhat.com>*
10106 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10108 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10110 *Martin Olsson, Neel Mehta*
10112 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10113 accommodate for stack sorting, always a write lock!).
10117 * On some versions of WIN32 Heap32Next is very slow. This can cause
10118 excessive delays in the RAND_poll(): over a minute. As a workaround
10119 include a time check in the inner Heap32Next loop too.
10123 * The code that handled flushing of data in SSL/TLS originally used the
10124 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10125 the problem outlined in PR#1949. The fix suggested there however can
10126 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10127 of Apache). So instead simplify the code to flush unconditionally.
10128 This should be fine since flushing with no data to flush is a no op.
10132 * Handle TLS versions 2.0 and later properly and correctly use the
10133 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10134 off ancient servers have a habit of sticking around for a while...
10138 * Modify compression code so it frees up structures without using the
10139 ex_data callbacks. This works around a problem where some applications
10140 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10141 restarting) then use compression (e.g. SSL with compression) later.
10142 This results in significant per-connection memory leaks and
10143 has caused some security issues including CVE-2008-1678 and
10148 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10149 change when encrypting or decrypting.
10153 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10154 connect and renegotiate with servers which do not support RI.
10155 Until RI is more widely deployed this option is enabled by default.
10159 * Add "missing" ssl ctrls to clear options and mode.
10163 * If client attempts to renegotiate and doesn't support RI respond with
10164 a no_renegotiation alert as required by RFC5746. Some renegotiating
10165 TLS clients will continue a connection gracefully when they receive
10166 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10167 waiting for a server hello which it will never receive. Now we treat a
10168 received no_renegotiation alert as a fatal error. This is because
10169 applications requesting a renegotiation might well expect it to succeed
10170 and would have no code in place to handle the server denying it so the
10171 only safe thing to do is to terminate the connection.
10175 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10176 peer supports secure renegotiation and 0 otherwise. Print out peer
10177 renegotiation support in s_client/s_server.
10181 * Replace the highly broken and deprecated SPKAC certification method with
10182 the updated NID creation version. This should correctly handle UTF8.
10186 * Implement RFC5746. Re-enable renegotiation but require the extension
10187 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10188 turns out to be a bad idea. It has been replaced by
10189 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10190 SSL_CTX_set_options(). This is really not recommended unless you
10191 know what you are doing.
10193 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10195 * Fixes to stateless session resumption handling. Use initial_ctx when
10196 issuing and attempting to decrypt tickets in case it has changed during
10197 servername handling. Use a non-zero length session ID when attempting
10198 stateless session resumption: this makes it possible to determine if
10199 a resumption has occurred immediately after receiving server hello
10200 (several places in OpenSSL subtly assume this) instead of later in
10205 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10206 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10207 fixes for a few places where the return code is not checked
10210 *Julia Lawall <julia@diku.dk>*
10212 * Add --strict-warnings option to Configure script to include devteam
10213 warnings in other configurations.
10217 * Add support for --libdir option and LIBDIR variable in makefiles. This
10218 makes it possible to install openssl libraries in locations which
10219 have names other than "lib", for example "/usr/lib64" which some
10222 *Steve Henson, based on patch from Jeremy Utley*
10224 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10225 X690 8.9.12 and can produce some misleading textual output of OIDs.
10227 *Steve Henson, reported by Dan Kaminsky*
10229 * Delete MD2 from algorithm tables. This follows the recommendation in
10230 several standards that it is not used in new applications due to
10231 several cryptographic weaknesses. For binary compatibility reasons
10232 the MD2 API is still compiled in by default.
10236 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10241 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10242 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10245 *Guenter <lists@gknw.net>*
10247 * Fix the server certificate chain building code to use X509_verify_cert(),
10248 it used to have an ad-hoc builder which was unable to cope with anything
10249 other than a simple chain.
10251 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10253 * Don't check self signed certificate signatures in X509_verify_cert()
10254 by default (a flag can override this): it just wastes time without
10255 adding any security. As a useful side effect self signed root CAs
10256 with non-FIPS digests are now usable in FIPS mode.
10260 * In dtls1_process_out_of_seq_message() the check if the current message
10261 is already buffered was missing. For every new message was memory
10262 allocated, allowing an attacker to perform an denial of service attack
10263 with sending out of seq handshake messages until there is no memory
10264 left. Additionally every future message was buffered, even if the
10265 sequence number made no sense and would be part of another handshake.
10266 So only messages with sequence numbers less than 10 in advance will be
10267 buffered. ([CVE-2009-1378])
10269 *Robin Seggelmann, discovered by Daniel Mentz*
10271 * Records are buffered if they arrive with a future epoch to be
10272 processed after finishing the corresponding handshake. There is
10273 currently no limitation to this buffer allowing an attacker to perform
10274 a DOS attack with sending records with future epochs until there is no
10275 memory left. This patch adds the pqueue_size() function to determine
10276 the size of a buffer and limits the record buffer to 100 entries.
10279 *Robin Seggelmann, discovered by Daniel Mentz*
10281 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10282 parent structure is freed. ([CVE-2009-1379])
10286 * Handle non-blocking I/O properly in SSL_shutdown() call.
10288 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10290 * Add `2.5.4.*` OIDs
10292 *Ilya O. <vrghost@gmail.com>*
10294 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10296 * Disable renegotiation completely - this fixes a severe security
10297 problem ([CVE-2009-3555]) at the cost of breaking all
10298 renegotiation. Renegotiation can be re-enabled by setting
10299 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10300 run-time. This is really not recommended unless you know what
10305 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10307 * Don't set val to NULL when freeing up structures, it is freed up by
10308 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10309 zeroing past the valid field. ([CVE-2009-0789])
10311 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10313 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10314 checked correctly. This would allow some invalid signed attributes to
10315 appear to verify correctly. ([CVE-2009-0591])
10317 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10319 * Reject UniversalString and BMPString types with invalid lengths. This
10320 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10321 a legal length. ([CVE-2009-0590])
10325 * Set S/MIME signing as the default purpose rather than setting it
10326 unconditionally. This allows applications to override it at the store
10331 * Permit restricted recursion of ASN1 strings. This is needed in practice
10332 to handle some structures.
10336 * Improve efficiency of mem_gets: don't search whole buffer each time
10339 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10341 * New -hex option for openssl rand.
10345 * Print out UTF8String and NumericString when parsing ASN1.
10349 * Support NumericString type for name components.
10353 * Allow CC in the environment to override the automatically chosen
10354 compiler. Note that nothing is done to ensure flags work with the
10359 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10361 * Properly check EVP_VerifyFinal() and similar return values
10364 *Ben Laurie, Bodo Moeller, Google Security Team*
10366 * Enable TLS extensions by default.
10370 * Allow the CHIL engine to be loaded, whether the application is
10371 multithreaded or not. (This does not release the developer from the
10372 obligation to set up the dynamic locking callbacks.)
10374 *Sander Temme <sander@temme.net>*
10376 * Use correct exit code if there is an error in dgst command.
10378 *Steve Henson; problem pointed out by Roland Dirlewanger*
10380 * Tweak Configure so that you need to say "experimental-jpake" to enable
10381 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10385 * Add experimental JPAKE support, including demo authentication in
10386 s_client and s_server.
10390 * Set the comparison function in v3_addr_canonize().
10392 *Rob Austein <sra@hactrn.net>*
10394 * Add support for XMPP STARTTLS in s_client.
10396 *Philip Paeps <philip@freebsd.org>*
10398 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10399 to ensure that even with this option, only ciphersuites in the
10400 server's preference list will be accepted. (Note that the option
10401 applies only when resuming a session, so the earlier behavior was
10402 just about the algorithm choice for symmetric cryptography.)
10406 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10408 * Fix NULL pointer dereference if a DTLS server received
10409 ChangeCipherSpec as first record ([CVE-2009-1386]).
10413 * Fix a state transition in s3_srvr.c and d1_srvr.c
10414 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10416 *Nagendra Modadugu*
10418 * The fix in 0.9.8c that supposedly got rid of unsafe
10419 double-checked locking was incomplete for RSA blinding,
10420 addressing just one layer of what turns out to have been
10421 doubly unsafe triple-checked locking.
10423 So now fix this for real by retiring the MONT_HELPER macro
10424 in crypto/rsa/rsa_eay.c.
10426 *Bodo Moeller; problem pointed out by Marius Schilder*
10428 * Various precautionary measures:
10430 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10432 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10433 (NB: This would require knowledge of the secret session ticket key
10434 to exploit, in which case you'd be SOL either way.)
10436 - Change bn_nist.c so that it will properly handle input BIGNUMs
10437 outside the expected range.
10439 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10442 *Neel Mehta, Bodo Moeller*
10444 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10445 the load fails. Useful for distros.
10447 *Ben Laurie and the FreeBSD team*
10449 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10453 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10457 * Expand ENGINE to support engine supplied SSL client certificate functions.
10459 This work was sponsored by Logica.
10463 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10464 keystores. Support for SSL/TLS client authentication too.
10465 Not compiled unless enable-capieng specified to Configure.
10467 This work was sponsored by Logica.
10471 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10472 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10473 attribute creation routines such as certificate requests and PKCS#12
10478 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10480 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10481 handshake which could lead to a client crash as found using the
10482 Codenomicon TLS test suite ([CVE-2008-1672])
10484 *Steve Henson, Mark Cox*
10486 * Fix double free in TLS server name extensions which could lead to
10487 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10491 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10493 Clear the error queue to ensure that error entries left from
10494 older function calls do not interfere with the correct operation.
10496 *Lutz Jaenicke, Erik de Castro Lopo*
10498 * Remove root CA certificates of commercial CAs:
10500 The OpenSSL project does not recommend any specific CA and does not
10501 have any policy with respect to including or excluding any CA.
10502 Therefore, it does not make any sense to ship an arbitrary selection
10503 of root CA certificates with the OpenSSL software.
10507 * RSA OAEP patches to fix two separate invalid memory reads.
10508 The first one involves inputs when 'lzero' is greater than
10509 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10510 before the beginning of from). The second one involves inputs where
10511 the 'db' section contains nothing but zeroes (there is a one-byte
10512 invalid read after the end of 'db').
10514 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10516 * Partial backport from 0.9.9-dev:
10518 Introduce bn_mul_mont (dedicated Montgomery multiplication
10519 procedure) as a candidate for BIGNUM assembler implementation.
10520 While 0.9.9-dev uses assembler for various architectures, only
10521 x86_64 is available by default here in the 0.9.8 branch, and
10522 32-bit x86 is available through a compile-time setting.
10524 To try the 32-bit x86 assembler implementation, use Configure
10525 option "enable-montasm" (which exists only for this backport).
10527 As "enable-montasm" for 32-bit x86 disclaims code stability
10528 anyway, in this constellation we activate additional code
10529 backported from 0.9.9-dev for further performance improvements,
10530 namely BN_from_montgomery_word. (To enable this otherwise,
10531 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10533 *Andy Polyakov (backport partially by Bodo Moeller)*
10535 * Add TLS session ticket callback. This allows an application to set
10536 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10537 values. This is useful for key rollover for example where several key
10538 sets may exist with different names.
10542 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10543 This was broken until now in 0.9.8 releases, such that the only way
10544 a registered ENGINE could be used (assuming it initialises
10545 successfully on the host) was to explicitly set it as the default
10546 for the relevant algorithms. This is in contradiction with 0.9.7
10547 behaviour and the documentation. With this fix, when an ENGINE is
10548 registered into a given algorithm's table of implementations, the
10549 'uptodate' flag is reset so that auto-discovery will be used next
10550 time a new context for that algorithm attempts to select an
10553 *Ian Lister (tweaked by Geoff Thorpe)*
10555 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10556 implementation in the following ways:
10558 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10561 Lack of BER streaming support means one pass streaming processing is
10562 only supported if data is detached: setting the streaming flag is
10563 ignored for embedded content.
10565 CMS support is disabled by default and must be explicitly enabled
10566 with the enable-cms configuration option.
10570 * Update the GMP engine glue to do direct copies between BIGNUM and
10571 mpz_t when openssl and GMP use the same limb size. Otherwise the
10572 existing "conversion via a text string export" trick is still used.
10574 *Paul Sheer <paulsheer@gmail.com>*
10576 * Zlib compression BIO. This is a filter BIO which compressed and
10577 uncompresses any data passed through it.
10581 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10582 RFC3394 compatible AES key wrapping.
10586 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10587 sets string data without copying. X509_ALGOR_set0() and
10588 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10589 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10590 from an X509_ATTRIBUTE structure optionally checking it occurs only
10591 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10596 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10597 to get the expected BN_FLG_CONSTTIME behavior.
10599 *Bodo Moeller (Google)*
10603 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10604 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10605 - added some more tests to do_tests.pl
10606 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10607 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10608 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10609 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10610 - various changes to netware.pl to enable gcc-cross builds on Win32
10612 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10613 - various changes to fix missing prototype warnings
10614 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10615 - added AES, WHIRLPOOL and CPUID assembler code to build files
10616 - added missing AES assembler make rules to mk1mf.pl
10617 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10619 *Guenter Knauf <eflash@gmx.net>*
10621 * Implement certificate status request TLS extension defined in RFC3546.
10622 A client can set the appropriate parameters and receive the encoded
10623 OCSP response via a callback. A server can query the supplied parameters
10624 and set the encoded OCSP response in the callback. Add simplified examples
10625 to s_client and s_server.
10629 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10631 * Fix various bugs:
10632 + Binary incompatibility of ssl_ctx_st structure
10633 + DTLS interoperation with non-compliant servers
10634 + Don't call get_session_cb() without proposed session
10635 + Fix ia64 assembler code
10637 *Andy Polyakov, Steve Henson*
10639 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10641 * DTLS Handshake overhaul. There were longstanding issues with
10642 OpenSSL DTLS implementation, which were making it impossible for
10643 RFC 4347 compliant client to communicate with OpenSSL server.
10644 Unfortunately just fixing these incompatibilities would "cut off"
10645 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10646 server keeps tolerating non RFC compliant syntax. The opposite is
10647 not true, 0.9.8f client can not communicate with earlier server.
10648 This update even addresses CVE-2007-4995.
10652 * Changes to avoid need for function casts in OpenSSL: some compilers
10653 (gcc 4.2 and later) reject their use.
10654 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10657 * Add RFC4507 support to OpenSSL. This includes the corrections in
10658 RFC4507bis. The encrypted ticket format is an encrypted encoded
10659 SSL_SESSION structure, that way new session features are automatically
10662 If a client application caches session in an SSL_SESSION structure
10663 support is transparent because tickets are now stored in the encoded
10666 The SSL_CTX structure automatically generates keys for ticket
10667 protection in servers so again support should be possible
10668 with no application modification.
10670 If a client or server wishes to disable RFC4507 support then the option
10671 SSL_OP_NO_TICKET can be set.
10673 Add a TLS extension debugging callback to allow the contents of any client
10674 or server extensions to be examined.
10676 This work was sponsored by Google.
10680 * Add initial support for TLS extensions, specifically for the server_name
10681 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10682 have new members for a hostname. The SSL data structure has an
10683 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10684 stored in that context to allow for session resumption, even after the
10685 SSL has been switched to a new SSL_CTX in reaction to a client's
10686 server_name extension.
10688 New functions (subject to change):
10690 SSL_get_servername()
10691 SSL_get_servername_type()
10694 New CTRL codes and macros (subject to change):
10696 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10697 - SSL_CTX_set_tlsext_servername_callback()
10698 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10699 - SSL_CTX_set_tlsext_servername_arg()
10700 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10702 openssl s_client has a new '-servername ...' option.
10704 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10705 '-key2 ...', '-servername_fatal' (subject to change). This allows
10706 testing the HostName extension for a specific single hostname ('-cert'
10707 and '-key' remain fallbacks for handshakes without HostName
10708 negotiation). If the unrecognized_name alert has to be sent, this by
10709 default is a warning; it becomes fatal with the '-servername_fatal'
10712 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10714 * Add AES and SSE2 assembly language support to VC++ build.
10718 * Mitigate attack on final subtraction in Montgomery reduction.
10722 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10723 (which previously caused an internal error).
10727 * Squeeze another 10% out of IGE mode when in != out.
10731 * AES IGE mode speedup.
10733 *Dean Gaudet (Google)*
10735 * Add the Korean symmetric 128-bit cipher SEED (see
10736 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10737 add SEED ciphersuites from RFC 4162:
10739 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10740 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10741 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10742 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10744 To minimize changes between patchlevels in the OpenSSL 0.9.8
10745 series, SEED remains excluded from compilation unless OpenSSL
10746 is configured with 'enable-seed'.
10748 *KISA, Bodo Moeller*
10750 * Mitigate branch prediction attacks, which can be practical if a
10751 single processor is shared, allowing a spy process to extract
10752 information. For detailed background information, see
10753 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10754 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10755 and Necessary Software Countermeasures"). The core of the change
10756 are new versions BN_div_no_branch() and
10757 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10758 respectively, which are slower, but avoid the security-relevant
10759 conditional branches. These are automatically called by BN_div()
10760 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10761 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10762 remove a conditional branch.
10764 BN_FLG_CONSTTIME is the new name for the previous
10765 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10766 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10767 in the exponent causes BN_mod_exp_mont() to use the alternative
10768 implementation in BN_mod_exp_mont_consttime().) The old name
10769 remains as a deprecated alias.
10771 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10772 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10773 constant-time implementations for more than just exponentiation.
10774 Here too the old name is kept as a deprecated alias.
10776 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10777 the BN_BLINDING structure gets an independent copy of the
10778 modulus. This means that the previous `BIGNUM *m` argument to
10779 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10780 essentially becomes `const BIGNUM *m`, although we can't actually
10781 change this in the header file before 0.9.9. It allows
10782 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10783 enable BN_FLG_CONSTTIME.
10785 *Matthew D Wood (Intel Corp)*
10787 * In the SSL/TLS server implementation, be strict about session ID
10788 context matching (which matters if an application uses a single
10789 external cache for different purposes). Previously,
10790 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10791 set. This did ensure strict client verification, but meant that,
10792 with applications using a single external cache for quite
10793 different requirements, clients could circumvent ciphersuite
10794 restrictions for a given session ID context by starting a session
10795 in a different context.
10799 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10800 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10801 authentication-only ciphersuites.
10805 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10806 not complete and could lead to a possible single byte overflow
10807 ([CVE-2007-5135]) [Ben Laurie]
10809 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10811 * Since AES128 and AES256 (and similarly Camellia128 and
10812 Camellia256) share a single mask bit in the logic of
10813 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10814 kludge to work properly if AES128 is available and AES256 isn't
10815 (or if Camellia128 is available and Camellia256 isn't).
10819 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10820 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10821 When a point or a seed is encoded in a BIT STRING, we need to
10822 prevent the removal of trailing zero bits to get the proper DER
10823 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10824 of a NamedBitList, for which trailing 0 bits need to be removed.)
10828 * Have SSL/TLS server implementation tolerate "mismatched" record
10829 protocol version while receiving ClientHello even if the
10830 ClientHello is fragmented. (The server can't insist on the
10831 particular protocol version it has chosen before the ServerHello
10832 message has informed the client about his choice.)
10836 * Add RFC 3779 support.
10838 *Rob Austein for ARIN, Ben Laurie*
10840 * Load error codes if they are not already present instead of using a
10841 static variable. This allows them to be cleanly unloaded and reloaded.
10842 Improve header file function name parsing.
10846 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10847 or CAPABILITY handshake as required by RFCs.
10849 *Goetz Babin-Ebell*
10851 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10853 * Introduce limits to prevent malicious keys being able to
10854 cause a denial of service. ([CVE-2006-2940])
10856 *Steve Henson, Bodo Moeller*
10858 * Fix ASN.1 parsing of certain invalid structures that can result
10859 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10861 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10862 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10864 * Fix SSL client code which could crash if connecting to a
10865 malicious SSLv2 server. ([CVE-2006-4343])
10867 *Tavis Ormandy and Will Drewry, Google Security Team*
10869 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10870 match only those. Before that, "AES256-SHA" would be interpreted
10871 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10872 the same strength classification in 0.9.7h) as we currently only
10873 have a single AES bit in the ciphersuite description bitmap.
10874 That change, however, also applied to ciphersuite strings such as
10875 "RC4-MD5" that intentionally matched multiple ciphersuites --
10876 namely, SSL 2.0 ciphersuites in addition to the more common ones
10877 from SSL 3.0/TLS 1.0.
10879 So we change the selection algorithm again: Naming an explicit
10880 ciphersuite selects this one ciphersuite, and any other similar
10881 ciphersuite (same bitmap) from *other* protocol versions.
10882 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10883 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10885 Since SSL 2.0 does not have any ciphersuites for which the
10886 128/256 bit distinction would be relevant, this works for now.
10887 The proper fix will be to use different bits for AES128 and
10888 AES256, which would have avoided the problems from the beginning;
10889 however, bits are scarce, so we can only do this in a new release
10890 (not just a patchlevel) when we can change the SSL_CIPHER
10891 definition to split the single 'unsigned long mask' bitmap into
10892 multiple values to extend the available space.
10896 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10898 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10899 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10901 * Add AES IGE and biIGE modes.
10905 * Change the Unix randomness entropy gathering to use poll() when
10906 possible instead of select(), since the latter has some
10907 undesirable limitations.
10909 *Darryl Miles via Richard Levitte and Bodo Moeller*
10911 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10912 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10913 cannot be implicitly activated as part of, e.g., the "AES" alias.
10914 However, please upgrade to OpenSSL 0.9.9[-dev] for
10915 non-experimental use of the ECC ciphersuites to get TLS extension
10916 support, which is required for curve and point format negotiation
10917 to avoid potential handshake problems.
10921 * Disable rogue ciphersuites:
10923 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10924 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10925 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10927 The latter two were purportedly from
10928 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10931 Also deactivate the remaining ciphersuites from
10932 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10933 unofficial, and the ID has long expired.
10937 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10938 dual-core machines) and other potential thread-safety issues.
10942 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10943 versions), which is now available for royalty-free use
10944 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10945 Also, add Camellia TLS ciphersuites from RFC 4132.
10947 To minimize changes between patchlevels in the OpenSSL 0.9.8
10948 series, Camellia remains excluded from compilation unless OpenSSL
10949 is configured with 'enable-camellia'.
10953 * Disable the padding bug check when compression is in use. The padding
10954 bug check assumes the first packet is of even length, this is not
10955 necessarily true if compression is enabled and can result in false
10956 positives causing handshake failure. The actual bug test is ancient
10957 code so it is hoped that implementations will either have fixed it by
10958 now or any which still have the bug do not support compression.
10962 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10964 * When applying a cipher rule check to see if string match is an explicit
10965 cipher suite and only match that one cipher suite if it is.
10969 * Link in manifests for VC++ if needed.
10971 *Austin Ziegler <halostatue@gmail.com>*
10973 * Update support for ECC-based TLS ciphersuites according to
10974 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10975 TLS extensions, which are supported starting with the 0.9.9
10976 branch, not in the OpenSSL 0.9.8 branch).
10980 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10981 opaque EVP_CIPHER_CTX handling.
10985 * Fixes and enhancements to zlib compression code. We now only use
10986 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10987 to conform with the standards mentioned here:
10988 <http://www.zlib.net/DLL_FAQ.txt>
10989 Static zlib linking now works on Windows and the new --with-zlib-include
10990 --with-zlib-lib options to Configure can be used to supply the location
10991 of the headers and library. Gracefully handle case where zlib library
10996 * Several fixes and enhancements to the OID generation code. The old code
10997 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10998 handle numbers larger than ULONG_MAX, truncated printing and had a
10999 non standard OBJ_obj2txt() behaviour.
11003 * Add support for building of engines under engine/ as shared libraries
11004 under VC++ build system.
11008 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11009 Hopefully, we will not see any false combination of paths any more.
11013 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11015 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11016 (part of SSL_OP_ALL). This option used to disable the
11017 countermeasure against man-in-the-middle protocol-version
11018 rollback in the SSL 2.0 server implementation, which is a bad
11019 idea. ([CVE-2005-2969])
11021 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11022 for Information Security, National Institute of Advanced Industrial
11023 Science and Technology [AIST], Japan)*
11025 * Add two function to clear and return the verify parameter flags.
11029 * Keep cipherlists sorted in the source instead of sorting them at
11030 runtime, thus removing the need for a lock.
11034 * Avoid some small subgroup attacks in Diffie-Hellman.
11036 *Nick Mathewson and Ben Laurie*
11038 * Add functions for well-known primes.
11042 * Extended Windows CE support.
11044 *Satoshi Nakamura and Andy Polyakov*
11046 * Initialize SSL_METHOD structures at compile time instead of during
11047 runtime, thus removing the need for a lock.
11051 * Make PKCS7_decrypt() work even if no certificate is supplied by
11052 attempting to decrypt each encrypted key in turn. Add support to
11057 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11059 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11062 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11066 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11067 key into the same file any more.
11071 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11075 * Add -utf8 command line and config file option to 'ca'.
11077 *Stefan <stf@udoma.org*
11079 * Removed the macro des_crypt(), as it seems to conflict with some
11080 libraries. Use DES_crypt().
11084 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11085 involves renaming the source and generated shared-libs for
11086 both. The engines will accept the corrected or legacy ids
11087 ('ncipher' and '4758_cca' respectively) when binding. NB,
11088 this only applies when building 'shared'.
11090 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11092 * Add attribute functions to EVP_PKEY structure. Modify
11093 PKCS12_create() to recognize a CSP name attribute and
11094 use it. Make -CSP option work again in pkcs12 utility.
11098 * Add new functionality to the bn blinding code:
11099 - automatic re-creation of the BN_BLINDING parameters after
11100 a fixed number of uses (currently 32)
11101 - add new function for parameter creation
11102 - introduce flags to control the update behaviour of the
11103 BN_BLINDING parameters
11104 - hide BN_BLINDING structure
11105 Add a second BN_BLINDING slot to the RSA structure to improve
11106 performance when a single RSA object is shared among several
11111 * Add support for DTLS.
11113 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11115 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11116 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11120 * Remove buggy and incomplete DH cert support from
11121 ssl/ssl_rsa.c and ssl/s3_both.c
11125 * Use SHA-1 instead of MD5 as the default digest algorithm for
11126 the `apps/openssl` commands.
11130 * Compile clean with "-Wall -Wmissing-prototypes
11131 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11132 DEBUG_SAFESTACK must also be set.
11136 * Change ./Configure so that certain algorithms can be disabled by default.
11137 The new counterpiece to "no-xxx" is "enable-xxx".
11139 The patented RC5 and MDC2 algorithms will now be disabled unless
11140 "enable-rc5" and "enable-mdc2", respectively, are specified.
11142 (IDEA remains enabled despite being patented. This is because IDEA
11143 is frequently required for interoperability, and there is no license
11144 fee for non-commercial use. As before, "no-idea" can be used to
11145 avoid this algorithm.)
11149 * Add processing of proxy certificates (see RFC 3820). This work was
11150 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11151 EGEE (Enabling Grids for E-science in Europe).
11155 * RC4 performance overhaul on modern architectures/implementations, such
11156 as Intel P4, IA-64 and AMD64.
11160 * New utility extract-section.pl. This can be used specify an alternative
11161 section number in a pod file instead of having to treat each file as
11162 a separate case in Makefile. This can be done by adding two lines to the
11165 =for comment openssl_section:XXX
11167 The blank line is mandatory.
11171 * New arguments -certform, -keyform and -pass for s_client and s_server
11172 to allow alternative format key and certificate files and passphrase
11177 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11178 update associated structures and add various utility functions.
11180 Add new policy related verify parameters, include policy checking in
11181 standard verify code. Enhance 'smime' application with extra parameters
11182 to support policy checking and print out.
11186 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11187 Nehemiah processors. These extensions support AES encryption in hardware
11188 as well as RNG (though RNG support is currently disabled).
11190 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11192 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11196 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11198 *Andy Polyakov and a number of other people*
11200 * Improved PowerPC platform support. Most notably BIGNUM assembler
11201 implementation contributed by IBM.
11203 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11205 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11206 exponent rather than 'unsigned long'. There is a corresponding change to
11207 the new 'rsa_keygen' element of the RSA_METHOD structure.
11209 *Jelte Jansen, Geoff Thorpe*
11211 * Functionality for creating the initial serial number file is now
11212 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11214 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11215 number file to 1, which is bound to cause problems. To avoid
11216 the problems while respecting compatibility between different 0.9.7
11217 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11218 CA.pl for serial number initialization. With the new release 0.9.8,
11219 we can fix the problem directly in the 'ca' utility.)
11223 * Reduced header interdependencies by declaring more opaque objects in
11224 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11225 give fewer recursive includes, which could break lazy source code - so
11226 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11227 developers should define this symbol when building and using openssl to
11228 ensure they track the recommended behaviour, interfaces, [etc], but
11229 backwards-compatible behaviour prevails when this isn't defined.
11233 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11237 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11238 This will generate a random key of the appropriate length based on the
11239 cipher context. The EVP_CIPHER can provide its own random key generation
11240 routine to support keys of a specific form. This is used in the des and
11241 3des routines to generate a key of the correct parity. Update S/MIME
11242 code to use new functions and hence generate correct parity DES keys.
11243 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11244 valid (weak or incorrect parity).
11248 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11249 as looking them up. This is useful when the verified structure may contain
11250 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11251 present unless the new PKCS7_NO_CRL flag is asserted.
11255 * Extend ASN1 oid configuration module. It now additionally accepts the
11258 shortName = some long name, 1.2.3.4
11262 * Reimplemented the BN_CTX implementation. There is now no more static
11263 limitation on the number of variables it can handle nor the depth of the
11264 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11265 information can now expand as required, and rather than having a single
11266 static array of bignums, BN_CTX now uses a linked-list of such arrays
11267 allowing it to expand on demand whilst maintaining the usefulness of
11268 BN_CTX's "bundling".
11272 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11273 to allow all RSA operations to function using a single BN_CTX.
11277 * Preliminary support for certificate policy evaluation and checking. This
11278 is initially intended to pass the tests outlined in "Conformance Testing
11279 of Relying Party Client Certificate Path Processing Logic" v1.07.
11283 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11284 remained unused and not that useful. A variety of other little bignum
11285 tweaks and fixes have also been made continuing on from the audit (see
11290 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11291 associated ASN1, EVP and SSL functions and old ASN1 macros.
11295 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11296 and this should never fail. So the return value from the use of
11297 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11298 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11302 * BN_CTX_get() should return zero-valued bignums, providing the same
11303 initialised value as BN_new().
11305 *Geoff Thorpe, suggested by Ulf Möller*
11307 * Support for inhibitAnyPolicy certificate extension.
11311 * An audit of the BIGNUM code is underway, for which debugging code is
11312 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11313 is considered valid when processing BIGNUMs, and causes execution to
11314 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11315 further steps are taken to deliberately pollute unused data in BIGNUM
11316 structures to try and expose faulty code further on. For now, openssl will
11317 (in its default mode of operation) continue to tolerate the inconsistent
11318 forms that it has tolerated in the past, but authors and packagers should
11319 consider trying openssl and their own applications when compiled with
11320 these debugging symbols defined. It will help highlight potential bugs in
11321 their own code, and will improve the test coverage for OpenSSL itself. At
11322 some point, these tighter rules will become openssl's default to improve
11323 maintainability, though the assert()s and other overheads will remain only
11324 in debugging configurations. See bn.h for more details.
11326 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11328 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11329 that can only be obtained through BN_CTX_new() (which implicitly
11330 initialises it). The presence of this function only made it possible
11331 to overwrite an existing structure (and cause memory leaks).
11335 * Because of the callback-based approach for implementing LHASH as a
11336 template type, lh_insert() adds opaque objects to hash-tables and
11337 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11338 to clean up those corresponding objects before destroying the hash table
11339 (and losing the object pointers). So some over-zealous constifications in
11340 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11341 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11342 prototyped to have "const" restrictions on the object pointers they are
11343 given (and so aren't required to cast them away any more).
11347 * The tmdiff.h API was so ugly and minimal that our own timing utility
11348 (speed) prefers to use its own implementation. The two implementations
11349 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11350 its object type properly exposed (MS_TM) instead of casting to/from
11351 `char *`. This may still change yet if someone realises MS_TM and
11353 aren't necessarily the greatest nomenclatures - but this is what was used
11354 internally to the implementation so I've used that for now.
11358 * Ensure that deprecated functions do not get compiled when
11359 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11360 the self-tests were still using deprecated key-generation functions so
11361 these have been updated also.
11365 * Reorganise PKCS#7 code to separate the digest location functionality
11366 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11367 New function PKCS7_set_digest() to set the digest type for PKCS#7
11368 digestedData type. Add additional code to correctly generate the
11369 digestedData type and add support for this type in PKCS7 initialization
11374 * New function PKCS7_set0_type_other() this initializes a PKCS7
11375 structure of type "other".
11379 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11380 sure the loop does correctly stop and breaking ("division by zero")
11381 modulus operations are not performed. The (pre-generated) prime
11382 table crypto/bn/bn_prime.h was already correct, but it could not be
11383 re-generated on some platforms because of the "division by zero"
11384 situation in the script.
11386 *Ralf S. Engelschall*
11388 * Update support for ECC-based TLS ciphersuites according to
11389 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11390 SHA-1 now is only used for "small" curves (where the
11391 representation of a field element takes up to 24 bytes); for
11392 larger curves, the field element resulting from ECDH is directly
11393 used as premaster secret.
11395 *Douglas Stebila (Sun Microsystems Laboratories)*
11397 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11398 curve secp160r1 to the tests.
11400 *Douglas Stebila (Sun Microsystems Laboratories)*
11402 * Add the possibility to load symbols globally with DSO.
11404 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11406 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11407 control of the error stack.
11411 * Add support for STORE in ENGINE.
11415 * Add the STORE type. The intention is to provide a common interface
11416 to certificate and key stores, be they simple file-based stores, or
11417 HSM-type store, or LDAP stores, or...
11418 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11422 * Add a generic structure called OPENSSL_ITEM. This can be used to
11423 pass a list of arguments to any function as well as provide a way
11424 for a function to pass data back to the caller.
11428 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11429 works like BUF_strdup() but can be used to duplicate a portion of
11430 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11435 * Add the function sk_find_ex() which works like sk_find(), but will
11436 return an index to an element even if an exact match couldn't be
11437 found. The index is guaranteed to point at the element where the
11438 searched-for key would be inserted to preserve sorting order.
11442 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11443 takes an extra flags argument for optional functionality. Currently,
11444 the following flags are defined:
11446 OBJ_BSEARCH_VALUE_ON_NOMATCH
11447 This one gets OBJ_bsearch_ex() to return a pointer to the first
11448 element where the comparing function returns a negative or zero
11451 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11452 This one gets OBJ_bsearch_ex() to return a pointer to the first
11453 element where the comparing function returns zero. This is useful
11454 if there are more than one element where the comparing function
11459 * Make it possible to create self-signed certificates with 'openssl ca'
11460 in such a way that the self-signed certificate becomes part of the
11461 CA database and uses the same mechanisms for serial number generation
11462 as all other certificate signing. The new flag '-selfsign' enables
11463 this functionality. Adapt CA.sh and CA.pl.in.
11467 * Add functionality to check the public key of a certificate request
11468 against a given private. This is useful to check that a certificate
11469 request can be signed by that key (self-signing).
11473 * Make it possible to have multiple active certificates with the same
11474 subject in the CA index file. This is done only if the keyword
11475 'unique_subject' is set to 'no' in the main CA section (default
11476 if 'CA_default') of the configuration file. The value is saved
11477 with the database itself in a separate index attribute file,
11478 named like the index file with '.attr' appended to the name.
11482 * Generate multi-valued AVAs using '+' notation in config files for
11487 * Support for nameConstraints certificate extension.
11491 * Support for policyConstraints certificate extension.
11495 * Support for policyMappings certificate extension.
11499 * Make sure the default DSA_METHOD implementation only uses its
11500 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11501 and change its own handlers to be NULL so as to remove unnecessary
11502 indirection. This lets alternative implementations fallback to the
11503 default implementation more easily.
11507 * Support for directoryName in GeneralName related extensions
11512 * Make it possible to link applications using Makefile.shared.
11513 Make that possible even when linking against static libraries!
11517 * Support for single pass processing for S/MIME signing. This now
11518 means that S/MIME signing can be done from a pipe, in addition
11519 cleartext signing (multipart/signed type) is effectively streaming
11520 and the signed data does not need to be all held in memory.
11522 This is done with a new flag PKCS7_STREAM. When this flag is set
11523 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11524 is done after the data is output (and digests calculated) in
11525 SMIME_write_PKCS7().
11529 * Add full support for -rpath/-R, both in shared libraries and
11530 applications, at least on the platforms where it's known how
11535 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11536 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11537 will now compute a table of multiples of the generator that
11538 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11539 faster (notably in the case of a single point multiplication,
11540 scalar * generator).
11542 *Nils Larsch, Bodo Moeller*
11544 * IPv6 support for certificate extensions. The various extensions
11545 which use the IP:a.b.c.d can now take IPv6 addresses using the
11546 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11551 * Added an ENGINE that implements RSA by performing private key
11552 exponentiations with the GMP library. The conversions to and from
11553 GMP's mpz_t format aren't optimised nor are any montgomery forms
11554 cached, and on x86 it appears OpenSSL's own performance has caught up.
11555 However there are likely to be other architectures where GMP could
11556 provide a boost. This ENGINE is not built in by default, but it can be
11557 specified at Configure time and should be accompanied by the necessary
11558 linker additions, eg;
11559 ./config -DOPENSSL_USE_GMP -lgmp
11563 * "openssl engine" will not display ENGINE/DSO load failure errors when
11564 testing availability of engines with "-t" - the old behaviour is
11565 produced by increasing the feature's verbosity with "-tt".
11569 * ECDSA routines: under certain error conditions uninitialized BN objects
11570 could be freed. Solution: make sure initialization is performed early
11571 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11576 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11577 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11578 software implementations. For DSA and DH, parameter generation can
11579 also be overridden by providing the appropriate method callbacks.
11583 * Change the "progress" mechanism used in key-generation and
11584 primality testing to functions that take a new BN_GENCB pointer in
11585 place of callback/argument pairs. The new API functions have `_ex`
11586 postfixes and the older functions are reimplemented as wrappers for
11587 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11588 declarations of the old functions to help (graceful) attempts to
11589 migrate to the new functions. Also, the new key-generation API
11590 functions operate on a caller-supplied key-structure and return
11591 success/failure rather than returning a key or NULL - this is to
11592 help make "keygen" another member function of RSA_METHOD etc.
11594 Example for using the new callback interface:
11596 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11597 void *my_arg = ...;
11600 BN_GENCB_set(&my_cb, my_callback, my_arg);
11602 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11603 /* For the meaning of a, b in calls to my_callback(), see the
11604 * documentation of the function that calls the callback.
11605 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11606 * my_callback should return 1 if it wants BN_is_prime_ex()
11607 * to continue, or 0 to stop.
11612 * Change the ZLIB compression method to be stateful, and make it
11613 available to TLS with the number defined in
11614 draft-ietf-tls-compression-04.txt.
11618 * Add the ASN.1 structures and functions for CertificatePair, which
11619 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11621 CertificatePair ::= SEQUENCE {
11622 forward [0] Certificate OPTIONAL,
11623 reverse [1] Certificate OPTIONAL,
11624 -- at least one of the pair shall be present -- }
11626 Also implement the PEM functions to read and write certificate
11627 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11629 This needed to be defined, mostly for the sake of the LDAP
11630 attribute crossCertificatePair, but may prove useful elsewhere as
11635 * Make it possible to inhibit symlinking of shared libraries in
11636 Makefile.shared, for Cygwin's sake.
11640 * Extend the BIGNUM API by creating a function
11641 void BN_set_negative(BIGNUM *a, int neg);
11642 and a macro that behave like
11643 int BN_is_negative(const BIGNUM *a);
11645 to avoid the need to access 'a->neg' directly in applications.
11649 * Implement fast modular reduction for pseudo-Mersenne primes
11650 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11651 EC_GROUP_new_curve_GFp() will now automatically use this
11654 *Nils Larsch <nla@trustcenter.de>*
11656 * Add new lock type (CRYPTO_LOCK_BN).
11660 * Change the ENGINE framework to automatically load engines
11661 dynamically from specific directories unless they could be
11662 found to already be built in or loaded. Move all the
11663 current engines except for the cryptodev one to a new
11664 directory engines/.
11665 The engines in engines/ are built as shared libraries if
11666 the "shared" options was given to ./Configure or ./config.
11667 Otherwise, they are inserted in libcrypto.a.
11668 /usr/local/ssl/engines is the default directory for dynamic
11669 engines, but that can be overridden at configure time through
11670 the usual use of --prefix and/or --openssldir, and at run
11671 time with the environment variable OPENSSL_ENGINES.
11673 *Geoff Thorpe and Richard Levitte*
11675 * Add Makefile.shared, a helper makefile to build shared
11676 libraries. Adapt Makefile.org.
11680 * Add version info to Win32 DLLs.
11682 *Peter 'Luna' Runestig" <peter@runestig.com>*
11684 * Add new 'medium level' PKCS#12 API. Certificates and keys
11685 can be added using this API to created arbitrary PKCS#12
11686 files while avoiding the low-level API.
11688 New options to PKCS12_create(), key or cert can be NULL and
11689 will then be omitted from the output file. The encryption
11690 algorithm NIDs can be set to -1 for no encryption, the mac
11691 iteration count can be set to 0 to omit the mac.
11693 Enhance pkcs12 utility by making the -nokeys and -nocerts
11694 options work when creating a PKCS#12 file. New option -nomac
11695 to omit the mac, NONE can be set for an encryption algorithm.
11696 New code is modified to use the enhanced PKCS12_create()
11697 instead of the low-level API.
11701 * Extend ASN1 encoder to support indefinite length constructed
11702 encoding. This can output sequences tags and octet strings in
11703 this form. Modify pk7_asn1.c to support indefinite length
11704 encoding. This is experimental and needs additional code to
11705 be useful, such as an ASN1 bio and some enhanced streaming
11708 Extend template encode functionality so that tagging is passed
11709 down to the template encoder.
11713 * Let 'openssl req' fail if an argument to '-newkey' is not
11714 recognized instead of using RSA as a default.
11718 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11719 As these are not official, they are not included in "ALL";
11720 the "ECCdraft" ciphersuite group alias can be used to select them.
11722 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11724 * Add ECDH engine support.
11726 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11728 * Add ECDH in new directory crypto/ecdh/.
11730 *Douglas Stebila (Sun Microsystems Laboratories)*
11732 * Let BN_rand_range() abort with an error after 100 iterations
11733 without success (which indicates a broken PRNG).
11737 * Change BN_mod_sqrt() so that it verifies that the input value
11738 is really the square of the return value. (Previously,
11739 BN_mod_sqrt would show GIGO behaviour.)
11743 * Add named elliptic curves over binary fields from X9.62, SECG,
11744 and WAP/WTLS; add OIDs that were still missing.
11746 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11748 * Extend the EC library for elliptic curves over binary fields
11749 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11752 EC_GF2m_simple_method
11756 EC_GROUP_new_curve_GF2m
11757 EC_GROUP_set_curve_GF2m
11758 EC_GROUP_get_curve_GF2m
11759 EC_POINT_set_affine_coordinates_GF2m
11760 EC_POINT_get_affine_coordinates_GF2m
11761 EC_POINT_set_compressed_coordinates_GF2m
11763 Point compression for binary fields is disabled by default for
11764 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11767 As binary polynomials are represented as BIGNUMs, various members
11768 of the EC_GROUP and EC_POINT data structures can be shared
11769 between the implementations for prime fields and binary fields;
11770 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11771 are essentially identical to their `..._GFp` counterparts.
11772 (For simplicity, the `..._GFp` prefix has been dropped from
11773 various internal method names.)
11775 An internal 'field_div' method (similar to 'field_mul' and
11776 'field_sqr') has been added; this is used only for binary fields.
11778 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11780 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11781 through methods ('mul', 'precompute_mult').
11783 The generic implementations (now internally called 'ec_wNAF_mul'
11784 and 'ec_wNAF_precomputed_mult') remain the default if these
11785 methods are undefined.
11787 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11789 * New function EC_GROUP_get_degree, which is defined through
11790 EC_METHOD. For curves over prime fields, this returns the bit
11791 length of the modulus.
11793 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11795 * New functions EC_GROUP_dup, EC_POINT_dup.
11796 (These simply call ..._new and ..._copy).
11798 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11800 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11801 Polynomials are represented as BIGNUMs (where the sign bit is not
11802 used) in the following functions [macros]:
11805 BN_GF2m_sub [= BN_GF2m_add]
11806 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11807 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11808 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11810 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11811 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11812 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11813 BN_GF2m_cmp [= BN_ucmp]
11815 (Note that only the 'mod' functions are actually for fields GF(2^m).
11816 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11818 For some functions, an the irreducible polynomial defining a
11819 field can be given as an 'unsigned int[]' with strictly
11820 decreasing elements giving the indices of those bits that are set;
11821 i.e., p[] represents the polynomial
11822 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11824 p[0] > p[1] > ... > p[k] = 0.
11825 This applies to the following functions:
11828 BN_GF2m_mod_mul_arr
11829 BN_GF2m_mod_sqr_arr
11830 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11831 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11832 BN_GF2m_mod_exp_arr
11833 BN_GF2m_mod_sqrt_arr
11834 BN_GF2m_mod_solve_quad_arr
11838 Conversion can be performed by the following functions:
11843 bntest.c has additional tests for binary polynomial arithmetic.
11845 Two implementations for BN_GF2m_mod_div() are available.
11846 The default algorithm simply uses BN_GF2m_mod_inv() and
11847 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11848 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11849 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11851 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11853 * Add new error code 'ERR_R_DISABLED' that can be used when some
11854 functionality is disabled at compile-time.
11856 *Douglas Stebila <douglas.stebila@sun.com>*
11858 * Change default behaviour of 'openssl asn1parse' so that more
11859 information is visible when viewing, e.g., a certificate:
11861 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11862 mode the content of non-printable OCTET STRINGs is output in a
11863 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11864 avoid the appearance of a printable string.
11866 *Nils Larsch <nla@trustcenter.de>*
11868 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11870 EC_GROUP_set_asn1_flag()
11871 EC_GROUP_get_asn1_flag()
11872 EC_GROUP_set_point_conversion_form()
11873 EC_GROUP_get_point_conversion_form()
11874 These control ASN1 encoding details:
11875 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11876 has been set to OPENSSL_EC_NAMED_CURVE.
11877 - Points are encoded in uncompressed form by default; options for
11878 asn1_for are as for point2oct, namely
11879 POINT_CONVERSION_COMPRESSED
11880 POINT_CONVERSION_UNCOMPRESSED
11881 POINT_CONVERSION_HYBRID
11883 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11885 EC_GROUP_set_seed()
11886 EC_GROUP_get0_seed()
11887 EC_GROUP_get_seed_len()
11888 This is used only for ASN1 purposes (so far).
11890 *Nils Larsch <nla@trustcenter.de>*
11892 * Add 'field_type' member to EC_METHOD, which holds the NID
11893 of the appropriate field type OID. The new function
11894 EC_METHOD_get_field_type() returns this value.
11896 *Nils Larsch <nla@trustcenter.de>*
11899 EC_POINT_point2bn()
11900 EC_POINT_bn2point()
11901 EC_POINT_point2hex()
11902 EC_POINT_hex2point()
11903 providing useful interfaces to EC_POINT_point2oct() and
11904 EC_POINT_oct2point().
11906 *Nils Larsch <nla@trustcenter.de>*
11908 * Change internals of the EC library so that the functions
11909 EC_GROUP_set_generator()
11910 EC_GROUP_get_generator()
11911 EC_GROUP_get_order()
11912 EC_GROUP_get_cofactor()
11913 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11914 to methods, which would lead to unnecessary code duplication when
11915 adding different types of curves.
11917 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11919 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11920 arithmetic, and such that modified wNAFs are generated
11921 (which avoid length expansion in many cases).
11925 * Add a function EC_GROUP_check_discriminant() (defined via
11926 EC_METHOD) that verifies that the curve discriminant is non-zero.
11928 Add a function EC_GROUP_check() that makes some sanity tests
11929 on a EC_GROUP, its generator and order. This includes
11930 EC_GROUP_check_discriminant().
11932 *Nils Larsch <nla@trustcenter.de>*
11934 * Add ECDSA in new directory crypto/ecdsa/.
11936 Add applications 'openssl ecparam' and 'openssl ecdsa'
11937 (these are based on 'openssl dsaparam' and 'openssl dsa').
11939 ECDSA support is also included in various other files across the
11940 library. Most notably,
11941 - 'openssl req' now has a '-newkey ecdsa:file' option;
11942 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11943 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11944 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11945 them suitable for ECDSA where domain parameters must be
11946 extracted before the specific public key;
11947 - ECDSA engine support has been added.
11949 *Nils Larsch <nla@trustcenter.de>*
11951 * Include some named elliptic curves, and add OIDs from X9.62,
11952 SECG, and WAP/WTLS. Each curve can be obtained from the new
11954 EC_GROUP_new_by_curve_name(),
11955 and the list of available named curves can be obtained with
11956 EC_get_builtin_curves().
11957 Also add a 'curve_name' member to EC_GROUP objects, which can be
11959 EC_GROUP_set_curve_name()
11960 EC_GROUP_get_curve_name()
11962 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11964 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11965 was actually never needed) and in BN_mul(). The removal in BN_mul()
11966 required a small change in bn_mul_part_recursive() and the addition
11967 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11968 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11969 bn_sub_words() and bn_add_words() except they take arrays with
11974 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11976 * Cleanse PEM buffers before freeing them since they may contain
11979 *Benjamin Bennett <ben@psc.edu>*
11981 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11982 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11983 authentication-only ciphersuites.
11987 * Since AES128 and AES256 share a single mask bit in the logic of
11988 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11989 kludge to work properly if AES128 is available and AES256 isn't.
11993 * Expand security boundary to match 1.1.1 module.
11997 * Remove redundant features: hash file source, editing of test vectors
11998 modify fipsld to use external fips_premain.c signature.
12002 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12003 run algorithm test programs.
12007 * Make algorithm test programs more tolerant of whitespace.
12011 * Have SSL/TLS server implementation tolerate "mismatched" record
12012 protocol version while receiving ClientHello even if the
12013 ClientHello is fragmented. (The server can't insist on the
12014 particular protocol version it has chosen before the ServerHello
12015 message has informed the client about his choice.)
12019 * Load error codes if they are not already present instead of using a
12020 static variable. This allows them to be cleanly unloaded and reloaded.
12024 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12026 * Introduce limits to prevent malicious keys being able to
12027 cause a denial of service. ([CVE-2006-2940])
12029 *Steve Henson, Bodo Moeller*
12031 * Fix ASN.1 parsing of certain invalid structures that can result
12032 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12034 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12035 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12037 * Fix SSL client code which could crash if connecting to a
12038 malicious SSLv2 server. ([CVE-2006-4343])
12040 *Tavis Ormandy and Will Drewry, Google Security Team*
12042 * Change ciphersuite string processing so that an explicit
12043 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12044 will no longer include "AES128-SHA"), and any other similar
12045 ciphersuite (same bitmap) from *other* protocol versions (so that
12046 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12047 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12048 changes from 0.9.8b and 0.9.8d.
12052 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12054 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12055 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12057 * Change the Unix randomness entropy gathering to use poll() when
12058 possible instead of select(), since the latter has some
12059 undesirable limitations.
12061 *Darryl Miles via Richard Levitte and Bodo Moeller*
12063 * Disable rogue ciphersuites:
12065 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12066 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12067 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12069 The latter two were purportedly from
12070 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12073 Also deactivate the remaining ciphersuites from
12074 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12075 unofficial, and the ID has long expired.
12079 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12080 dual-core machines) and other potential thread-safety issues.
12084 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12086 * Adapt fipsld and the build system to link against the validated FIPS
12087 module in FIPS mode.
12091 * Fixes for VC++ 2005 build under Windows.
12095 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12096 from a Windows bash shell such as MSYS. It is autodetected from the
12097 "config" script when run from a VC++ environment. Modify standard VC++
12098 build to use fipscanister.o from the GNU make build.
12102 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12104 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12105 The value now differs depending on if you build for FIPS or not.
12106 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12107 safely run with a non-FIPSed libcrypto, as it may crash because of
12108 the difference induced by this change.
12112 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12114 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12115 (part of SSL_OP_ALL). This option used to disable the
12116 countermeasure against man-in-the-middle protocol-version
12117 rollback in the SSL 2.0 server implementation, which is a bad
12118 idea. ([CVE-2005-2969])
12120 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12121 for Information Security, National Institute of Advanced Industrial
12122 Science and Technology [AIST, Japan)]*
12124 * Minimal support for X9.31 signatures and PSS padding modes. This is
12125 mainly for FIPS compliance and not fully integrated at this stage.
12129 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12130 the exponentiation using a fixed-length exponent. (Otherwise,
12131 the information leaked through timing could expose the secret key
12132 after many signatures; cf. Bleichenbacher's attack on DSA with
12137 * Make a new fixed-window mod_exp implementation the default for
12138 RSA, DSA, and DH private-key operations so that the sequence of
12139 squares and multiplies and the memory access pattern are
12140 independent of the particular secret key. This will mitigate
12141 cache-timing and potential related attacks.
12143 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12144 and this is automatically used by BN_mod_exp_mont() if the new flag
12145 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12146 will use this BN flag for private exponents unless the flag
12147 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12148 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12150 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12152 * Change the client implementation for SSLv23_method() and
12153 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12154 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12155 (Previously, the SSL 2.0 backwards compatible Client Hello
12156 message format would be used even with SSL_OP_NO_SSLv2.)
12160 * Add support for smime-type MIME parameter in S/MIME messages which some
12165 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12166 a threadsafe manner. Modify rsa code to use new function and add calls
12167 to dsa and dh code (which had race conditions before).
12171 * Include the fixed error library code in the C error file definitions
12172 instead of fixing them up at runtime. This keeps the error code
12173 structures constant.
12177 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12179 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12182 * Fixes for newer kerberos headers. NB: the casts are needed because
12183 the 'length' field is signed on one version and unsigned on another
12184 with no (?) obvious way to tell the difference, without these VC++
12185 complains. Also the "definition" of FAR (blank) is no longer included
12186 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12187 some needed definitions.
12191 * Undo Cygwin change.
12195 * Added support for proxy certificates according to RFC 3820.
12196 Because they may be a security thread to unaware applications,
12197 they must be explicitly allowed in run-time. See
12198 docs/HOWTO/proxy_certificates.txt for further information.
12202 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12204 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12205 server and client random values. Previously
12206 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12207 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12209 This change has negligible security impact because:
12211 1. Server and client random values still have 24 bytes of pseudo random
12214 2. Server and client random values are sent in the clear in the initial
12217 3. The master secret is derived using the premaster secret (48 bytes in
12218 size for static RSA ciphersuites) as well as client server and random
12221 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12224 *Stephen Henson, reported by UK NISCC*
12226 * Use Windows randomness collection on Cygwin.
12230 * Fix hang in EGD/PRNGD query when communication socket is closed
12231 prematurely by EGD/PRNGD.
12233 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12235 * Prompt for pass phrases when appropriate for PKCS12 input format.
12239 * Back-port of selected performance improvements from development
12240 branch, as well as improved support for PowerPC platforms.
12244 * Add lots of checks for memory allocation failure, error codes to indicate
12245 failure and freeing up memory if a failure occurs.
12247 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12249 * Add new -passin argument to dgst.
12253 * Perform some character comparisons of different types in X509_NAME_cmp:
12254 this is needed for some certificates that re-encode DNs into UTF8Strings
12255 (in violation of RFC3280) and can't or won't issue name rollover
12260 * Make an explicit check during certificate validation to see that
12261 the CA setting in each certificate on the chain is correct. As a
12262 side effect always do the following basic checks on extensions,
12263 not just when there's an associated purpose to the check:
12265 - if there is an unhandled critical extension (unless the user
12266 has chosen to ignore this fault)
12267 - if the path length has been exceeded (if one is set at all)
12268 - that certain extensions fit the associated purpose (if one has
12273 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12275 * Avoid a race condition when CRLs are checked in a multi threaded
12276 environment. This would happen due to the reordering of the revoked
12277 entries during signature checking and serial number lookup. Now the
12278 encoding is cached and the serial number sort performed under a lock.
12279 Add new STACK function sk_is_sorted().
12283 * Add Delta CRL to the extension code.
12287 * Various fixes to s3_pkt.c so alerts are sent properly.
12289 *David Holmes <d.holmes@f5.com>*
12291 * Reduce the chances of duplicate issuer name and serial numbers (in
12292 violation of RFC3280) using the OpenSSL certificate creation utilities.
12293 This is done by creating a random 64 bit value for the initial serial
12294 number when a serial number file is created or when a self signed
12295 certificate is created using 'openssl req -x509'. The initial serial
12296 number file is created using 'openssl x509 -next_serial' in CA.pl
12297 rather than being initialized to 1.
12301 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12303 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12304 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12306 *Joe Orton, Steve Henson*
12308 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12311 *Joe Orton, Steve Henson*
12313 * Make it possible to have multiple active certificates with the same
12314 subject in the CA index file. This is done only if the keyword
12315 'unique_subject' is set to 'no' in the main CA section (default
12316 if 'CA_default') of the configuration file. The value is saved
12317 with the database itself in a separate index attribute file,
12318 named like the index file with '.attr' appended to the name.
12322 * X509 verify fixes. Disable broken certificate workarounds when
12323 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12324 keyUsage extension present. Don't accept CRLs with unhandled critical
12325 extensions: since verify currently doesn't process CRL extensions this
12326 rejects a CRL with *any* critical extensions. Add new verify error codes
12331 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12332 A clarification of RFC2560 will require the use of OCTET STRINGs and
12333 some implementations cannot handle the current raw format. Since OpenSSL
12334 copies and compares OCSP nonces as opaque blobs without any attempt at
12335 parsing them this should not create any compatibility issues.
12339 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12340 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12341 this HMAC (and other) operations are several times slower than OpenSSL
12346 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12348 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12350 * Use the correct content when signing type "other".
12354 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12356 * Fix various bugs revealed by running the NISCC test suite:
12358 Stop out of bounds reads in the ASN1 code when presented with
12359 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12361 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12363 If verify callback ignores invalid public key errors don't try to check
12364 certificate signature with the NULL public key.
12368 * New -ignore_err option in ocsp application to stop the server
12369 exiting on the first error in a request.
12373 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12374 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12379 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12380 extra data after the compression methods not only for TLS 1.0
12381 but also for SSL 3.0 (as required by the specification).
12383 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12385 * Change X509_certificate_type() to mark the key as exported/exportable
12386 when it's 512 *bits* long, not 512 bytes.
12390 * Change AES_cbc_encrypt() so it outputs exact multiple of
12391 blocks during encryption.
12395 * Various fixes to base64 BIO and non blocking I/O. On write
12396 flushes were not handled properly if the BIO retried. On read
12397 data was not being buffered properly and had various logic bugs.
12398 This also affects blocking I/O when the data being decoded is a
12403 * Various S/MIME bugfixes and compatibility changes:
12404 output correct application/pkcs7 MIME type if
12405 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12406 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12407 of files as .eml work). Correctly handle very long lines in MIME
12412 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12414 * Countermeasure against the Klima-Pokorny-Rosa extension of
12415 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12416 a protocol version number mismatch like a decryption error
12417 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12421 * Turn on RSA blinding by default in the default implementation
12422 to avoid a timing attack. Applications that don't want it can call
12423 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12424 They would be ill-advised to do so in most cases.
12426 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12428 * Change RSA blinding code so that it works when the PRNG is not
12429 seeded (in this case, the secret RSA exponent is abused as
12430 an unpredictable seed -- if it is not unpredictable, there
12431 is no point in blinding anyway). Make RSA blinding thread-safe
12432 by remembering the creator's thread ID in rsa->blinding and
12433 having all other threads use local one-time blinding factors
12434 (this requires more computation than sharing rsa->blinding, but
12435 avoids excessive locking; and if an RSA object is not shared
12436 between threads, blinding will still be very fast).
12440 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12441 ENGINE as defaults for all supported algorithms irrespective of
12442 the 'flags' parameter. 'flags' is now honoured, so applications
12443 should make sure they are passing it correctly.
12447 * Target "mingw" now allows native Windows code to be generated in
12448 the Cygwin environment as well as with the MinGW compiler.
12452 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12454 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12455 via timing by performing a MAC computation even if incorrect
12456 block cipher padding has been found. This is a countermeasure
12457 against active attacks where the attacker has to distinguish
12458 between bad padding and a MAC verification error. ([CVE-2003-0078])
12460 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12461 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12462 Martin Vuagnoux (EPFL, Ilion)*
12464 * Make the no-err option work as intended. The intention with no-err
12465 is not to have the whole error stack handling routines removed from
12466 libcrypto, it's only intended to remove all the function name and
12467 reason texts, thereby removing some of the footprint that may not
12468 be interesting if those errors aren't displayed anyway.
12470 NOTE: it's still possible for any application or module to have its
12471 own set of error texts inserted. The routines are there, just not
12472 used by default when no-err is given.
12476 * Add support for FreeBSD on IA64.
12478 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12480 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12481 Kerberos function mit_des_cbc_cksum(). Before this change,
12482 the value returned by DES_cbc_cksum() was like the one from
12483 mit_des_cbc_cksum(), except the bytes were swapped.
12485 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12487 * Allow an application to disable the automatic SSL chain building.
12488 Before this a rather primitive chain build was always performed in
12489 ssl3_output_cert_chain(): an application had no way to send the
12490 correct chain if the automatic operation produced an incorrect result.
12492 Now the chain builder is disabled if either:
12494 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12496 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12498 The reasoning behind this is that an application would not want the
12499 auto chain building to take place if extra chain certificates are
12500 present and it might also want a means of sending no additional
12501 certificates (for example the chain has two certificates and the
12506 * Add the possibility to build without the ENGINE framework.
12508 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12510 * Under Win32 gmtime() can return NULL: check return value in
12511 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12515 * DSA routines: under certain error conditions uninitialized BN objects
12516 could be freed. Solution: make sure initialization is performed early
12517 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12518 Nils Larsch <nla@trustcenter.de> via PR#459)
12522 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12523 checked on reconnect on the client side, therefore session resumption
12524 could still fail with a "ssl session id is different" error. This
12525 behaviour is masked when SSL_OP_ALL is used due to
12526 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12527 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12528 followup to PR #377.
12532 * IA-32 assembler support enhancements: unified ELF targets, support
12533 for SCO/Caldera platforms, fix for Cygwin shared build.
12537 * Add support for FreeBSD on sparc64. As a consequence, support for
12538 FreeBSD on non-x86 processors is separate from x86 processors on
12539 the config script, much like the NetBSD support.
12541 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12543 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12545 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12548 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12549 code (06) was taken as the first octet of the session ID and the last
12550 octet was ignored consequently. As a result SSLv2 client side session
12551 caching could not have worked due to the session ID mismatch between
12553 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12558 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12559 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12564 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12565 seems that in spite of existing for more than a year, many application
12566 author have done nothing to provide the necessary callbacks, which
12567 means that this particular engine will not work properly anywhere.
12568 This is a very unfortunate situation which forces us, in the name
12569 of usability, to give the hw_ncipher.c a static lock, which is part
12571 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12572 appear in 0.9.8 or later. We EXPECT application authors to have
12573 dealt properly with this when 0.9.8 is released (unless we actually
12574 make such changes in the libcrypto locking code that changes will
12575 have to be made anyway).
12579 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12580 octets have been read, EOF or an error occurs. Without this change
12581 some truncated ASN1 structures will not produce an error.
12585 * Disable Heimdal support, since it hasn't been fully implemented.
12586 Still give the possibility to force the use of Heimdal, but with
12587 warnings and a request that patches get sent to openssl-dev.
12591 * Add the VC-CE target, introduce the WINCE sysname, and add
12592 INSTALL.WCE and appropriate conditionals to make it build.
12594 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12596 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12597 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12598 edit numbers of the version.
12600 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12602 * Introduce safe string copy and catenation functions
12603 (BUF_strlcpy() and BUF_strlcat()).
12605 *Ben Laurie (CHATS) and Richard Levitte*
12607 * Avoid using fixed-size buffers for one-line DNs.
12609 *Ben Laurie (CHATS)*
12611 * Add BUF_MEM_grow_clean() to avoid information leakage when
12612 resizing buffers containing secrets, and use where appropriate.
12614 *Ben Laurie (CHATS)*
12616 * Avoid using fixed size buffers for configuration file location.
12618 *Ben Laurie (CHATS)*
12620 * Avoid filename truncation for various CA files.
12622 *Ben Laurie (CHATS)*
12624 * Use sizeof in preference to magic numbers.
12626 *Ben Laurie (CHATS)*
12628 * Avoid filename truncation in cert requests.
12630 *Ben Laurie (CHATS)*
12632 * Add assertions to check for (supposedly impossible) buffer
12635 *Ben Laurie (CHATS)*
12637 * Don't cache truncated DNS entries in the local cache (this could
12638 potentially lead to a spoofing attack).
12640 *Ben Laurie (CHATS)*
12642 * Fix various buffers to be large enough for hex/decimal
12643 representations in a platform independent manner.
12645 *Ben Laurie (CHATS)*
12647 * Add CRYPTO_realloc_clean() to avoid information leakage when
12648 resizing buffers containing secrets, and use where appropriate.
12650 *Ben Laurie (CHATS)*
12652 * Add BIO_indent() to avoid much slightly worrying code to do
12655 *Ben Laurie (CHATS)*
12657 * Convert sprintf()/BIO_puts() to BIO_printf().
12659 *Ben Laurie (CHATS)*
12661 * buffer_gets() could terminate with the buffer only half
12664 *Ben Laurie (CHATS)*
12666 * Add assertions to prevent user-supplied crypto functions from
12667 overflowing internal buffers by having large block sizes, etc.
12669 *Ben Laurie (CHATS)*
12671 * New OPENSSL_assert() macro (similar to assert(), but enabled
12674 *Ben Laurie (CHATS)*
12676 * Eliminate unused copy of key in RC4.
12678 *Ben Laurie (CHATS)*
12680 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12682 *Ben Laurie (CHATS)*
12684 * Fix off-by-one error in EGD path.
12686 *Ben Laurie (CHATS)*
12688 * If RANDFILE path is too long, ignore instead of truncating.
12690 *Ben Laurie (CHATS)*
12692 * Eliminate unused and incorrectly sized X.509 structure
12695 *Ben Laurie (CHATS)*
12697 * Eliminate unused and dangerous function knumber().
12699 *Ben Laurie (CHATS)*
12701 * Eliminate unused and dangerous structure, KSSL_ERR.
12703 *Ben Laurie (CHATS)*
12705 * Protect against overlong session ID context length in an encoded
12706 session object. Since these are local, this does not appear to be
12709 *Ben Laurie (CHATS)*
12711 * Change from security patch (see 0.9.6e below) that did not affect
12712 the 0.9.6 release series:
12714 Remote buffer overflow in SSL3 protocol - an attacker could
12715 supply an oversized master key in Kerberos-enabled versions.
12718 *Ben Laurie (CHATS)*
12720 * Change the SSL kerb5 codes to match RFC 2712.
12724 * Make -nameopt work fully for req and add -reqopt switch.
12726 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12728 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12730 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12732 * Make sure tests can be performed even if the corresponding algorithms
12733 have been removed entirely. This was also the last step to make
12734 OpenSSL compilable with DJGPP under all reasonable conditions.
12736 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12738 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12739 to allow version independent disabling of normally unselected ciphers,
12740 which may be activated as a side-effect of selecting a single cipher.
12742 (E.g., cipher list string "RSA" enables ciphersuites that are left
12743 out of "ALL" because they do not provide symmetric encryption.
12744 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12746 *Lutz Jaenicke, Bodo Moeller*
12748 * Add appropriate support for separate platform-dependent build
12749 directories. The recommended way to make a platform-dependent
12750 build directory is the following (tested on Linux), maybe with
12753 # Place yourself outside of the OpenSSL source tree. In
12754 # this example, the environment variable OPENSSL_SOURCE
12755 # is assumed to contain the absolute OpenSSL source directory.
12756 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12757 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12758 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12759 mkdir -p `dirname $F`
12760 ln -s $OPENSSL_SOURCE/$F $F
12763 To be absolutely sure not to disturb the source tree, a "make clean"
12764 is a good thing. If it isn't successful, don't worry about it,
12765 it probably means the source directory is very clean.
12769 * Make sure any ENGINE control commands make local copies of string
12770 pointers passed to them whenever necessary. Otherwise it is possible
12771 the caller may have overwritten (or deallocated) the original string
12772 data when a later ENGINE operation tries to use the stored values.
12774 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12776 * Improve diagnostics in file reading and command-line digests.
12778 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12780 * Add AES modes CFB and OFB to the object database. Correct an
12781 error in AES-CFB decryption.
12785 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12786 allows existing EVP_CIPHER_CTX structures to be reused after
12787 calling `EVP_*Final()`. This behaviour is used by encryption
12788 BIOs and some applications. This has the side effect that
12789 applications must explicitly clean up cipher contexts with
12790 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12794 * Check the values of dna and dnb in bn_mul_recursive before calling
12795 bn_mul_comba (a non zero value means the a or b arrays do not contain
12796 n2 elements) and fallback to bn_mul_normal if either is not zero.
12800 * Fix escaping of non-ASCII characters when using the -subj option
12801 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12805 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12806 form for "surname", serialNumber has no short form.
12807 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12808 therefore remove "mail" short name for "internet 7".
12809 The OID for unique identifiers in X509 certificates is
12810 x500UniqueIdentifier, not uniqueIdentifier.
12811 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12815 * Add an "init" command to the ENGINE config module and auto initialize
12816 ENGINEs. Without any "init" command the ENGINE will be initialized
12817 after all ctrl commands have been executed on it. If init=1 the
12818 ENGINE is initialized at that point (ctrls before that point are run
12819 on the uninitialized ENGINE and after on the initialized one). If
12820 init=0 then the ENGINE will not be initialized at all.
12824 * Fix the 'app_verify_callback' interface so that the user-defined
12825 argument is actually passed to the callback: In the
12826 SSL_CTX_set_cert_verify_callback() prototype, the callback
12827 declaration has been changed from
12830 int (*cb)(X509_STORE_CTX *,void *);
12831 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12832 i=s->ctx->app_verify_callback(&ctx)
12833 has been changed into
12834 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12836 To update applications using SSL_CTX_set_cert_verify_callback(),
12837 a dummy argument can be added to their callback functions.
12839 *D. K. Smetters <smetters@parc.xerox.com>*
12841 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12843 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12845 * Add and OPENSSL_LOAD_CONF define which will cause
12846 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12847 This allows older applications to transparently support certain
12848 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12849 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12850 load the config file and OPENSSL_add_all_algorithms_conf() which will
12851 always load it have also been added.
12855 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12856 Adjust NIDs and EVP layer.
12858 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12860 * Config modules support in openssl utility.
12862 Most commands now load modules from the config file,
12863 though in a few (such as version) this isn't done
12864 because it couldn't be used for anything.
12866 In the case of ca and req the config file used is
12867 the same as the utility itself: that is the -config
12868 command line option can be used to specify an
12873 * Move default behaviour from OPENSSL_config(). If appname is NULL
12874 use "openssl_conf" if filename is NULL use default openssl config file.
12878 * Add an argument to OPENSSL_config() to allow the use of an alternative
12879 config section name. Add a new flag to tolerate a missing config file
12880 and move code to CONF_modules_load_file().
12884 * Support for crypto accelerator cards from Accelerated Encryption
12885 Processing, www.aep.ie. (Use engine 'aep')
12886 The support was copied from 0.9.6c [engine] and adapted/corrected
12887 to work with the new engine framework.
12889 *AEP Inc. and Richard Levitte*
12891 * Support for SureWare crypto accelerator cards from Baltimore
12892 Technologies. (Use engine 'sureware')
12893 The support was copied from 0.9.6c [engine] and adapted
12894 to work with the new engine framework.
12898 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12899 make the newer ENGINE framework commands for the CHIL engine work.
12901 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12903 * Make it possible to produce shared libraries on ReliantUNIX.
12905 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12907 * Add the configuration target debug-linux-ppro.
12908 Make 'openssl rsa' use the general key loading routines
12909 implemented in `apps.c`, and make those routines able to
12910 handle the key format FORMAT_NETSCAPE and the variant
12913 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12915 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12917 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12919 * Add -keyform to rsautl, and document -engine.
12921 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12923 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12924 BIO_R_NO_SUCH_FILE error code rather than the generic
12925 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12929 * Add new functions
12930 ERR_peek_last_error
12931 ERR_peek_last_error_line
12932 ERR_peek_last_error_line_data.
12933 These are similar to
12935 ERR_peek_error_line
12936 ERR_peek_error_line_data,
12937 but report on the latest error recorded rather than the first one
12938 still in the error queue.
12940 *Ben Laurie, Bodo Moeller*
12942 * default_algorithms option in ENGINE config module. This allows things
12944 default_algorithms = ALL
12945 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12949 * Preliminary ENGINE config module.
12953 * New experimental application configuration code.
12957 * Change the AES code to follow the same name structure as all other
12958 symmetric ciphers, and behave the same way. Move everything to
12959 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12961 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12963 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12965 *Ben Laurie and Theo de Raadt*
12967 * Add option to output public keys in req command.
12969 *Massimiliano Pala madwolf@openca.org*
12971 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12972 (up to about 10% better than before for P-192 and P-224).
12976 * New functions/macros
12978 SSL_CTX_set_msg_callback(ctx, cb)
12979 SSL_CTX_set_msg_callback_arg(ctx, arg)
12980 SSL_set_msg_callback(ssl, cb)
12981 SSL_set_msg_callback_arg(ssl, arg)
12983 to request calling a callback function
12985 void cb(int write_p, int version, int content_type,
12986 const void *buf, size_t len, SSL *ssl, void *arg)
12988 whenever a protocol message has been completely received
12989 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12990 protocol version according to which the SSL library interprets
12991 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12992 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12993 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12994 specification (change_cipher_spec(20), alert(21), handshake(22)).
12995 'buf' and 'len' point to the actual message, 'ssl' to the
12996 SSL object, and 'arg' is the application-defined value set by
12997 SSL[_CTX]_set_msg_callback_arg().
12999 'openssl s_client' and 'openssl s_server' have new '-msg' options
13000 to enable a callback that displays all protocol messages.
13004 * Change the shared library support so shared libraries are built as
13005 soon as the corresponding static library is finished, and thereby get
13006 openssl and the test programs linked against the shared library.
13007 This still only happens when the keyword "shard" has been given to
13008 the configuration scripts.
13010 NOTE: shared library support is still an experimental thing, and
13011 backward binary compatibility is still not guaranteed.
13013 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13015 * Add support for Subject Information Access extension.
13017 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13019 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13020 additional bytes when new memory had to be allocated, not just
13021 when reusing an existing buffer.
13025 * New command line and configuration option 'utf8' for the req command.
13026 This allows field values to be specified as UTF8 strings.
13030 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13031 runs for the former and machine-readable output for the latter.
13035 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13036 of the e-mail address in the DN (i.e., it will go into a certificate
13037 extension only). The new configuration file option 'email_in_dn = no'
13038 has the same effect.
13040 *Massimiliano Pala madwolf@openca.org*
13042 * Change all functions with names starting with `des_` to be starting
13043 with `DES_` instead. Add wrappers that are compatible with libdes,
13044 but are named `_ossl_old_des_*`. Finally, add macros that map the
13045 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13046 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13047 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13050 Since we provide two compatibility mappings, the user needs to
13051 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13052 compatibility is desired. The default (i.e., when that macro
13053 isn't defined) is OpenSSL 0.9.6c compatibility.
13055 There are also macros that enable and disable the support of old
13056 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13057 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13058 are defined, the default will apply: to support the old des routines.
13060 In either case, one must include openssl/des.h to get the correct
13061 definitions. Do not try to just include openssl/des_old.h, that
13064 NOTE: This is a major break of an old API into a new one. Software
13065 authors are encouraged to switch to the `DES_` style functions. Some
13066 time in the future, des_old.h and the libdes compatibility functions
13067 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13068 default), and then completely removed.
13072 * Test for certificates which contain unsupported critical extensions.
13073 If such a certificate is found during a verify operation it is
13074 rejected by default: this behaviour can be overridden by either
13075 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13076 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13077 X509_supported_extension() has also been added which returns 1 if a
13078 particular extension is supported.
13082 * Modify the behaviour of EVP cipher functions in similar way to digests
13083 to retain compatibility with existing code.
13087 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13088 compatibility with existing code. In particular the 'ctx' parameter does
13089 not have to be to be initialized before the call to EVP_DigestInit() and
13090 it is tidied up after a call to EVP_DigestFinal(). New function
13091 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13092 EVP_MD_CTX_copy() changed to not require the destination to be
13093 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13094 requires the destination to be valid.
13096 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13097 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13101 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13102 so that complete 'Handshake' protocol structures are kept in memory
13103 instead of overwriting 'msg_type' and 'length' with 'body' data.
13107 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13109 *Massimo Santin via Richard Levitte*
13111 * Major restructuring to the underlying ENGINE code. This includes
13112 reduction of linker bloat, separation of pure "ENGINE" manipulation
13113 (initialisation, etc) from functionality dealing with implementations
13114 of specific crypto interfaces. This change also introduces integrated
13115 support for symmetric ciphers and digest implementations - so ENGINEs
13116 can now accelerate these by providing EVP_CIPHER and EVP_MD
13117 implementations of their own. This is detailed in
13118 [crypto/engine/README.md](crypto/engine/README.md)
13119 as it couldn't be adequately described here. However, there are a few
13120 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13121 were changed in the original introduction of ENGINE code have now
13122 reverted back - the hooking from this code to ENGINE is now a good
13123 deal more passive and at run-time, operations deal directly with
13124 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13125 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13126 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13127 they were not being used by the framework as there is no concept of a
13128 BIGNUM_METHOD and they could not be generalised to the new
13129 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13130 ENGINE_cpy() has been removed as it cannot be consistently defined in
13135 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13139 * Change mkdef.pl to sort symbols that get the same entry number,
13140 and make sure the automatically generated functions `ERR_load_*`
13141 become part of libeay.num as well.
13145 * New function SSL_renegotiate_pending(). This returns true once
13146 renegotiation has been requested (either SSL_renegotiate() call
13147 or HelloRequest/ClientHello received from the peer) and becomes
13148 false once a handshake has been completed.
13149 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13150 sends a HelloRequest, but does not ensure that a handshake takes
13151 place. SSL_renegotiate_pending() is useful for checking if the
13152 client has followed the request.)
13156 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13157 By default, clients may request session resumption even during
13158 renegotiation (if session ID contexts permit); with this option,
13159 session resumption is possible only in the first handshake.
13161 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13162 more bits available for options that should not be part of
13163 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13167 * Add some demos for certificate and certificate request creation.
13171 * Make maximum certificate chain size accepted from the peer application
13172 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13173 "Douglas E. Engert" <deengert@anl.gov>.
13177 * Add support for shared libraries for Unixware-7
13178 (Boyd Lynn Gerber <gerberb@zenez.com>).
13182 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13183 be done prior to destruction. Use this to unload error strings from
13184 ENGINEs that load their own error strings. NB: This adds two new API
13185 functions to "get" and "set" this destroy handler in an ENGINE.
13189 * Alter all existing ENGINE implementations (except "openssl" and
13190 "openbsd") to dynamically instantiate their own error strings. This
13191 makes them more flexible to be built both as statically-linked ENGINEs
13192 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13193 Also, add stub code to each that makes building them as self-contained
13194 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13198 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13199 implementations into applications that are completely implemented in
13200 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13201 commands that can be used to configure what shared-library to load and
13202 to control aspects of the way it is handled. Also, made an update to
13203 the [README-Engine.md](README-Engine.md) file
13204 that brings its information up-to-date and
13205 provides some information and instructions on the "dynamic" ENGINE
13206 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13210 * Make it possible to unload ranges of ERR strings with a new
13211 "ERR_unload_strings" function.
13215 * Add a copy() function to EVP_MD.
13219 * Make EVP_MD routines take a context pointer instead of just the
13220 md_data void pointer.
13224 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13225 that the digest can only process a single chunk of data
13226 (typically because it is provided by a piece of
13227 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13228 is only going to provide a single chunk of data, and hence the
13229 framework needn't accumulate the data for oneshot drivers.
13233 * As with "ERR", make it possible to replace the underlying "ex_data"
13234 functions. This change also alters the storage and management of global
13235 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13236 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13237 index counters. The API functions that use this state have been changed
13238 to take a "class_index" rather than pointers to the class's local STACK
13239 and counter, and there is now an API function to dynamically create new
13240 classes. This centralisation allows us to (a) plug a lot of the
13241 thread-safety problems that existed, and (b) makes it possible to clean
13242 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13243 such data would previously have always leaked in application code and
13244 workarounds were in place to make the memory debugging turn a blind eye
13245 to it. Application code that doesn't use this new function will still
13246 leak as before, but their memory debugging output will announce it now
13247 rather than letting it slide.
13249 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13250 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13251 has a return value to indicate success or failure.
13255 * Make it possible to replace the underlying "ERR" functions such that the
13256 global state (2 LHASH tables and 2 locks) is only used by the "default"
13257 implementation. This change also adds two functions to "get" and "set"
13258 the implementation prior to it being automatically set the first time
13259 any other ERR function takes place. Ie. an application can call "get",
13260 pass the return value to a module it has just loaded, and that module
13261 can call its own "set" function using that value. This means the
13262 module's "ERR" operations will use (and modify) the error state in the
13263 application and not in its own statically linked copy of OpenSSL code.
13267 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13268 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13269 the operation, and provides a more encapsulated way for external code
13270 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13271 to use these functions rather than manually incrementing the counts.
13273 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13277 * Add EVP test program.
13281 * Add symmetric cipher support to ENGINE. Expect the API to change!
13285 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13286 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13287 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13288 These allow a CRL to be built without having to access X509_CRL fields
13289 directly. Modify 'ca' application to use new functions.
13293 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13294 bug workarounds. Rollback attack detection is a security feature.
13295 The problem will only arise on OpenSSL servers when TLSv1 is not
13296 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13297 Software authors not wanting to support TLSv1 will have special reasons
13298 for their choice and can explicitly enable this option.
13300 *Bodo Moeller, Lutz Jaenicke*
13302 * Rationalise EVP so it can be extended: don't include a union of
13303 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13304 (similar to those existing for EVP_CIPHER_CTX).
13309 EVP_MD_CTX_init(&md); /* new function call */
13310 EVP_DigestInit(&md, EVP_sha1());
13311 EVP_DigestUpdate(&md, in, len);
13312 EVP_DigestFinal(&md, out, NULL);
13313 EVP_MD_CTX_cleanup(&md); /* new function call */
13317 * Make DES key schedule conform to the usual scheme, as well as
13318 correcting its structure. This means that calls to DES functions
13319 now have to pass a pointer to a des_key_schedule instead of a
13320 plain des_key_schedule (which was actually always a pointer
13323 des_key_schedule ks;
13325 des_set_key_checked(..., &ks);
13326 des_ncbc_encrypt(..., &ks, ...);
13328 (Note that a later change renames 'des_...' into 'DES_...'.)
13332 * Initial reduction of linker bloat: the use of some functions, such as
13333 PEM causes large amounts of unused functions to be linked in due to
13334 poor organisation. For example pem_all.c contains every PEM function
13335 which has a knock on effect of linking in large amounts of (unused)
13336 ASN1 code. Grouping together similar functions and splitting unrelated
13337 functions prevents this.
13341 * Cleanup of EVP macros.
13345 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13346 correct `_ecb suffix`.
13350 * Add initial OCSP responder support to ocsp application. The
13351 revocation information is handled using the text based index
13352 use by the ca application. The responder can either handle
13353 requests generated internally, supplied in files (for example
13354 via a CGI script) or using an internal minimal server.
13358 * Add configuration choices to get zlib compression for TLS.
13362 * Changes to Kerberos SSL for RFC 2712 compliance:
13363 1. Implemented real KerberosWrapper, instead of just using
13364 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13365 2. Implemented optional authenticator field of KerberosWrapper.
13367 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13368 and authenticator structs; see crypto/krb5/.
13370 Generalized Kerberos calls to support multiple Kerberos libraries.
13371 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13372 via Richard Levitte*
13374 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13375 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13376 values for each of the key sizes rather than having just
13377 parameters (and 'speed' generating keys each time).
13381 * Speed up EVP routines.
13384 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13385 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13386 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13387 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13389 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13390 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13391 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13394 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13396 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13400 * Added the OS2-EMX target.
13402 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13404 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13405 New functions to support `NCONF` routines in extension code.
13406 New function `CONF_set_nconf()`
13407 to allow functions which take an `NCONF` to also handle the old `LHASH`
13408 structure: this means that the old `CONF` compatible routines can be
13409 retained (in particular w.rt. extensions) without having to duplicate the
13410 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13414 * Enhance the general user interface with mechanisms for inner control
13415 and with possibilities to have yes/no kind of prompts.
13419 * Change all calls to low-level digest routines in the library and
13420 applications to use EVP. Add missing calls to HMAC_cleanup() and
13421 don't assume HMAC_CTX can be copied using memcpy().
13423 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13425 * Add the possibility to control engines through control names but with
13426 arbitrary arguments instead of just a string.
13427 Change the key loaders to take a UI_METHOD instead of a callback
13428 function pointer. NOTE: this breaks binary compatibility with earlier
13429 versions of OpenSSL [engine].
13430 Adapt the nCipher code for these new conditions and add a card insertion
13435 * Enhance the general user interface with mechanisms to better support
13436 dialog box interfaces, application-defined prompts, the possibility
13437 to use defaults (for example default passwords from somewhere else)
13438 and interrupts/cancellations.
13442 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13443 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13447 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13448 tidy up some unnecessarily weird code in 'sk_new()').
13450 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13452 * Change the key loading routines for ENGINEs to use the same kind
13453 callback (pem_password_cb) as all other routines that need this
13458 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13459 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13460 than this minimum value is recommended.
13464 * New random seeder for OpenVMS, using the system process statistics
13465 that are easily reachable.
13469 * Windows apparently can't transparently handle global
13470 variables defined in DLLs. Initialisations such as:
13472 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13474 won't compile. This is used by the any applications that need to
13475 declare their own ASN1 modules. This was fixed by adding the option
13476 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13477 needed for static libraries under Win32.
13481 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13482 setting of purpose and trust fields. New X509_STORE trust and
13483 purpose functions and tidy up setting in other SSL functions.
13487 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13488 structure. These are inherited by X509_STORE_CTX when it is
13489 initialised. This allows various defaults to be set in the
13490 X509_STORE structure (such as flags for CRL checking and custom
13491 purpose or trust settings) for functions which only use X509_STORE_CTX
13492 internally such as S/MIME.
13494 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13495 trust settings if they are not set in X509_STORE. This allows X509_STORE
13496 purposes and trust (in S/MIME for example) to override any set by default.
13498 Add command line options for CRL checking to smime, s_client and s_server
13503 * Initial CRL based revocation checking. If the CRL checking flag(s)
13504 are set then the CRL is looked up in the X509_STORE structure and
13505 its validity and signature checked, then if the certificate is found
13506 in the CRL the verify fails with a revoked error.
13508 Various new CRL related callbacks added to X509_STORE_CTX structure.
13510 Command line options added to 'verify' application to support this.
13512 This needs some additional work, such as being able to handle multiple
13513 CRLs with different times, extension based lookup (rather than just
13514 by subject name) and ultimately more complete V2 CRL extension
13519 * Add a general user interface API (crypto/ui/). This is designed
13520 to replace things like des_read_password and friends (backward
13521 compatibility functions using this new API are provided).
13522 The purpose is to remove prompting functions from the DES code
13523 section as well as provide for prompting through dialog boxes in
13524 a window system and the like.
13528 * Add "ex_data" support to ENGINE so implementations can add state at a
13529 per-structure level rather than having to store it globally.
13533 * Make it possible for ENGINE structures to be copied when retrieved by
13534 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13535 This causes the "original" ENGINE structure to act like a template,
13536 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13537 operational state can be localised to each ENGINE structure, despite the
13538 fact they all share the same "methods". New ENGINE structures returned in
13539 this case have no functional references and the return value is the single
13540 structural reference. This matches the single structural reference returned
13541 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13546 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13547 needs to match any other type at all we need to manually clear the
13552 * Changes to the "openssl engine" utility to include;
13553 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13554 about an ENGINE's available control commands.
13555 - executing control commands from command line arguments using the
13556 '-pre' and '-post' switches. '-post' is only used if '-t' is
13557 specified and the ENGINE is successfully initialised. The syntax for
13558 the individual commands are colon-separated, for example;
13559 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13563 * New dynamic control command support for ENGINEs. ENGINEs can now
13564 declare their own commands (numbers), names (strings), descriptions,
13565 and input types for run-time discovery by calling applications. A
13566 subset of these commands are implicitly classed as "executable"
13567 depending on their input type, and only these can be invoked through
13568 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13569 can be based on user input, config files, etc). The distinction is
13570 that "executable" commands cannot return anything other than a boolean
13571 result and can only support numeric or string input, whereas some
13572 discoverable commands may only be for direct use through
13573 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13574 pointers, or other custom uses. The "executable" commands are to
13575 support parameterisations of ENGINE behaviour that can be
13576 unambiguously defined by ENGINEs and used consistently across any
13577 OpenSSL-based application. Commands have been added to all the
13578 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13579 control over shared-library paths without source code alterations.
13583 * Changed all ENGINE implementations to dynamically allocate their
13584 ENGINEs rather than declaring them statically. Apart from this being
13585 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13586 this also allows the implementations to compile without using the
13587 internal engine_int.h header.
13591 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13592 'const' value. Any code that should be able to modify a RAND_METHOD
13593 should already have non-const pointers to it (ie. they should only
13594 modify their own ones).
13598 * Made a variety of little tweaks to the ENGINE code.
13599 - "atalla" and "ubsec" string definitions were moved from header files
13600 to C code. "nuron" string definitions were placed in variables
13601 rather than hard-coded - allowing parameterisation of these values
13602 later on via ctrl() commands.
13603 - Removed unused "#if 0"'d code.
13604 - Fixed engine list iteration code so it uses ENGINE_free() to release
13605 structural references.
13606 - Constified the RAND_METHOD element of ENGINE structures.
13607 - Constified various get/set functions as appropriate and added
13608 missing functions (including a catch-all ENGINE_cpy that duplicates
13609 all ENGINE values onto a new ENGINE except reference counts/state).
13610 - Removed NULL parameter checks in get/set functions. Setting a method
13611 or function to NULL is a way of cancelling out a previously set
13612 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13613 and doesn't justify the extra error symbols and code.
13614 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13615 flags from engine_int.h to engine.h.
13616 - Changed prototypes for ENGINE handler functions (init(), finish(),
13617 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13621 * Implement binary inversion algorithm for BN_mod_inverse in addition
13622 to the algorithm using long division. The binary algorithm can be
13623 used only if the modulus is odd. On 32-bit systems, it is faster
13624 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13625 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13626 up to 450 bits. In 64-bit environments, the binary algorithm
13627 appears to be advantageous for much longer moduli; here we use it
13628 for moduli up to 2048 bits.
13632 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13633 could not support the combine flag in choice fields.
13637 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13638 extensions from a certificate request to the certificate.
13642 * Allow multiple 'certopt' and 'nameopt' options to be separated
13643 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13644 file: this allows the display of the certificate about to be
13645 signed to be customised, to allow certain fields to be included
13646 or excluded and extension details. The old system didn't display
13647 multicharacter strings properly, omitted fields not in the policy
13648 and couldn't display additional details such as extensions.
13652 * Function EC_POINTs_mul for multiple scalar multiplication
13653 of an arbitrary number of elliptic curve points
13654 \sum scalars[i]*points[i],
13655 optionally including the generator defined for the EC_GROUP:
13656 scalar*generator + \sum scalars[i]*points[i].
13658 EC_POINT_mul is a simple wrapper function for the typical case
13659 that the point list has just one item (besides the optional
13664 * First EC_METHODs for curves over GF(p):
13666 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13667 operations and provides various method functions that can also
13668 operate with faster implementations of modular arithmetic.
13670 EC_GFp_mont_method() reuses most functions that are part of
13671 EC_GFp_simple_method, but uses Montgomery arithmetic.
13673 *Bodo Moeller; point addition and point doubling
13674 implementation directly derived from source code provided by
13675 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13677 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13678 crypto/ec/ec_lib.c):
13680 Curves are EC_GROUP objects (with an optional group generator)
13681 based on EC_METHODs that are built into the library.
13683 Points are EC_POINT objects based on EC_GROUP objects.
13685 Most of the framework would be able to handle curves over arbitrary
13686 finite fields, but as there are no obvious types for fields other
13687 than GF(p), some functions are limited to that for now.
13691 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13692 that the file contains a complete HTTP response.
13696 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13697 change the def and num file printf format specifier from "%-40sXXX"
13698 to "%-39s XXX". The latter will always guarantee a space after the
13699 field while the former will cause them to run together if the field
13700 is 40 of more characters long.
13704 * Constify the cipher and digest 'method' functions and structures
13705 and modify related functions to take constant EVP_MD and EVP_CIPHER
13710 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13711 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13715 * Modify `EVP_Digest*()` routines so they now return values. Although the
13716 internal software routines can never fail additional hardware versions
13721 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13723 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13724 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13727 ERR_R_NESTED_ASN1_ERROR
13729 ERR_R_MISSING_ASN1_EOS
13730 were 4 .. 9, conflicting with
13731 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13733 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13734 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13736 Add new error code 'ERR_R_INTERNAL_ERROR'.
13740 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13745 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13746 sets the subject name for a new request or supersedes the
13747 subject name in a given request. Formats that can be parsed are
13748 'CN=Some Name, OU=myOU, C=IT'
13750 'CN=Some Name/OU=myOU/C=IT'.
13752 Add options '-batch' and '-verbose' to 'openssl req'.
13754 *Massimiliano Pala <madwolf@hackmasters.net>*
13756 * Introduce the possibility to access global variables through
13757 functions on platform were that's the best way to handle exporting
13758 global variables in shared libraries. To enable this functionality,
13759 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13760 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13761 is normally done by Configure or something similar).
13763 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13764 in the source file (foo.c) like this:
13766 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13767 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13769 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13770 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13772 OPENSSL_DECLARE_GLOBAL(int,foo);
13773 #define foo OPENSSL_GLOBAL_REF(foo)
13774 OPENSSL_DECLARE_GLOBAL(double,bar);
13775 #define bar OPENSSL_GLOBAL_REF(bar)
13777 The #defines are very important, and therefore so is including the
13778 header file everywhere where the defined globals are used.
13780 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13781 of ASN.1 items, but that structure is a bit different.
13783 The largest change is in util/mkdef.pl which has been enhanced with
13784 better and easier to understand logic to choose which symbols should
13785 go into the Windows .def files as well as a number of fixes and code
13786 cleanup (among others, algorithm keywords are now sorted
13787 lexicographically to avoid constant rewrites).
13791 * In BN_div() keep a copy of the sign of 'num' before writing the
13792 result to 'rm' because if rm==num the value will be overwritten
13793 and produce the wrong result if 'num' is negative: this caused
13794 problems with BN_mod() and BN_nnmod().
13798 * Function OCSP_request_verify(). This checks the signature on an
13799 OCSP request and verifies the signer certificate. The signer
13800 certificate is just checked for a generic purpose and OCSP request
13805 * Add OCSP_check_validity() function to check the validity of OCSP
13806 responses. OCSP responses are prepared in real time and may only
13807 be a few seconds old. Simply checking that the current time lies
13808 between thisUpdate and nextUpdate max reject otherwise valid responses
13809 caused by either OCSP responder or client clock inaccuracy. Instead
13810 we allow thisUpdate and nextUpdate to fall within a certain period of
13811 the current time. The age of the response can also optionally be
13812 checked. Two new options -validity_period and -status_age added to
13817 * If signature or public key algorithm is unrecognized print out its
13818 OID rather that just UNKNOWN.
13822 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13823 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13824 ID to be generated from the issuer certificate alone which can then be
13825 passed to OCSP_id_issuer_cmp().
13829 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13830 ASN1 modules to export functions returning ASN1_ITEM pointers
13831 instead of the ASN1_ITEM structures themselves. This adds several
13832 new macros which allow the underlying ASN1 function/structure to
13833 be accessed transparently. As a result code should not use ASN1_ITEM
13834 references directly (such as &X509_it) but instead use the relevant
13835 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13836 use of the new ASN1 code on platforms where exporting structures
13837 is problematical (for example in shared libraries) but exporting
13838 functions returning pointers to structures is not.
13842 * Add support for overriding the generation of SSL/TLS session IDs.
13843 These callbacks can be registered either in an SSL_CTX or per SSL.
13844 The purpose of this is to allow applications to control, if they wish,
13845 the arbitrary values chosen for use as session IDs, particularly as it
13846 can be useful for session caching in multiple-server environments. A
13847 command-line switch for testing this (and any client code that wishes
13848 to use such a feature) has been added to "s_server".
13850 *Geoff Thorpe, Lutz Jaenicke*
13852 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13853 of the form `#if defined(...) || defined(...) || ...` and
13854 `#if !defined(...) && !defined(...) && ...`. This also avoids
13855 the growing number of special cases it was previously handling.
13859 * Make all configuration macros available for application by making
13860 sure they are available in opensslconf.h, by giving them names starting
13861 with `OPENSSL_` to avoid conflicts with other packages and by making
13862 sure e_os2.h will cover all platform-specific cases together with
13864 Additionally, it is now possible to define configuration/platform-
13865 specific names (called "system identities"). In the C code, these
13866 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13867 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13868 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13873 * New option -set_serial to 'req' and 'x509' this allows the serial
13874 number to use to be specified on the command line. Previously self
13875 signed certificates were hard coded with serial number 0 and the
13876 CA options of 'x509' had to use a serial number in a file which was
13881 * New options to 'ca' utility to support V2 CRL entry extensions.
13882 Currently CRL reason, invalidity date and hold instruction are
13883 supported. Add new CRL extensions to V3 code and some new objects.
13887 * New function EVP_CIPHER_CTX_set_padding() this is used to
13888 disable standard block padding (aka PKCS#5 padding) in the EVP
13889 API, which was previously mandatory. This means that the data is
13890 not padded in any way and so the total length much be a multiple
13891 of the block size, otherwise an error occurs.
13895 * Initial (incomplete) OCSP SSL support.
13899 * New function OCSP_parse_url(). This splits up a URL into its host,
13900 port and path components: primarily to parse OCSP URLs. New -url
13901 option to ocsp utility.
13905 * New nonce behavior. The return value of OCSP_check_nonce() now
13906 reflects the various checks performed. Applications can decide
13907 whether to tolerate certain situations such as an absent nonce
13908 in a response when one was present in a request: the ocsp application
13909 just prints out a warning. New function OCSP_add1_basic_nonce()
13910 this is to allow responders to include a nonce in a response even if
13911 the request is nonce-less.
13915 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13916 skipped when using openssl x509 multiple times on a single input file,
13917 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13921 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13922 set string type: to handle setting ASN1_TIME structures. Fix ca
13923 utility to correctly initialize revocation date of CRLs.
13927 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13928 the clients preferred ciphersuites and rather use its own preferences.
13929 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13930 Internet Explorer by ensuring unchanged hash method during stepup.
13931 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13935 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13936 to aes and add a new 'exist' option to print out symbols that don't
13941 * Additional options to ocsp utility to allow flags to be set and
13942 additional certificates supplied.
13946 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13947 OCSP client a number of certificate to only verify the response
13952 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13953 handle the new API. Currently only ECB, CBC modes supported. Add new
13956 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13957 Encryption Standard (AES) Ciphersuites for Transport Layer
13958 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13959 not enabled by default and were not part of the "ALL" ciphersuite
13960 alias because they were not yet official; they could be
13961 explicitly requested by specifying the "AESdraft" ciphersuite
13962 group alias. In the final release of OpenSSL 0.9.7, the group
13963 alias is called "AES" and is part of "ALL".)
13965 *Ben Laurie, Steve Henson, Bodo Moeller*
13967 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13968 request to response.
13972 * Functions for OCSP responders. OCSP_request_onereq_count(),
13973 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13974 extract information from a certificate request. OCSP_response_create()
13975 creates a response and optionally adds a basic response structure.
13976 OCSP_basic_add1_status() adds a complete single response to a basic
13977 response and returns the OCSP_SINGLERESP structure just added (to allow
13978 extensions to be included for example). OCSP_basic_add1_cert() adds a
13979 certificate to a basic response and OCSP_basic_sign() signs a basic
13980 response with various flags. New helper functions ASN1_TIME_check()
13981 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13982 (converts ASN1_TIME to GeneralizedTime).
13986 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13987 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13988 structure from a certificate. X509_pubkey_digest() digests the public_key
13989 contents: this is used in various key identifiers.
13993 * Make sk_sort() tolerate a NULL argument.
13995 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13997 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13998 passed by the function are trusted implicitly. If any of them signed the
13999 response then it is assumed to be valid and is not verified.
14003 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14004 to data. This was previously part of the PKCS7 ASN1 code. This
14005 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14006 *Steve Henson, reported by Kenneth R. Robinette
14007 <support@securenetterm.com>*
14009 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14010 routines: without these tracing memory leaks is very painful.
14011 Fix leaks in PKCS12 and PKCS7 routines.
14015 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14016 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14017 effectively meant GeneralizedTime would never be used. Now it
14018 is initialised to -1 but X509_time_adj() now has to check the value
14019 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14020 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14021 *Steve Henson, reported by Kenneth R. Robinette
14022 <support@securenetterm.com>*
14024 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14025 result in a zero length in the ASN1_INTEGER structure which was
14026 not consistent with the structure when d2i_ASN1_INTEGER() was used
14027 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14028 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14029 where it did not print out a minus for negative ASN1_INTEGER.
14033 * Add summary printout to ocsp utility. The various functions which
14034 convert status values to strings have been renamed to:
14035 OCSP_response_status_str(), OCSP_cert_status_str() and
14036 OCSP_crl_reason_str() and are no longer static. New options
14037 to verify nonce values and to disable verification. OCSP response
14038 printout format cleaned up.
14042 * Add additional OCSP certificate checks. These are those specified
14043 in RFC2560. This consists of two separate checks: the CA of the
14044 certificate being checked must either be the OCSP signer certificate
14045 or the issuer of the OCSP signer certificate. In the latter case the
14046 OCSP signer certificate must contain the OCSP signing extended key
14047 usage. This check is performed by attempting to match the OCSP
14048 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14049 in the OCSP_CERTID structures of the response.
14053 * Initial OCSP certificate verification added to OCSP_basic_verify()
14054 and related routines. This uses the standard OpenSSL certificate
14055 verify routines to perform initial checks (just CA validity) and
14056 to obtain the certificate chain. Then additional checks will be
14057 performed on the chain. Currently the root CA is checked to see
14058 if it is explicitly trusted for OCSP signing. This is used to set
14059 a root CA as a global signing root: that is any certificate that
14060 chains to that CA is an acceptable OCSP signing certificate.
14064 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14065 extensions from a separate configuration file.
14066 As when reading extensions from the main configuration file,
14067 the '-extensions ...' option may be used for specifying the
14070 *Massimiliano Pala <madwolf@comune.modena.it>*
14072 * New OCSP utility. Allows OCSP requests to be generated or
14073 read. The request can be sent to a responder and the output
14074 parsed, outputted or printed in text form. Not complete yet:
14075 still needs to check the OCSP response validity.
14079 * New subcommands for 'openssl ca':
14080 `openssl ca -status <serial>` prints the status of the cert with
14081 the given serial number (according to the index file).
14082 `openssl ca -updatedb` updates the expiry status of certificates
14085 *Massimiliano Pala <madwolf@comune.modena.it>*
14087 * New '-newreq-nodes' command option to CA.pl. This is like
14088 '-newreq', but calls 'openssl req' with the '-nodes' option
14089 so that the resulting key is not encrypted.
14091 *Damien Miller <djm@mindrot.org>*
14093 * New configuration for the GNU Hurd.
14095 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14097 * Initial code to implement OCSP basic response verify. This
14098 is currently incomplete. Currently just finds the signer's
14099 certificate and verifies the signature on the response.
14103 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14104 value of OPENSSLDIR. This is available via the new '-d' option
14105 to 'openssl version', and is also included in 'openssl version -a'.
14109 * Allowing defining memory allocation callbacks that will be given
14110 file name and line number information in additional arguments
14111 (a `const char*` and an int). The basic functionality remains, as
14112 well as the original possibility to just replace malloc(),
14113 realloc() and free() by functions that do not know about these
14114 additional arguments. To register and find out the current
14115 settings for extended allocation functions, the following
14116 functions are provided:
14118 CRYPTO_set_mem_ex_functions
14119 CRYPTO_set_locked_mem_ex_functions
14120 CRYPTO_get_mem_ex_functions
14121 CRYPTO_get_locked_mem_ex_functions
14123 These work the same way as CRYPTO_set_mem_functions and friends.
14124 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14125 extended allocation function is enabled.
14126 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14127 a conventional allocation function is enabled.
14129 *Richard Levitte, Bodo Moeller*
14131 * Finish off removing the remaining LHASH function pointer casts.
14132 There should no longer be any prototype-casting required when using
14133 the LHASH abstraction, and any casts that remain are "bugs". See
14134 the callback types and macros at the head of lhash.h for details
14135 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14139 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14140 If /dev/[u]random devices are not available or do not return enough
14141 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14143 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14144 /etc/entropy will be queried once each in this sequence, querying stops
14145 when enough entropy was collected without querying more sockets.
14149 * Change the Unix RAND_poll() variant to be able to poll several
14150 random devices, as specified by DEVRANDOM, until a sufficient amount
14151 of data has been collected. We spend at most 10 ms on each file
14152 (select timeout) and read in non-blocking mode. DEVRANDOM now
14153 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14154 (previously it was just the string "/dev/urandom"), so on typical
14155 platforms the 10 ms delay will never occur.
14156 Also separate out the Unix variant to its own file, rand_unix.c.
14157 For VMS, there's a currently-empty rand_vms.c.
14161 * Move OCSP client related routines to ocsp_cl.c. These
14162 provide utility functions which an application needing
14163 to issue a request to an OCSP responder and analyse the
14164 response will typically need: as opposed to those which an
14165 OCSP responder itself would need which will be added later.
14167 OCSP_request_sign() signs an OCSP request with an API similar
14168 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14169 response. OCSP_response_get1_basic() extracts basic response
14170 from response. OCSP_resp_find_status(): finds and extracts status
14171 information from an OCSP_CERTID structure (which will be created
14172 when the request structure is built). These are built from lower
14173 level functions which work on OCSP_SINGLERESP structures but
14174 won't normally be used unless the application wishes to examine
14175 extensions in the OCSP response for example.
14177 Replace nonce routines with a pair of functions.
14178 OCSP_request_add1_nonce() adds a nonce value and optionally
14179 generates a random value. OCSP_check_nonce() checks the
14180 validity of the nonce in an OCSP response.
14184 * Change function OCSP_request_add() to OCSP_request_add0_id().
14185 This doesn't copy the supplied OCSP_CERTID and avoids the
14186 need to free up the newly created id. Change return type
14187 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14188 This can then be used to add extensions to the request.
14189 Deleted OCSP_request_new(), since most of its functionality
14190 is now in OCSP_REQUEST_new() (and the case insensitive name
14191 clash) apart from the ability to set the request name which
14192 will be added elsewhere.
14196 * Update OCSP API. Remove obsolete extensions argument from
14197 various functions. Extensions are now handled using the new
14198 OCSP extension code. New simple OCSP HTTP function which
14199 can be used to send requests and parse the response.
14203 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14204 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14205 uses the special reorder version of SET OF to sort the attributes
14206 and reorder them to match the encoded order. This resolves a long
14207 standing problem: a verify on a PKCS7 structure just after signing
14208 it used to fail because the attribute order did not match the
14209 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14210 it uses the received order. This is necessary to tolerate some broken
14211 software that does not order SET OF. This is handled by encoding
14212 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14213 to produce the required SET OF.
14217 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14218 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14219 files to get correct declarations of the ASN.1 item variables.
14223 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14224 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14225 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14226 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14227 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14228 ASN1_ITEM and no wrapper functions.
14232 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14233 replace the old function pointer based I/O routines. Change most of
14234 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14238 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14239 lines, recognize more "algorithms" that can be deselected, and make
14240 it complain about algorithm deselection that isn't recognised.
14244 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14245 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14246 to use new functions. Add NO_ASN1_OLD which can be set to remove
14247 some old style ASN1 functions: this can be used to determine if old
14248 code will still work when these eventually go away.
14252 * New extension functions for OCSP structures, these follow the
14253 same conventions as certificates and CRLs.
14257 * New function X509V3_add1_i2d(). This automatically encodes and
14258 adds an extension. Its behaviour can be customised with various
14259 flags to append, replace or delete. Various wrappers added for
14260 certificates and CRLs.
14264 * Fix to avoid calling the underlying ASN1 print routine when
14265 an extension cannot be parsed. Correct a typo in the
14266 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14270 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14271 entries for variables.
14275 * Add functionality to `apps/openssl.c` for detecting locking
14276 problems: As the program is single-threaded, all we have
14277 to do is register a locking callback using an array for
14278 storing which locks are currently held by the program.
14282 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14283 SSL_get_ex_data_X509_STORE_idx(), which is used in
14284 ssl_verify_cert_chain() and thus can be called at any time
14285 during TLS/SSL handshakes so that thread-safety is essential.
14286 Unfortunately, the ex_data design is not at all suited
14287 for multi-threaded use, so it probably should be abolished.
14291 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14293 *Broadcom, tweaked and integrated by Geoff Thorpe*
14295 * Move common extension printing code to new function
14296 X509V3_print_extensions(). Reorganise OCSP print routines and
14297 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14301 * New function X509_signature_print() to remove duplication in some
14306 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14307 set (this was treated exactly the same as SET OF previously). This
14308 is used to reorder the STACK representing the structure to match the
14309 encoding. This will be used to get round a problem where a PKCS7
14310 structure which was signed could not be verified because the STACK
14311 order did not reflect the encoded order.
14315 * Reimplement the OCSP ASN1 module using the new code.
14319 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14320 for its ASN1 operations. The old style function pointers still exist
14321 for now but they will eventually go away.
14325 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14326 completely replaces the old ASN1 functionality with a table driven
14327 encoder and decoder which interprets an ASN1_ITEM structure describing
14328 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14329 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14330 has also been converted to the new form.
14334 * Change BN_mod_exp_recp so that negative moduli are tolerated
14335 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14336 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14337 for negative moduli.
14341 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14342 of not touching the result's sign bit.
14346 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14351 * Changed the LHASH code to use prototypes for callbacks, and created
14352 macros to declare and implement thin (optionally static) functions
14353 that provide type-safety and avoid function pointer casting for the
14354 type-specific callbacks.
14358 * Added Kerberos Cipher Suites to be used with TLS, as written in
14360 *Veers Staats <staatsvr@asc.hpc.mil>,
14361 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14363 * Reformat the FAQ so the different questions and answers can be divided
14364 in sections depending on the subject.
14368 * Have the zlib compression code load ZLIB.DLL dynamically under
14373 * New function BN_mod_sqrt for computing square roots modulo a prime
14374 (using the probabilistic Tonelli-Shanks algorithm unless
14375 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14376 be handled deterministically).
14378 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14380 * Make BN_mod_inverse faster by explicitly handling small quotients
14381 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14382 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14386 * New function BN_kronecker.
14390 * Fix BN_gcd so that it works on negative inputs; the result is
14391 positive unless both parameters are zero.
14392 Previously something reasonably close to an infinite loop was
14393 possible because numbers could be growing instead of shrinking
14394 in the implementation of Euclid's algorithm.
14398 * Fix BN_is_word() and BN_is_one() macros to take into account the
14399 sign of the number in question.
14401 Fix BN_is_word(a,w) to work correctly for w == 0.
14403 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14404 because its test if the absolute value of 'a' equals 'w'.
14405 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14406 it exists mostly for use in the implementations of BN_is_zero(),
14407 BN_is_one(), and BN_is_word().
14411 * New function BN_swap.
14415 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14416 the exponentiation functions are more likely to produce reasonable
14417 results on negative inputs.
14421 * Change BN_mod_mul so that the result is always non-negative.
14422 Previously, it could be negative if one of the factors was negative;
14423 I don't think anyone really wanted that behaviour.
14427 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14428 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14429 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14430 and add new functions:
14439 BN_mod_lshift1_quick
14441 BN_mod_lshift_quick
14443 These functions always generate non-negative results.
14445 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14446 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14448 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14449 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14450 be reduced modulo `m`.
14452 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14455 The following entry accidentally appeared in the CHANGES file
14456 distributed with OpenSSL 0.9.7. The modifications described in
14457 it do *not* apply to OpenSSL 0.9.7.
14459 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14460 was actually never needed) and in BN_mul(). The removal in BN_mul()
14461 required a small change in bn_mul_part_recursive() and the addition
14462 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14463 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14464 bn_sub_words() and bn_add_words() except they take arrays with
14470 * In 'openssl passwd', verify passwords read from the terminal
14471 unless the '-salt' option is used (which usually means that
14472 verification would just waste user's time since the resulting
14473 hash is going to be compared with some given password hash)
14474 or the new '-noverify' option is used.
14476 This is an incompatible change, but it does not affect
14477 non-interactive use of 'openssl passwd' (passwords on the command
14478 line, '-stdin' option, '-in ...' option) and thus should not
14479 cause any problems.
14483 * Remove all references to RSAref, since there's no more need for it.
14487 * Make DSO load along a path given through an environment variable
14488 (SHLIB_PATH) with shl_load().
14492 * Constify the ENGINE code as a result of BIGNUM constification.
14493 Also constify the RSA code and most things related to it. In a
14494 few places, most notable in the depth of the ASN.1 code, ugly
14495 casts back to non-const were required (to be solved at a later
14500 * Make it so the openssl application has all engines loaded by default.
14504 * Constify the BIGNUM routines a little more.
14508 * Add the following functions:
14510 ENGINE_load_cswift()
14512 ENGINE_load_atalla()
14513 ENGINE_load_nuron()
14514 ENGINE_load_builtin_engines()
14516 That way, an application can itself choose if external engines that
14517 are built-in in OpenSSL shall ever be used or not. The benefit is
14518 that applications won't have to be linked with libdl or other dso
14519 libraries unless it's really needed.
14521 Changed 'openssl engine' to load all engines on demand.
14522 Changed the engine header files to avoid the duplication of some
14523 declarations (they differed!).
14527 * 'openssl engine' can now list capabilities.
14531 * Better error reporting in 'openssl engine'.
14535 * Never call load_dh_param(NULL) in s_server.
14539 * Add engine application. It can currently list engines by name and
14540 identity, and test if they are actually available.
14544 * Improve RPM specification file by forcing symbolic linking and making
14545 sure the installed documentation is also owned by root.root.
14547 *Damien Miller <djm@mindrot.org>*
14549 * Give the OpenSSL applications more possibilities to make use of
14550 keys (public as well as private) handled by engines.
14554 * Add OCSP code that comes from CertCo.
14558 * Add VMS support for the Rijndael code.
14562 * Added untested support for Nuron crypto accelerator.
14566 * Add support for external cryptographic devices. This code was
14567 previously distributed separately as the "engine" branch.
14569 *Geoff Thorpe, Richard Levitte*
14571 * Rework the filename-translation in the DSO code. It is now possible to
14572 have far greater control over how a "name" is turned into a filename
14573 depending on the operating environment and any oddities about the
14574 different shared library filenames on each system.
14578 * Support threads on FreeBSD-elf in Configure.
14582 * Fix for SHA1 assembly problem with MASM: it produces
14583 warnings about corrupt line number information when assembling
14584 with debugging information. This is caused by the overlapping
14587 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14590 NCONF_get_number() has no error checking at all. As a replacement,
14591 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14592 promoted strongly. The old NCONF_get_number is kept around for
14593 binary backward compatibility.
14594 Make it possible for methods to load from something other than a BIO,
14595 by providing a function pointer that is given a name instead of a BIO.
14596 For example, this could be used to load configuration data from an
14601 * Fix for non blocking accept BIOs. Added new I/O special reason
14602 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14603 with non blocking I/O was not possible because no retry code was
14604 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14609 * Added the beginnings of Rijndael support.
14613 * Fix for bug in DirectoryString mask setting. Add support for
14614 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14615 to allow certificate printing to more controllable, additional
14616 'certopt' option to 'x509' to allow new printing options to be
14621 * Clean old EAY MD5 hack from e_os.h.
14625 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14627 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14628 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14630 *Joe Orton, Steve Henson*
14632 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14634 * Fix additional bug revealed by the NISCC test suite:
14636 Stop bug triggering large recursion when presented with
14637 certain ASN.1 tags ([CVE-2003-0851])
14641 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14643 * Fix various bugs revealed by running the NISCC test suite:
14645 Stop out of bounds reads in the ASN1 code when presented with
14646 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14648 If verify callback ignores invalid public key errors don't try to check
14649 certificate signature with the NULL public key.
14653 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14654 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14659 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14660 extra data after the compression methods not only for TLS 1.0
14661 but also for SSL 3.0 (as required by the specification).
14663 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14665 * Change X509_certificate_type() to mark the key as exported/exportable
14666 when it's 512 *bits* long, not 512 bytes.
14670 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14672 * Countermeasure against the Klima-Pokorny-Rosa extension of
14673 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14674 a protocol version number mismatch like a decryption error
14675 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14679 * Turn on RSA blinding by default in the default implementation
14680 to avoid a timing attack. Applications that don't want it can call
14681 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14682 They would be ill-advised to do so in most cases.
14684 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14686 * Change RSA blinding code so that it works when the PRNG is not
14687 seeded (in this case, the secret RSA exponent is abused as
14688 an unpredictable seed -- if it is not unpredictable, there
14689 is no point in blinding anyway). Make RSA blinding thread-safe
14690 by remembering the creator's thread ID in rsa->blinding and
14691 having all other threads use local one-time blinding factors
14692 (this requires more computation than sharing rsa->blinding, but
14693 avoids excessive locking; and if an RSA object is not shared
14694 between threads, blinding will still be very fast).
14698 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14700 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14701 via timing by performing a MAC computation even if incorrect
14702 block cipher padding has been found. This is a countermeasure
14703 against active attacks where the attacker has to distinguish
14704 between bad padding and a MAC verification error. ([CVE-2003-0078])
14706 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14707 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14708 Martin Vuagnoux (EPFL, Ilion)*
14710 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14712 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14713 memory from its contents. This is done with a counter that will
14714 place alternating values in each byte. This can be used to solve
14715 two issues: 1) the removal of calls to memset() by highly optimizing
14716 compilers, and 2) cleansing with other values than 0, since those can
14717 be read through on certain media, for example a swap space on disk.
14721 * Bugfix: client side session caching did not work with external caching,
14722 because the session->cipher setting was not restored when reloading
14723 from the external cache. This problem was masked, when
14724 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14725 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14729 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14730 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14732 *Zeev Lieber <zeev-l@yahoo.com>*
14734 * Undo an undocumented change introduced in 0.9.6e which caused
14735 repeated calls to OpenSSL_add_all_ciphers() and
14736 OpenSSL_add_all_digests() to be ignored, even after calling
14741 * Change the default configuration reader to deal with last line not
14742 being properly terminated.
14746 * Change X509_NAME_cmp() so it applies the special rules on handling
14747 DN values that are of type PrintableString, as well as RDNs of type
14748 emailAddress where the value has the type ia5String.
14750 *stefank@valicert.com via Richard Levitte*
14752 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14753 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14754 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14755 the bitwise-OR of the two for use by the majority of applications
14756 wanting this behaviour, and update the docs. The documented
14757 behaviour and actual behaviour were inconsistent and had been
14758 changing anyway, so this is more a bug-fix than a behavioural
14761 *Geoff Thorpe, diagnosed by Nadav Har'El*
14763 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14764 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14768 * Fix initialization code race conditions in
14769 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14770 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14771 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14772 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14773 ssl2_get_cipher_by_char(),
14774 ssl3_get_cipher_by_char().
14776 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14778 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14779 the cached sessions are flushed, as the remove_cb() might use ex_data
14780 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14781 (see [openssl.org #212]).
14783 *Geoff Thorpe, Lutz Jaenicke*
14785 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14786 length, instead of the encoding length to d2i_ASN1_OBJECT.
14790 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14792 * [In 0.9.6g-engine release:]
14793 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14795 *Lynn Gazis <lgazis@rainbow.com>*
14797 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14799 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14800 and get fix the header length calculation.
14801 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14802 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14804 * Use proper error handling instead of 'assertions' in buffer
14805 overflow checks added in 0.9.6e. This prevents DoS (the
14806 assertions could call abort()).
14808 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14810 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14812 * Add various sanity checks to asn1_get_length() to reject
14813 the ASN1 length bytes if they exceed sizeof(long), will appear
14814 negative or the content length exceeds the length of the
14817 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14819 * Fix cipher selection routines: ciphers without encryption had no flags
14820 for the cipher strength set and where therefore not handled correctly
14821 by the selection routines (PR #130).
14825 * Fix EVP_dsa_sha macro.
14830 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14831 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14832 that was added in OpenSSL 0.9.6d.
14834 As the countermeasure turned out to be incompatible with some
14835 broken SSL implementations, the new option is part of SSL_OP_ALL.
14836 SSL_OP_ALL is usually employed when compatibility with weird SSL
14837 implementations is desired (e.g. '-bugs' option to 's_client' and
14838 's_server'), so the new option is automatically set in many
14843 * Changes in security patch:
14845 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14846 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14847 Air Force Materiel Command, USAF, under agreement number
14850 * Add various sanity checks to asn1_get_length() to reject
14851 the ASN1 length bytes if they exceed sizeof(long), will appear
14852 negative or the content length exceeds the length of the
14853 supplied buffer. ([CVE-2002-0659])
14855 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14857 * Assertions for various potential buffer overflows, not known to
14858 happen in practice.
14860 *Ben Laurie (CHATS)*
14862 * Various temporary buffers to hold ASCII versions of integers were
14863 too small for 64 bit platforms. ([CVE-2002-0655])
14864 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14866 * Remote buffer overflow in SSL3 protocol - an attacker could
14867 supply an oversized session ID to a client. ([CVE-2002-0656])
14869 *Ben Laurie (CHATS)*
14871 * Remote buffer overflow in SSL2 protocol - an attacker could
14872 supply an oversized client master key. ([CVE-2002-0656])
14874 *Ben Laurie (CHATS)*
14876 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14878 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14879 encoded as NULL) with id-dsa-with-sha1.
14881 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14883 * Check various `X509_...()` return values in `apps/req.c`.
14885 *Nils Larsch <nla@trustcenter.de>*
14887 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14888 an end-of-file condition would erroneously be flagged, when the CRLF
14889 was just at the end of a processed block. The bug was discovered when
14890 processing data through a buffering memory BIO handing the data to a
14891 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14892 <ptsekov@syntrex.com> and Nedelcho Stanev.
14896 * Implement a countermeasure against a vulnerability recently found
14897 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14898 before application data chunks to avoid the use of known IVs
14899 with data potentially chosen by the attacker.
14903 * Fix length checks in ssl3_get_client_hello().
14907 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14908 to prevent ssl3_read_internal() from incorrectly assuming that
14909 ssl3_read_bytes() found application data while handshake
14910 processing was enabled when in fact s->s3->in_read_app_data was
14911 merely automatically cleared during the initial handshake.
14913 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14915 * Fix object definitions for Private and Enterprise: they were not
14916 recognized in their shortname (=lowercase) representation. Extend
14917 obj_dat.pl to issue an error when using undefined keywords instead
14918 of silently ignoring the problem (Svenning Sorensen
14919 <sss@sss.dnsalias.net>).
14923 * Fix DH_generate_parameters() so that it works for 'non-standard'
14924 generators, i.e. generators other than 2 and 5. (Previously, the
14925 code did not properly initialise the 'add' and 'rem' values to
14926 BN_generate_prime().)
14928 In the new general case, we do not insist that 'generator' is
14929 actually a primitive root: This requirement is rather pointless;
14930 a generator of the order-q subgroup is just as good, if not
14935 * Map new X509 verification errors to alerts. Discovered and submitted by
14936 Tom Wu <tom@arcot.com>.
14940 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14941 returning non-zero before the data has been completely received
14942 when using non-blocking I/O.
14944 *Bodo Moeller; problem pointed out by John Hughes*
14946 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14948 *Ben Laurie, Lutz Jaenicke*
14950 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14951 Yoram Zahavi <YoramZ@gilian.com>).
14955 * Add information about CygWin 1.3 and on, and preserve proper
14956 configuration for the versions before that.
14958 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14960 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14961 check whether we deal with a copy of a session and do not delete from
14962 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14963 <izhar@checkpoint.com>.
14967 * Do not store session data into the internal session cache, if it
14968 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14969 flag is set). Proposed by Aslam <aslam@funk.com>.
14973 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14978 * [In 0.9.6d-engine release:]
14979 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14981 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14983 * Add the configuration target linux-s390x.
14985 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14987 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14988 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14989 variable as an indication that a ClientHello message has been
14990 received. As the flag value will be lost between multiple
14991 invocations of ssl3_accept when using non-blocking I/O, the
14992 function may not be aware that a handshake has actually taken
14993 place, thus preventing a new session from being added to the
14996 To avoid this problem, we now set s->new_session to 2 instead of
14997 using a local variable.
14999 *Lutz Jaenicke, Bodo Moeller*
15001 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15002 if the SSL_R_LENGTH_MISMATCH error is detected.
15004 *Geoff Thorpe, Bodo Moeller*
15006 * New 'shared_ldflag' column in Configure platform table.
15010 * Fix EVP_CIPHER_mode macro.
15012 *"Dan S. Camper" <dan@bti.net>*
15014 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15015 type, we must throw them away by setting rr->length to 0.
15017 *D P Chang <dpc@qualys.com>*
15019 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15021 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15022 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15023 worked incorrectly for those cases where range = `10..._2` and
15024 `3*range` is two bits longer than range.)
15028 * Only add signing time to PKCS7 structures if it is not already
15033 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15034 OBJ_ld_ce should be OBJ_id_ce.
15035 Also some ip-pda OIDs in crypto/objects/objects.txt were
15036 incorrect (cf. RFC 3039).
15038 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15040 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15041 returns early because it has nothing to do.
15043 *Andy Schneider <andy.schneider@bjss.co.uk>*
15045 * [In 0.9.6c-engine release:]
15046 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15048 *Andy Schneider <andy.schneider@bjss.co.uk>*
15050 * [In 0.9.6c-engine release:]
15051 Add support for Cryptographic Appliance's keyserver technology.
15052 (Use engine 'keyclient')
15054 *Cryptographic Appliances and Geoff Thorpe*
15056 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15057 is called via tools/c89.sh because arguments have to be
15058 rearranged (all '-L' options must appear before the first object
15061 *Richard Shapiro <rshapiro@abinitio.com>*
15063 * [In 0.9.6c-engine release:]
15064 Add support for Broadcom crypto accelerator cards, backported
15067 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15069 * [In 0.9.6c-engine release:]
15070 Add support for SureWare crypto accelerator cards from
15071 Baltimore Technologies. (Use engine 'sureware')
15073 *Baltimore Technologies and Mark Cox*
15075 * [In 0.9.6c-engine release:]
15076 Add support for crypto accelerator cards from Accelerated
15077 Encryption Processing, www.aep.ie. (Use engine 'aep')
15079 *AEP Inc. and Mark Cox*
15081 * Add a configuration entry for gcc on UnixWare.
15083 *Gary Benson <gbenson@redhat.com>*
15085 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15086 messages are stored in a single piece (fixed-length part and
15087 variable-length part combined) and fix various bugs found on the way.
15091 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15092 instead. BIO_gethostbyname() does not know what timeouts are
15093 appropriate, so entries would stay in cache even when they have
15095 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15097 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15098 faced with a pathologically small ClientHello fragment that does
15099 not contain client_version: Instead of aborting with an error,
15100 simply choose the highest available protocol version (i.e.,
15101 TLS 1.0 unless it is disabled). In practice, ClientHello
15102 messages are never sent like this, but this change gives us
15103 strictly correct behaviour at least for TLS.
15107 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15108 never resets s->method to s->ctx->method when called from within
15109 one of the SSL handshake functions.
15111 *Bodo Moeller; problem pointed out by Niko Baric*
15113 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15114 (sent using the client's version number) if client_version is
15115 smaller than the protocol version in use. Also change
15116 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15117 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15118 the client will at least see that alert.
15122 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15127 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15128 client receives HelloRequest while in a handshake.
15130 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15132 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15133 should end in 'break', not 'goto end' which circumvents various
15134 cleanups done in state SSL_ST_OK. But session related stuff
15135 must be disabled for SSL_ST_OK in the case that we just sent a
15138 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15139 before just sending a HelloRequest.
15141 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15143 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15144 reveal whether illegal block cipher padding was found or a MAC
15145 verification error occurred. (Neither SSLerr() codes nor alerts
15146 are directly visible to potential attackers, but the information
15147 may leak via logfiles.)
15149 Similar changes are not required for the SSL 2.0 implementation
15150 because the number of padding bytes is sent in clear for SSL 2.0,
15151 and the extra bytes are just ignored. However ssl/s2_pkt.c
15152 failed to verify that the purported number of padding bytes is in
15157 * Add OpenUNIX-8 support including shared libraries
15158 (Boyd Lynn Gerber <gerberb@zenez.com>).
15162 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15163 'wristwatch attack' using huge encoding parameters (cf.
15164 James H. Manger's CRYPTO 2001 paper). Note that the
15165 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15166 encoding parameters and hence was not vulnerable.
15170 * BN_sqr() bug fix.
15172 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15174 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15175 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15176 followed by modular reduction.
15178 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15180 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15181 equivalent based on BN_pseudo_rand() instead of BN_rand().
15185 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15186 This function was broken, as the check for a new client hello message
15187 to handle SGC did not allow these large messages.
15188 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15192 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15196 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15197 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15201 * Rework the configuration and shared library support for Tru64 Unix.
15202 The configuration part makes use of modern compiler features and
15203 still retains old compiler behavior for those that run older versions
15204 of the OS. The shared library support part includes a variant that
15205 uses the RPATH feature, and is available through the special
15206 configuration target "alpha-cc-rpath", which will never be selected
15209 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15211 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15212 with the same message size as in ssl3_get_certificate_request().
15213 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15214 messages might inadvertently be reject as too long.
15216 *Petr Lampa <lampa@fee.vutbr.cz>*
15218 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15222 * Modified SSL library such that the verify_callback that has been set
15223 specifically for an SSL object with SSL_set_verify() is actually being
15224 used. Before the change, a verify_callback set with this function was
15225 ignored and the verify_callback() set in the SSL_CTX at the time of
15226 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15227 to allow the necessary settings.
15231 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15232 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15233 done automatically (in contradiction to the requirements of the C
15234 standard). This made problems when used from OpenSSH.
15238 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15239 dh->length and always used
15241 BN_rand_range(priv_key, dh->p).
15243 BN_rand_range() is not necessary for Diffie-Hellman, and this
15244 specific range makes Diffie-Hellman unnecessarily inefficient if
15245 dh->length (recommended exponent length) is much smaller than the
15246 length of dh->p. We could use BN_rand_range() if the order of
15247 the subgroup was stored in the DH structure, but we only have
15252 BN_rand(priv_key, l, ...)
15254 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15261 RSA_eay_public_encrypt
15262 RSA_eay_private_decrypt
15263 RSA_eay_private_encrypt (signing)
15264 RSA_eay_public_decrypt (signature verification)
15266 (default implementations for RSA_public_encrypt,
15267 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15268 always reject numbers >= n.
15272 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15273 to synchronize access to 'locking_thread'. This is necessary on
15274 systems where access to 'locking_thread' (an 'unsigned long'
15275 variable) is not atomic.
15279 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15280 *before* setting the 'crypto_lock_rand' flag. The previous code had
15281 a race condition if 0 is a valid thread ID.
15283 *Travis Vitek <vitek@roguewave.com>*
15285 * Add support for shared libraries under Irix.
15287 *Albert Chin-A-Young <china@thewrittenword.com>*
15289 * Add configuration option to build on Linux on both big-endian and
15290 little-endian MIPS.
15292 *Ralf Baechle <ralf@uni-koblenz.de>*
15294 * Add the possibility to create shared libraries on HP-UX.
15298 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15300 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15301 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15302 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15303 PRNG state recovery was possible based on the output of
15304 one PRNG request appropriately sized to gain knowledge on
15305 'md' followed by enough consecutive 1-byte PRNG requests
15306 to traverse all of 'state'.
15308 1. When updating 'md_local' (the current thread's copy of 'md')
15309 during PRNG output generation, hash all of the previous
15310 'md_local' value, not just the half used for PRNG output.
15312 2. Make the number of bytes from 'state' included into the hash
15313 independent from the number of PRNG bytes requested.
15315 The first measure alone would be sufficient to avoid
15316 Markku-Juhani's attack. (Actually it had never occurred
15317 to me that the half of 'md_local' used for chaining was the
15318 half from which PRNG output bytes were taken -- I had always
15319 assumed that the secret half would be used.) The second
15320 measure makes sure that additional data from 'state' is never
15321 mixed into 'md_local' in small portions; this heuristically
15322 further strengthens the PRNG.
15326 * Fix crypto/bn/asm/mips3.s.
15330 * When only the key is given to "enc", the IV is undefined. Print out
15331 an error message in this case.
15335 * Handle special case when X509_NAME is empty in X509 printing routines.
15339 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15340 positive and less than q.
15344 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15345 used: it isn't thread safe and the add_lock_callback should handle
15348 *Paul Rose <Paul.Rose@bridge.com>*
15350 * Verify that incoming data obeys the block size in
15351 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15357 *Ulf Möller, Bodo Möller*
15359 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15360 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15361 when fixing the server behaviour for backwards-compatible 'client
15362 hello' messages. (Note that the attack is impractical against
15363 SSL 3.0 and TLS 1.0 anyway because length and version checking
15364 means that the probability of guessing a valid ciphertext is
15365 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15368 Before 0.9.5, the countermeasure (hide the error by generating a
15369 random 'decryption result') did not work properly because
15370 ERR_clear_error() was missing, meaning that SSL_get_error() would
15371 detect the supposedly ignored error.
15373 Both problems are now fixed.
15377 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15378 (previously it was 1024).
15382 * Fix for compatibility mode trust settings: ignore trust settings
15383 unless some valid trust or reject settings are present.
15387 * Fix for blowfish EVP: its a variable length cipher.
15391 * Fix various bugs related to DSA S/MIME verification. Handle missing
15392 parameters in DSA public key structures and return an error in the
15393 DSA routines if parameters are absent.
15397 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15398 in the current directory if neither $RANDFILE nor $HOME was set.
15399 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15400 caused some confusion to Windows users who haven't defined $HOME.
15401 Thus RAND_file_name() is changed again: e_os.h can define a
15402 DEFAULT_HOME, which will be used if $HOME is not set.
15403 For Windows, we use "C:"; on other platforms, we still require
15404 environment variables.
15406 * Move 'if (!initialized) RAND_poll()' into regions protected by
15407 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15408 having multiple threads call RAND_poll() concurrently.
15412 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15413 combination of a flag and a thread ID variable.
15414 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15415 flag), *other* threads can enter ssleay_add_bytes without obeying
15416 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15417 that they do not hold after the first thread unsets add_do_not_lock).
15421 * Change bctest again: '-x' expressions are not available in all
15422 versions of 'test'.
15426 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15428 * Fix a couple of memory leaks in PKCS7_dataDecode()
15430 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15432 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15433 the default extension for executables, if any. Also, make the perl
15434 scripts that use symlink() to test if it really exists and use "cp"
15435 if it doesn't. All this made OpenSSL compilable and installable in
15440 * Fix for asn1_GetSequence() for indefinite length constructed data.
15441 If SEQUENCE is length is indefinite just set c->slen to the total
15442 amount of data available.
15444 *Steve Henson, reported by shige@FreeBSD.org*
15446 *This change does not apply to 0.9.7.*
15448 * Change bctest to avoid here-documents inside command substitution
15449 (workaround for FreeBSD /bin/sh bug).
15450 For compatibility with Ultrix, avoid shell functions (introduced
15451 in the bctest version that searches along $PATH).
15455 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15456 with des_encrypt() defined on some operating systems, like Solaris
15461 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15462 On the Importance of Eliminating Errors in Cryptographic
15463 Computations, J. Cryptology 14 (2001) 2, 101-119,
15464 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15468 * MIPS assembler BIGNUM division bug fix.
15472 * Disabled incorrect Alpha assembler code.
15476 * Fix PKCS#7 decode routines so they correctly update the length
15477 after reading an EOC for the EXPLICIT tag.
15481 *This change does not apply to 0.9.7.*
15483 * Fix bug in PKCS#12 key generation routines. This was triggered
15484 if a 3DES key was generated with a 0 initial byte. Include
15485 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15486 (but broken) behaviour.
15490 * Enhance bctest to search for a working bc along $PATH and print
15493 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15495 * Fix memory leaks in err.c: free err_data string if necessary;
15496 don't write to the wrong index in ERR_set_error_data.
15500 * Implement ssl23_peek (analogous to ssl23_read), which previously
15505 * Replace rdtsc with `_emit` statements for VC++ version 5.
15507 *Jeremy Cooper <jeremy@baymoo.org>*
15509 * Make it possible to reuse SSLv2 sessions.
15513 * In copy_email() check for >= 0 as a return value for
15514 X509_NAME_get_index_by_NID() since 0 is a valid index.
15516 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15518 * Avoid coredump with unsupported or invalid public keys by checking if
15519 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15520 PKCS7_verify() fails with non detached data.
15524 * Don't use getenv in library functions when run as setuid/setgid.
15525 New function OPENSSL_issetugid().
15529 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15530 due to incorrect handling of multi-threading:
15532 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15534 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15536 3. Count how many times MemCheck_off() has been called so that
15537 nested use can be treated correctly. This also avoids
15538 inband-signalling in the previous code (which relied on the
15539 assumption that thread ID 0 is impossible).
15543 * Add "-rand" option also to s_client and s_server.
15547 * Fix CPU detection on Irix 6.x.
15548 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15549 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15551 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15556 *This change does not apply to 0.9.7.*
15558 * Use the cached encoding of an X509_NAME structure rather than
15559 copying it. This is apparently the reason for the libsafe "errors"
15560 but the code is actually correct.
15564 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15565 Bleichenbacher's DSA attack.
15566 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15567 to be set and top=0 forces the highest bit to be set; top=-1 is new
15568 and leaves the highest bit random.
15570 *Ulf Moeller, Bodo Moeller*
15572 * In the `NCONF_...`-based implementations for `CONF_...` queries
15573 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15574 a temporary CONF structure with the data component set to NULL
15575 (which gives segmentation faults in lh_retrieve).
15576 Instead, use NULL for the CONF pointer in CONF_get_string and
15577 CONF_get_number (which may use environment variables) and directly
15578 return NULL from CONF_get_section.
15582 * Fix potential buffer overrun for EBCDIC.
15586 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15587 keyUsage if basicConstraints absent for a CA.
15591 * Make SMIME_write_PKCS7() write mail header values with a format that
15592 is more generally accepted (no spaces before the semicolon), since
15593 some programs can't parse those values properly otherwise. Also make
15594 sure BIO's that break lines after each write do not create invalid
15599 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15600 macros previously used would not encode an empty SEQUENCE OF
15601 and break the signature.
15605 *This change does not apply to 0.9.7.*
15607 * Zero the premaster secret after deriving the master secret in
15612 * Add some EVP_add_digest_alias registrations (as found in
15613 OpenSSL_add_all_digests()) to SSL_library_init()
15614 aka OpenSSL_add_ssl_algorithms(). This provides improved
15615 compatibility with peers using X.509 certificates
15616 with unconventional AlgorithmIdentifier OIDs.
15620 * Fix for Irix with NO_ASM.
15622 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15624 * ./config script fixes.
15626 *Ulf Moeller, Richard Levitte*
15628 * Fix 'openssl passwd -1'.
15632 * Change PKCS12_key_gen_asc() so it can cope with non null
15633 terminated strings whose length is passed in the passlen
15634 parameter, for example from PEM callbacks. This was done
15635 by adding an extra length parameter to asc2uni().
15637 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15639 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15640 call failed, free the DSA structure.
15644 * Fix to uni2asc() to cope with zero length Unicode strings.
15645 These are present in some PKCS#12 files.
15649 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15650 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15651 when writing a 32767 byte record.
15653 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15655 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15656 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15658 (RSA objects have a reference count access to which is protected
15659 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15660 so they are meant to be shared between threads.)
15661 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15662 "Reddie, Steven" <Steven.Reddie@ca.com>*
15664 * Fix a deadlock in CRYPTO_mem_leaks().
15668 * Use better test patterns in bntest.
15672 * rand_win.c fix for Borland C.
15676 * BN_rshift bugfix for n == 0.
15680 * Add a 'bctest' script that checks for some known 'bc' bugs
15681 so that 'make test' does not abort just because 'bc' is broken.
15685 * Store verify_result within SSL_SESSION also for client side to
15686 avoid potential security hole. (Re-used sessions on the client side
15687 always resulted in verify_result==X509_V_OK, not using the original
15688 result of the server certificate verification.)
15692 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15693 SSL3_RT_APPLICATION_DATA, return 0.
15694 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15699 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15700 releases, have been re-implemented by renaming the previous
15701 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15702 and ssl3_read_internal, respectively, and adding 'peek' parameters
15703 to them. The new ssl[23]_{read,peek} functions are calls to
15704 ssl[23]_read_internal with the 'peek' flag set appropriately.
15705 A 'peek' parameter has also been added to ssl3_read_bytes, which
15706 does the actual work for ssl3_read_internal.
15710 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15711 the method-specific "init()" handler. Also clean up ex_data after
15712 calling the method-specific "finish()" handler. Previously, this was
15713 happening the other way round.
15717 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15718 The previous value, 12, was not always sufficient for BN_mod_exp().
15722 * Make sure that shared libraries get the internal name engine with
15723 the full version number and not just 0. This should mark the
15724 shared libraries as not backward compatible. Of course, this should
15725 be changed again when we can guarantee backward binary compatibility.
15729 * Fix typo in get_cert_by_subject() in by_dir.c
15731 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15733 * Rework the system to generate shared libraries:
15735 - Make note of the expected extension for the shared libraries and
15736 if there is a need for symbolic links from for example libcrypto.so.0
15737 to libcrypto.so.0.9.7. There is extended info in Configure for
15740 - Make as few rebuilds of the shared libraries as possible.
15742 - Still avoid linking the OpenSSL programs with the shared libraries.
15744 - When installing, install the shared libraries separately from the
15749 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15751 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15752 and not in SSL_clear because the latter is also used by the
15753 accept/connect functions; previously, the settings made by
15754 SSL_set_read_ahead would be lost during the handshake.
15756 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15758 * Correct util/mkdef.pl to be selective about disabled algorithms.
15759 Previously, it would create entries for disabled algorithms no
15764 * Added several new manual pages for SSL_* function.
15768 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15770 * In ssl23_get_client_hello, generate an error message when faced
15771 with an initial SSL 3.0/TLS record that is too small to contain the
15772 first two bytes of the ClientHello message, i.e. client_version.
15773 (Note that this is a pathologic case that probably has never happened
15774 in real life.) The previous approach was to use the version number
15775 from the record header as a substitute; but our protocol choice
15776 should not depend on that one because it is not authenticated
15777 by the Finished messages.
15781 * More robust randomness gathering functions for Windows.
15783 *Jeffrey Altman <jaltman@columbia.edu>*
15785 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15786 not set then we don't setup the error code for issuer check errors
15787 to avoid possibly overwriting other errors which the callback does
15788 handle. If an application does set the flag then we assume it knows
15789 what it is doing and can handle the new informational codes
15794 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15795 a general "ANY" type, as such it should be able to decode anything
15796 including tagged types. However it didn't check the class so it would
15797 wrongly interpret tagged types in the same way as their universal
15798 counterpart and unknown types were just rejected. Changed so that the
15799 tagged and unknown types are handled in the same way as a SEQUENCE:
15800 that is the encoding is stored intact. There is also a new type
15801 "V_ASN1_OTHER" which is used when the class is not universal, in this
15802 case we have no idea what the actual type is so we just lump them all
15807 * On VMS, stdout may very well lead to a file that is written to
15808 in a record-oriented fashion. That means that every write() will
15809 write a separate record, which will be read separately by the
15810 programs trying to read from it. This can be very confusing.
15812 The solution is to put a BIO filter in the way that will buffer
15813 text until a linefeed is reached, and then write everything a
15814 line at a time, so every record written will be an actual line,
15815 not chunks of lines and not (usually doesn't happen, but I've
15816 seen it once) several lines in one record. BIO_f_linebuffer() is
15819 Currently, it's a VMS-only method, because that's where it has
15820 been tested well enough.
15824 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15825 it can return incorrect results.
15826 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15827 but it was in 0.9.6-beta[12].)
15831 * Disable the check for content being present when verifying detached
15832 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15833 include zero length content when signing messages.
15837 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15838 BIO_ctrl (for BIO pairs).
15842 * Add DSO method for VMS.
15846 * Bug fix: Montgomery multiplication could produce results with the
15851 * Add RPM specification openssl.spec and modify it to build three
15852 packages. The default package contains applications, application
15853 documentation and run-time libraries. The devel package contains
15854 include files, static libraries and function documentation. The
15855 doc package contains the contents of the doc directory. The original
15856 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15860 * Add a large number of documentation files for many SSL routines.
15862 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15864 * Add a configuration entry for Sony News 4.
15866 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15868 * Don't set the two most significant bits to one when generating a
15869 random number < q in the DSA library.
15873 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15874 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15875 the underlying transport is blocking) if a handshake took place.
15876 (The default behaviour is needed by applications such as s_client
15877 and s_server that use select() to determine when to use SSL_read;
15878 but for applications that know in advance when to expect data, it
15879 just makes things more complicated.)
15883 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15888 * Add a few more EBCDIC conditionals that make `req` and `x509`
15889 work better on such systems.
15891 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15893 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15894 Update PKCS12_parse() so it copies the friendlyName and the
15895 keyid to the certificates aux info.
15899 * Fix bug in PKCS7_verify() which caused an infinite loop
15900 if there was more than one signature.
15902 *Sven Uszpelkat <su@celocom.de>*
15904 * Major change in util/mkdef.pl to include extra information
15905 about each symbol, as well as presenting variables as well
15906 as functions. This change means that there's n more need
15907 to rebuild the .num files when some algorithms are excluded.
15911 * Allow the verify time to be set by an application,
15912 rather than always using the current time.
15916 * Phase 2 verify code reorganisation. The certificate
15917 verify code now looks up an issuer certificate by a
15918 number of criteria: subject name, authority key id
15919 and key usage. It also verifies self signed certificates
15920 by the same criteria. The main comparison function is
15921 X509_check_issued() which performs these checks.
15923 Lot of changes were necessary in order to support this
15924 without completely rewriting the lookup code.
15926 Authority and subject key identifier are now cached.
15928 The LHASH 'certs' is X509_STORE has now been replaced
15929 by a STACK_OF(X509_OBJECT). This is mainly because an
15930 LHASH can't store or retrieve multiple objects with
15931 the same hash value.
15933 As a result various functions (which were all internal
15934 use only) have changed to handle the new X509_STORE
15935 structure. This will break anything that messed round
15936 with X509_STORE internally.
15938 The functions X509_STORE_add_cert() now checks for an
15939 exact match, rather than just subject name.
15941 The X509_STORE API doesn't directly support the retrieval
15942 of multiple certificates matching a given criteria, however
15943 this can be worked round by performing a lookup first
15944 (which will fill the cache with candidate certificates)
15945 and then examining the cache for matches. This is probably
15946 the best we can do without throwing out X509_LOOKUP
15947 entirely (maybe later...).
15949 The X509_VERIFY_CTX structure has been enhanced considerably.
15951 All certificate lookup operations now go via a get_issuer()
15952 callback. Although this currently uses an X509_STORE it
15953 can be replaced by custom lookups. This is a simple way
15954 to bypass the X509_STORE hackery necessary to make this
15955 work and makes it possible to use more efficient techniques
15956 in future. A very simple version which uses a simple
15957 STACK for its trusted certificate store is also provided
15958 using X509_STORE_CTX_trusted_stack().
15960 The verify_cb() and verify() callbacks now have equivalents
15961 in the X509_STORE_CTX structure.
15963 X509_STORE_CTX also has a 'flags' field which can be used
15964 to customise the verify behaviour.
15968 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15969 excludes S/MIME capabilities.
15973 * When a certificate request is read in keep a copy of the
15974 original encoding of the signed data and use it when outputting
15975 again. Signatures then use the original encoding rather than
15976 a decoded, encoded version which may cause problems if the
15977 request is improperly encoded.
15981 * For consistency with other BIO_puts implementations, call
15982 buffer_write(b, ...) directly in buffer_puts instead of calling
15985 In BIO_puts, increment b->num_write as in BIO_write.
15987 *Peter.Sylvester@EdelWeb.fr*
15989 * Fix BN_mul_word for the case where the word is 0. (We have to use
15990 BN_zero, we may not return a BIGNUM with an array consisting of
15991 words set to zero.)
15995 * Avoid calling abort() from within the library when problems are
15996 detected, except if preprocessor symbols have been defined
15997 (such as REF_CHECK, BN_DEBUG etc.).
16001 * New openssl application 'rsautl'. This utility can be
16002 used for low-level RSA operations. DER public key
16003 BIO/fp routines also added.
16007 * New Configure entry and patches for compiling on QNX 4.
16009 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16011 * A demo state-machine implementation was sponsored by
16012 Nuron (<http://www.nuron.com/>) and is now available in
16013 demos/state_machine.
16017 * New options added to the 'dgst' utility for signature
16018 generation and verification.
16022 * Unrecognized PKCS#7 content types are now handled via a
16023 catch all ASN1_TYPE structure. This allows unsupported
16024 types to be stored as a "blob" and an application can
16025 encode and decode it manually.
16029 * Fix various signed/unsigned issues to make a_strex.c
16030 compile under VC++.
16032 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16034 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16035 length if passed a buffer. ASN1_INTEGER_to_BN failed
16036 if passed a NULL BN and its argument was negative.
16038 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16040 * Modification to PKCS#7 encoding routines to output definite
16041 length encoding. Since currently the whole structures are in
16042 memory there's not real point in using indefinite length
16043 constructed encoding. However if OpenSSL is compiled with
16044 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16048 * Added BIO_vprintf() and BIO_vsnprintf().
16052 * Added more prefixes to parse for in the strings written
16053 through a logging bio, to cover all the levels that are available
16054 through syslog. The prefixes are now:
16056 PANIC, EMERG, EMR => LOG_EMERG
16057 ALERT, ALR => LOG_ALERT
16058 CRIT, CRI => LOG_CRIT
16059 ERROR, ERR => LOG_ERR
16060 WARNING, WARN, WAR => LOG_WARNING
16061 NOTICE, NOTE, NOT => LOG_NOTICE
16062 INFO, INF => LOG_INFO
16063 DEBUG, DBG => LOG_DEBUG
16065 and as before, if none of those prefixes are present at the
16066 beginning of the string, LOG_ERR is chosen.
16068 On Win32, the `LOG_*` levels are mapped according to this:
16070 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16071 LOG_WARNING => EVENTLOG_WARNING_TYPE
16072 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16076 * Made it possible to reconfigure with just the configuration
16077 argument "reconf" or "reconfigure". The command line arguments
16078 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16079 and are retrieved from there when reconfiguring.
16085 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16087 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16091 * The obj_dat.pl script was messing up the sorting of object
16092 names. The reason was that it compared the quoted version
16093 of strings as a result "OCSP" > "OCSP Signing" because
16094 " > SPACE. Changed script to store unquoted versions of
16095 names and add quotes on output. It was also omitting some
16096 names from the lookup table if they were given a default
16097 value (that is if SN is missing it is given the same
16098 value as LN and vice versa), these are now added on the
16099 grounds that if an object has a name we should be able to
16100 look it up. Finally added warning output when duplicate
16101 short or long names are found.
16105 * Changes needed for Tandem NSK.
16107 *Scott Uroff <scott@xypro.com>*
16109 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16110 RSA_padding_check_SSLv23(), special padding was never detected
16111 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16112 version rollback attacks was not effective.
16114 In s23_clnt.c, don't use special rollback-attack detection padding
16115 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16116 client; similarly, in s23_srvr.c, don't do the rollback check if
16117 SSL 2.0 is the only protocol enabled in the server.
16121 * Make it possible to get hexdumps of unprintable data with 'openssl
16122 asn1parse'. By implication, the functions ASN1_parse_dump() and
16123 BIO_dump_indent() are added.
16127 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16128 these print out strings and name structures based on various
16129 flags including RFC2253 support and proper handling of
16130 multibyte characters. Added options to the 'x509' utility
16131 to allow the various flags to be set.
16135 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16136 Also change the functions X509_cmp_current_time() and
16137 X509_gmtime_adj() work with an ASN1_TIME structure,
16138 this will enable certificates using GeneralizedTime in validity
16139 dates to be checked.
16143 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16144 negative public key encodings) on by default,
16145 NO_NEG_PUBKEY_BUG can be set to disable it.
16149 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16150 content octets. An i2c_ASN1_OBJECT is unnecessary because
16151 the encoding can be trivially obtained from the structure.
16155 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16156 not read locks (`CRYPTO_r_[un]lock`).
16160 * A first attempt at creating official support for shared
16161 libraries through configuration. I've kept it so the
16162 default is static libraries only, and the OpenSSL programs
16163 are always statically linked for now, but there are
16164 preparations for dynamic linking in place.
16165 This has been tested on Linux and Tru64.
16169 * Randomness polling function for Win9x, as described in:
16170 Peter Gutmann, Software Generation of Practically Strong
16175 * Fix so PRNG is seeded in req if using an already existing
16180 * New options to smime application. -inform and -outform
16181 allow alternative formats for the S/MIME message including
16182 PEM and DER. The -content option allows the content to be
16183 specified separately. This should allow things like Netscape
16184 form signing output easier to verify.
16188 * Fix the ASN1 encoding of tags using the 'long form'.
16192 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16193 STRING types. These convert content octets to and from the
16194 underlying type. The actual tag and length octets are
16195 already assumed to have been read in and checked. These
16196 are needed because all other string types have virtually
16197 identical handling apart from the tag. By having versions
16198 of the ASN1 functions that just operate on content octets
16199 IMPLICIT tagging can be handled properly. It also allows
16200 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16201 and ASN1_INTEGER are identical apart from the tag.
16205 * Change the handling of OID objects as follows:
16207 - New object identifiers are inserted in objects.txt, following
16208 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16209 - objects.pl is used to process obj_mac.num and create a new
16211 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16214 This is currently kind of a hack, and the perl code in objects.pl
16215 isn't very elegant, but it works as I intended. The simplest way
16216 to check that it worked correctly is to look in obj_dat.h and
16217 check the array nid_objs and make sure the objects haven't moved
16218 around (this is important!). Additions are OK, as well as
16219 consistent name changes.
16223 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16227 * Addition of the command line parameter '-rand file' to 'openssl req'.
16228 The given file adds to whatever has already been seeded into the
16229 random pool through the RANDFILE configuration file option or
16230 environment variable, or the default random state file.
16234 * mkstack.pl now sorts each macro group into lexical order.
16235 Previously the output order depended on the order the files
16236 appeared in the directory, resulting in needless rewriting
16241 * Patches to make OpenSSL compile under Win32 again. Mostly
16242 work arounds for the VC++ problem that it treats func() as
16243 func(void). Also stripped out the parts of mkdef.pl that
16244 added extra typesafe functions: these no longer exist.
16248 * Reorganisation of the stack code. The macros are now all
16249 collected in safestack.h . Each macro is defined in terms of
16250 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16251 DEBUG_SAFESTACK is now handled in terms of function casts,
16252 this has the advantage of retaining type safety without the
16253 use of additional functions. If DEBUG_SAFESTACK is not defined
16254 then the non typesafe macros are used instead. Also modified the
16255 mkstack.pl script to handle the new form. Needs testing to see
16256 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16257 the default if no major problems. Similar behaviour for ASN1_SET_OF
16258 and PKCS12_STACK_OF.
16262 * When some versions of IIS use the 'NET' form of private key the
16263 key derivation algorithm is different. Normally MD5(password) is
16264 used as a 128 bit RC4 key. In the modified case
16265 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16266 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16267 as the old Netscape_RSA functions except they have an additional
16268 'sgckey' parameter which uses the modified algorithm. Also added
16269 an -sgckey command line option to the rsa utility. Thanks to
16270 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16271 algorithm to openssl-dev.
16275 * The evp_local.h macros were using 'c.##kname' which resulted in
16276 invalid expansion on some systems (SCO 5.0.5 for example).
16277 Corrected to 'c.kname'.
16279 *Phillip Porch <root@theporch.com>*
16281 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16282 a STACK of email addresses from a certificate or request, these look
16283 in the subject name and the subject alternative name extensions and
16284 omit any duplicate addresses.
16288 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16289 This makes DSA verification about 2 % faster.
16293 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16294 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16295 plus overhead for 1024 bit moduli).
16296 This makes exponentiations about 0.5 % faster for 1024 bit
16297 exponents (as measured by "openssl speed rsa2048").
16301 * Rename memory handling macros to avoid conflicts with other
16303 Malloc => OPENSSL_malloc
16304 Malloc_locked => OPENSSL_malloc_locked
16305 Realloc => OPENSSL_realloc
16306 Free => OPENSSL_free
16310 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16311 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16315 * CygWin32 support.
16317 *John Jarvie <jjarvie@newsguy.com>*
16319 * The type-safe stack code has been rejigged. It is now only compiled
16320 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16321 by default all type-specific stack functions are "#define"d back to
16322 standard stack functions. This results in more streamlined output
16323 but retains the type-safety checking possibilities of the original
16328 * The STACK code has been cleaned up, and certain type declarations
16329 that didn't make a lot of sense have been brought in line. This has
16330 also involved a cleanup of sorts in safestack.h to more correctly
16331 map type-safe stack functions onto their plain stack counterparts.
16332 This work has also resulted in a variety of "const"ifications of
16333 lots of the code, especially `_cmp` operations which should normally
16334 be prototyped with "const" parameters anyway.
16338 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16339 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16340 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16341 where all of 'md' is used each time the PRNG is used, but 'state'
16342 is used only indexed by a cyclic counter. As entropy may not be
16343 well distributed from the beginning, 'md' is important as a
16344 chaining variable. However, the output function chains only half
16345 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16346 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16347 in all of 'state' being rewritten, with the new values depending
16348 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16352 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16353 the handshake is continued after ssl_verify_cert_chain();
16354 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16355 can lead to 'unexplainable' connection aborts later.
16357 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16359 * Major EVP API cipher revision.
16360 Add hooks for extra EVP features. This allows various cipher
16361 parameters to be set in the EVP interface. Support added for variable
16362 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16363 setting of RC2 and RC5 parameters.
16365 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16368 Remove lots of duplicated code from the EVP library. For example *every*
16369 cipher init() function handles the 'iv' in the same way according to the
16370 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16371 for CFB and OFB modes they zero ctx->num.
16373 New functionality allows removal of S/MIME code RC2 hack.
16375 Most of the routines have the same form and so can be declared in terms
16378 By shifting this to the top level EVP_CipherInit() it can be removed from
16379 all individual ciphers. If the cipher wants to handle IVs or keys
16380 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16383 Change lots of functions like EVP_EncryptUpdate() to now return a
16384 value: although software versions of the algorithms cannot fail
16385 any installed hardware versions can.
16389 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16390 this option is set, tolerate broken clients that send the negotiated
16391 protocol version number instead of the requested protocol version
16396 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16397 i.e. non-zero for export ciphersuites, zero otherwise.
16398 Previous versions had this flag inverted, inconsistent with
16399 rsa_tmp_cb (..._TMP_RSA_CB).
16401 *Bodo Moeller; problem reported by Amit Chopra*
16403 * Add missing DSA library text string. Work around for some IIS
16404 key files with invalid SEQUENCE encoding.
16408 * Add a document (doc/standards.txt) that list all kinds of standards
16409 and so on that are implemented in OpenSSL.
16413 * Enhance c_rehash script. Old version would mishandle certificates
16414 with the same subject name hash and wouldn't handle CRLs at all.
16415 Added -fingerprint option to crl utility, to support new c_rehash
16420 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16424 * Fix for SSL server purpose checking. Server checking was
16425 rejecting certificates which had extended key usage present
16426 but no ssl client purpose.
16428 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16430 * Make PKCS#12 code work with no password. The PKCS#12 spec
16431 is a little unclear about how a blank password is handled.
16432 Since the password in encoded as a BMPString with terminating
16433 double NULL a zero length password would end up as just the
16434 double NULL. However no password at all is different and is
16435 handled differently in the PKCS#12 key generation code. NS
16436 treats a blank password as zero length. MSIE treats it as no
16437 password on export: but it will try both on import. We now do
16438 the same: PKCS12_parse() tries zero length and no password if
16439 the password is set to "" or NULL (NULL is now a valid password:
16440 it wasn't before) as does the pkcs12 application.
16444 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16445 perror when PEM_read_bio_X509_REQ fails, the error message must
16446 be obtained from the error queue.
16450 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16451 it in ERR_remove_state if appropriate, and change ERR_get_state
16452 accordingly to avoid race conditions (this is necessary because
16453 thread_hash is no longer constant once set).
16457 * Bugfix for linux-elf makefile.one.
16461 * RSA_get_default_method() will now cause a default
16462 RSA_METHOD to be chosen if one doesn't exist already.
16463 Previously this was only set during a call to RSA_new()
16464 or RSA_new_method(NULL) meaning it was possible for
16465 RSA_get_default_method() to return NULL.
16469 * Added native name translation to the existing DSO code
16470 that will convert (if the flag to do so is set) filenames
16471 that are sufficiently small and have no path information
16472 into a canonical native form. Eg. "blah" converted to
16473 "libblah.so" or "blah.dll" etc.
16477 * New function ERR_error_string_n(e, buf, len) which is like
16478 ERR_error_string(e, buf), but writes at most 'len' bytes
16479 including the 0 terminator. For ERR_error_string_n, 'buf'
16482 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16484 * CONF library reworked to become more general. A new CONF
16485 configuration file reader "class" is implemented as well as a
16486 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16487 old `CONF_*` functions are still there, but are reimplemented to
16488 work in terms of the new functions. Also, a set of functions
16489 to handle the internal storage of the configuration data is
16490 provided to make it easier to write new configuration file
16491 reader "classes" (I can definitely see something reading a
16492 configuration file in XML format, for example), called `_CONF_*`,
16493 or "the configuration storage API"...
16495 The new configuration file reading functions are:
16497 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16498 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16500 NCONF_default, NCONF_WIN32
16502 NCONF_dump_fp, NCONF_dump_bio
16504 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16505 NCONF_new creates a new CONF object. This works in the same way
16506 as other interfaces in OpenSSL, like the BIO interface.
16507 `NCONF_dump_*` dump the internal storage of the configuration file,
16508 which is useful for debugging. All other functions take the same
16509 arguments as the old `CONF_*` functions with the exception of the
16510 first that must be a `CONF *` instead of a `LHASH *`.
16512 To make it easier to use the new classes with the old `CONF_*` functions,
16513 the function CONF_set_default_method is provided.
16517 * Add '-tls1' option to 'openssl ciphers', which was already
16518 mentioned in the documentation but had not been implemented.
16519 (This option is not yet really useful because even the additional
16520 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16524 * Initial DSO code added into libcrypto for letting OpenSSL (and
16525 OpenSSL-based applications) load shared libraries and bind to
16526 them in a portable way.
16528 *Geoff Thorpe, with contributions from Richard Levitte*
16530 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16532 * Make sure _lrotl and _lrotr are only used with MSVC.
16534 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16535 (the default implementation of RAND_status).
16537 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16538 to '-clrext' (= clear extensions), as intended and documented.
16539 *Bodo Moeller; inconsistency pointed out by Michael Attili
16540 <attili@amaxo.com>*
16542 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16543 was larger than the MD block size.
16545 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16547 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16548 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16549 using the passed key: if the passed key was a private key the result
16550 of X509_print(), for example, would be to print out all the private key
16555 * des_quad_cksum() byte order bug fix.
16556 *Ulf Möller, using the problem description in krb4-0.9.7, where
16557 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16559 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16562 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16564 * For easily testing in shell scripts whether some command
16565 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16566 returns with exit code 0 iff no command of the given name is available.
16567 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16568 the output goes to stdout and nothing is printed to stderr.
16569 Additional arguments are always ignored.
16571 Since for each cipher there is a command of the same name,
16572 the 'no-cipher' compilation switches can be tested this way.
16574 ('openssl no-XXX' is not able to detect pseudo-commands such
16575 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16579 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16583 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16584 is set; it will be thrown away anyway because each handshake creates
16586 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16587 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16588 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16589 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16593 * New s_client option -ign_eof: EOF at stdin is ignored, and
16594 'Q' and 'R' lose their special meanings (quit/renegotiate).
16595 This is part of what -quiet does; unlike -quiet, -ign_eof
16596 does not suppress any output.
16600 * Add compatibility options to the purpose and trust code. The
16601 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16602 accepts a certificate or CA, this was the previous behaviour,
16603 with all the associated security issues.
16605 X509_TRUST_COMPAT is the old trust behaviour: only and
16606 automatically trust self signed roots in certificate store. A
16607 new trust setting X509_TRUST_DEFAULT is used to specify that
16608 a purpose has no associated trust setting and it should instead
16609 use the value in the default purpose.
16613 * Fix the PKCS#8 DSA private key code so it decodes keys again
16614 and fix a memory leak.
16618 * In util/mkerr.pl (which implements 'make errors'), preserve
16619 reason strings from the previous version of the .c file, as
16620 the default to have only downcase letters (and digits) in
16621 automatically generated reasons codes is not always appropriate.
16625 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16626 using strerror. Previously, ERR_reason_error_string() returned
16627 library names as reason strings for SYSerr; but SYSerr is a special
16628 case where small numbers are errno values, not library numbers.
16632 * Add '-dsaparam' option to 'openssl dhparam' application. This
16633 converts DSA parameters into DH parameters. (When creating parameters,
16634 DSA_generate_parameters is used.)
16638 * Include 'length' (recommended exponent length) in C code generated
16639 by 'openssl dhparam -C'.
16643 * The second argument to set_label in perlasm was already being used
16644 so couldn't be used as a "file scope" flag. Moved to third argument
16649 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16650 instead of RAND_bytes for encryption IVs and salts.
16654 * Include RAND_status() into RAND_METHOD instead of implementing
16655 it only for md_rand.c Otherwise replacing the PRNG by calling
16656 RAND_set_rand_method would be impossible.
16660 * Don't let DSA_generate_key() enter an infinite loop if the random
16661 number generation fails.
16665 * New 'rand' application for creating pseudo-random output.
16669 * Added configuration support for Linux/IA64
16671 *Rolf Haberrecker <rolf@suse.de>*
16673 * Assembler module support for Mingw32.
16677 * Shared library support for HPUX (in shlib/).
16679 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16681 * Shared library support for Solaris gcc.
16683 *Lutz Behnke <behnke@trustcenter.de>*
16685 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16687 * PKCS7_encrypt() was adding text MIME headers twice because they
16688 were added manually and by SMIME_crlf_copy().
16692 * In bntest.c don't call BN_rand with zero bits argument.
16694 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16696 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16697 case was implemented. This caused BN_div_recp() to fail occasionally.
16701 * Add an optional second argument to the set_label() in the perl
16702 assembly language builder. If this argument exists and is set
16703 to 1 it signals that the assembler should use a symbol whose
16704 scope is the entire file, not just the current function. This
16705 is needed with MASM which uses the format label:: for this scope.
16707 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16709 * Change the ASN1 types so they are typedefs by default. Before
16710 almost all types were #define'd to ASN1_STRING which was causing
16711 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16716 * Change names of new functions to the new get1/get0 naming
16717 convention: After 'get1', the caller owns a reference count
16718 and has to call `..._free`; 'get0' returns a pointer to some
16719 data structure without incrementing reference counters.
16720 (Some of the existing 'get' functions increment a reference
16721 counter, some don't.)
16722 Similarly, 'set1' and 'add1' functions increase reference
16723 counters or duplicate objects.
16727 * Allow for the possibility of temp RSA key generation failure:
16728 the code used to assume it always worked and crashed on failure.
16732 * Fix potential buffer overrun problem in BIO_printf().
16733 *Ulf Möller, using public domain code by Patrick Powell; problem
16734 pointed out by David Sacerdote <das33@cornell.edu>*
16736 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16737 RAND_egd() and RAND_status(). In the command line application,
16738 the EGD socket can be specified like a seed file using RANDFILE
16743 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16744 Some CAs (e.g. Verisign) distribute certificates in this form.
16748 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16749 list to exclude them. This means that no special compilation option
16750 is needed to use anonymous DH: it just needs to be included in the
16755 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16756 EVP_MD_type. The old functionality is available in a new macro called
16757 EVP_MD_md(). Change code that uses it and update docs.
16761 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16762 where the `void *` argument is replaced by a function pointer argument.
16763 Previously `void *` was abused to point to functions, which works on
16764 many platforms, but is not correct. As these functions are usually
16765 called by macros defined in OpenSSL header files, most source code
16766 should work without changes.
16770 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16771 sections with information on -D... compiler switches used for
16772 compiling the library so that applications can see them. To enable
16773 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16774 must be defined. E.g.,
16775 #define OPENSSL_ALGORITHM_DEFINES
16776 #include <openssl/opensslconf.h>
16777 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16779 *Richard Levitte, Ulf and Bodo Möller*
16781 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16786 * Change the 'other' type in certificate aux info to a STACK_OF
16787 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16788 the required ASN1 format: arbitrary types determined by an OID.
16792 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16793 argument to 'req'. This is not because the function is newer or
16794 better than others it just uses the work 'NEW' in the certificate
16795 request header lines. Some software needs this.
16799 * Reorganise password command line arguments: now passwords can be
16800 obtained from various sources. Delete the PEM_cb function and make
16801 it the default behaviour: i.e. if the callback is NULL and the
16802 usrdata argument is not NULL interpret it as a null terminated pass
16803 phrase. If usrdata and the callback are NULL then the pass phrase
16804 is prompted for as usual.
16808 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16809 the support is automatically enabled. The resulting binaries will
16810 autodetect the card and use it if present.
16812 *Ben Laurie and Compaq Inc.*
16814 * Work around for Netscape hang bug. This sends certificate request
16815 and server done in one record. Since this is perfectly legal in the
16816 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16817 the bugs/SSLv3 entry for more info.
16821 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16825 * Add -rand argument to smime and pkcs12 applications and read/write
16830 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16834 * Add command line password options to the remaining applications.
16838 * Bug fix for BN_div_recp() for numerators with an even number of
16843 * More tests in bntest.c, and changed test_bn output.
16847 * ./config recognizes MacOS X now.
16851 * Bug fix for BN_div() when the first words of num and divisor are
16852 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16856 * Add support for various broken PKCS#8 formats, and command line
16857 options to produce them.
16861 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16862 get temporary BIGNUMs from a BN_CTX.
16866 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16871 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16872 include a #define from the old name to the new. The original intent
16873 was that statically linked binaries could for example just call
16874 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16875 link with digests. This never worked because SSLeay_add_all_digests()
16876 and SSLeay_add_all_ciphers() were in the same source file so calling
16877 one would link with the other. They are now in separate source files.
16881 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16885 * Use a less unusual form of the Miller-Rabin primality test (it used
16886 a binary algorithm for exponentiation integrated into the Miller-Rabin
16887 loop, our standard modexp algorithms are faster).
16891 * Support for the EBCDIC character set completed.
16893 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16895 * Source code cleanups: use const where appropriate, eliminate casts,
16896 use `void *` instead of `char *` in lhash.
16900 * Bugfix: ssl3_send_server_key_exchange was not restartable
16901 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16902 this the server could overwrite ephemeral keys that the client
16907 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16908 using 50 iterations of the Rabin-Miller test.
16910 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16911 iterations of the Rabin-Miller test as required by the appendix
16912 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16913 As BN_is_prime_fasttest includes trial division, DSA parameter
16914 generation becomes much faster.
16916 This implies a change for the callback functions in DSA_is_prime
16917 and DSA_generate_parameters: The callback function is called once
16918 for each positive witness in the Rabin-Miller test, not just
16919 occasionally in the inner loop; and the parameters to the
16920 callback function now provide an iteration count for the outer
16921 loop rather than for the current invocation of the inner loop.
16922 DSA_generate_parameters additionally can call the callback
16923 function with an 'iteration count' of -1, meaning that a
16924 candidate has passed the trial division test (when q is generated
16925 from an application-provided seed, trial division is skipped).
16929 * New function BN_is_prime_fasttest that optionally does trial
16930 division before starting the Rabin-Miller test and has
16931 an additional BN_CTX * argument (whereas BN_is_prime always
16932 has to allocate at least one BN_CTX).
16933 'callback(1, -1, cb_arg)' is called when a number has passed the
16934 trial division stage.
16938 * Fix for bug in CRL encoding. The validity dates weren't being handled
16943 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16947 * New function BN_pseudo_rand().
16951 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16952 bignum version of BN_from_montgomery() with the working code from
16953 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16958 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16959 made it impossible to use the same SSL_SESSION data structure in
16960 SSL2 clients in multiple threads.
16964 * The return value of RAND_load_file() no longer counts bytes obtained
16965 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16966 to seed the PRNG (previously an explicit byte count was required).
16968 *Ulf Möller, Bodo Möller*
16970 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16971 used `char *` instead of `void *` and had casts all over the place.
16975 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16979 * Retain source code compatibility for BN_prime_checks macro:
16980 BN_is_prime(..., BN_prime_checks, ...) now uses
16981 BN_prime_checks_for_size to determine the appropriate number of
16982 Rabin-Miller iterations.
16986 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16987 DH_CHECK_P_NOT_SAFE_PRIME.
16988 (Check if this is true? OpenPGP calls them "strong".)
16992 * Merge the functionality of "dh" and "gendh" programs into a new program
16993 "dhparam". The old programs are retained for now but will handle DH keys
16994 (instead of parameters) in future.
16998 * Make the ciphers, s_server and s_client programs check the return values
16999 when a new cipher list is set.
17003 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17004 ciphers. Before when the 56bit ciphers were enabled the sorting was
17007 The syntax for the cipher sorting has been extended to support sorting by
17008 cipher-strength (using the strength_bits hard coded in the tables).
17009 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17011 Fix a bug in the cipher-command parser: when supplying a cipher command
17012 string with an "undefined" symbol (neither command nor alphanumeric
17013 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17014 an error is flagged.
17016 Due to the strength-sorting extension, the code of the
17017 ssl_create_cipher_list() function was completely rearranged. I hope that
17018 the readability was also increased :-)
17020 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17022 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17023 for the first serial number and places 2 in the serial number file. This
17024 avoids problems when the root CA is created with serial number zero and
17025 the first user certificate has the same issuer name and serial number
17030 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17031 the new code. Add documentation for this stuff.
17035 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17036 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17037 structures and behave in an analogous way to the X509v3 functions:
17038 they shouldn't be called directly but wrapper functions should be used
17041 So we also now have some wrapper functions that call the X509at functions
17042 when passed certificate requests. (TO DO: similar things can be done with
17043 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17044 things. Some of these need some d2i or i2d and print functionality
17045 because they handle more complex structures.)
17049 * Add missing #ifndefs that caused missing symbols when building libssl
17050 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17051 NO_RSA in `ssl/s2*.c`.
17053 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17055 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17056 has a return value which indicates the quality of the random data
17057 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17058 error queue. New function RAND_pseudo_bytes() generates output that is
17059 guaranteed to be unique but not unpredictable. RAND_add is like
17060 RAND_seed, but takes an extra argument for an entropy estimate
17061 (RAND_seed always assumes full entropy).
17065 * Do more iterations of Rabin-Miller probable prime test (specifically,
17066 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17067 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17068 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17069 false-positive rate of at most 2^-80 for random input.
17073 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17077 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17078 in the 0.9.5 release), this returns the chain
17079 from an X509_CTX structure with a dup of the stack and all
17080 the X509 reference counts upped: so the stack will exist
17081 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17084 Also make SSL_SESSION_print() print out the verify return
17089 * Add manpage for the pkcs12 command. Also change the default
17090 behaviour so MAC iteration counts are used unless the new
17091 -nomaciter option is used. This improves file security and
17092 only older versions of MSIE (4.0 for example) need it.
17096 * Honor the no-xxx Configure options when creating .DEF files.
17100 * Add PKCS#10 attributes to field table: challengePassword,
17101 unstructuredName and unstructuredAddress. These are taken from
17102 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17103 international characters are used.
17105 More changes to X509_ATTRIBUTE code: allow the setting of types
17106 based on strings. Remove the 'loc' parameter when adding
17107 attributes because these will be a SET OF encoding which is sorted
17112 * Initial changes to the 'req' utility to allow request generation
17113 automation. This will allow an application to just generate a template
17114 file containing all the field values and have req construct the
17117 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17118 used all over the place including certificate requests and PKCS#7
17119 structures. They are currently handled manually where necessary with
17120 some primitive wrappers for PKCS#7. The new functions behave in a
17121 manner analogous to the X509 extension functions: they allow
17122 attributes to be looked up by NID and added.
17124 Later something similar to the X509V3 code would be desirable to
17125 automatically handle the encoding, decoding and printing of the
17126 more complex types. The string types like challengePassword can
17127 be handled by the string table functions.
17129 Also modified the multi byte string table handling. Now there is
17130 a 'global mask' which masks out certain types. The table itself
17131 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17132 is useful when for example there is only one permissible type
17133 (as in countryName) and using the mask might result in no valid
17138 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17139 SSL_get_peer_finished to allow applications to obtain the latest
17140 Finished messages sent to the peer or expected from the peer,
17141 respectively. (SSL_get_peer_finished is usually the Finished message
17142 actually received from the peer, otherwise the protocol will be aborted.)
17144 As the Finished message are message digests of the complete handshake
17145 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17146 be used for external authentication procedures when the authentication
17147 provided by SSL/TLS is not desired or is not enough.
17151 * Enhanced support for Alpha Linux is added. Now ./config checks if
17152 the host supports BWX extension and if Compaq C is present on the
17153 $PATH. Just exploiting of the BWX extension results in 20-30%
17154 performance kick for some algorithms, e.g. DES and RC4 to mention
17155 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17160 * Add support for MS "fast SGC". This is arguably a violation of the
17161 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17162 weak crypto and after checking the certificate is SGC a second one
17163 with strong crypto. MS SGC stops the first handshake after receiving
17164 the server certificate message and sends a second client hello. Since
17165 a server will typically do all the time consuming operations before
17166 expecting any further messages from the client (server key exchange
17167 is the most expensive) there is little difference between the two.
17169 To get OpenSSL to support MS SGC we have to permit a second client
17170 hello message after we have sent server done. In addition we have to
17171 reset the MAC if we do get this second client hello.
17175 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17176 if a DER encoded private key is RSA or DSA traditional format. Changed
17177 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17178 format DER encoded private key. Newer code should use PKCS#8 format which
17179 has the key type encoded in the ASN1 structure. Added DER private key
17180 support to pkcs8 application.
17184 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17185 ciphersuites has been selected (as required by the SSL 3/TLS 1
17186 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17187 is set, we interpret this as a request to violate the specification
17188 (the worst that can happen is a handshake failure, and 'correct'
17189 behaviour would result in a handshake failure anyway).
17193 * In SSL_CTX_add_session, take into account that there might be multiple
17194 SSL_SESSION structures with the same session ID (e.g. when two threads
17195 concurrently obtain them from an external cache).
17196 The internal cache can handle only one SSL_SESSION with a given ID,
17197 so if there's a conflict, we now throw out the old one to achieve
17202 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17203 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17204 some routines that use cipher OIDs: some ciphers do not have OIDs
17205 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17210 * Simplify the trust setting structure and code. Now we just have
17211 two sequences of OIDs for trusted and rejected settings. These will
17212 typically have values the same as the extended key usage extension
17213 and any application specific purposes.
17215 The trust checking code now has a default behaviour: it will just
17216 check for an object with the same NID as the passed id. Functions can
17217 be provided to override either the default behaviour or the behaviour
17218 for a given id. SSL client, server and email already have functions
17219 in place for compatibility: they check the NID and also return "trusted"
17220 if the certificate is self signed.
17224 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17225 traditional format into an EVP_PKEY structure.
17229 * Add a password callback function PEM_cb() which either prompts for
17230 a password if usr_data is NULL or otherwise assumes it is a null
17231 terminated password. Allow passwords to be passed on command line
17232 environment or config files in a few more utilities.
17236 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17237 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17238 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17239 Update documentation.
17243 * Support for ASN1 "NULL" type. This could be handled before by using
17244 ASN1_TYPE but there wasn't any function that would try to read a NULL
17245 and produce an error if it couldn't. For compatibility we also have
17246 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17247 don't allocate anything because they don't need to.
17251 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17254 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17256 * Rebuild of the memory allocation routines used by OpenSSL code and
17257 possibly others as well. The purpose is to make an interface that
17258 provide hooks so anyone can build a separate set of allocation and
17259 deallocation routines to be used by OpenSSL, for example memory
17260 pool implementations, or something else, which was previously hard
17261 since Malloc(), Realloc() and Free() were defined as macros having
17262 the values malloc, realloc and free, respectively (except for Win32
17263 compilations). The same is provided for memory debugging code.
17264 OpenSSL already comes with functionality to find memory leaks, but
17265 this gives people a chance to debug other memory problems.
17267 With these changes, a new set of functions and macros have appeared:
17269 CRYPTO_set_mem_debug_functions() [F]
17270 CRYPTO_get_mem_debug_functions() [F]
17271 CRYPTO_dbg_set_options() [F]
17272 CRYPTO_dbg_get_options() [F]
17273 CRYPTO_malloc_debug_init() [M]
17275 The memory debug functions are NULL by default, unless the library
17276 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17277 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17278 gives the standard debugging functions that come with OpenSSL) or
17279 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17280 provided by the library user) must be used. When the standard
17281 debugging functions are used, CRYPTO_dbg_set_options can be used to
17282 request additional information:
17283 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17284 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17286 Also, things like CRYPTO_set_mem_functions will always give the
17287 expected result (the new set of functions is used for allocation
17288 and deallocation) at all times, regardless of platform and compiler
17291 To finish it up, some functions that were never use in any other
17292 way than through macros have a new API and new semantic:
17294 CRYPTO_dbg_malloc()
17295 CRYPTO_dbg_realloc()
17298 All macros of value have retained their old syntax.
17300 *Richard Levitte and Bodo Moeller*
17302 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17303 ordering of SMIMECapabilities wasn't in "strength order" and there
17304 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17309 * Some ASN1 types with illegal zero length encoding (INTEGER,
17310 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17312 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17314 * Merge in my S/MIME library for OpenSSL. This provides a simple
17315 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17316 functionality to handle multipart/signed properly) and a utility
17317 called 'smime' to call all this stuff. This is based on code I
17318 originally wrote for Celo who have kindly allowed it to be
17319 included in OpenSSL.
17323 * Add variants des_set_key_checked and des_set_key_unchecked of
17324 des_set_key (aka des_key_sched). Global variable des_check_key
17325 decides which of these is called by des_set_key; this way
17326 des_check_key behaves as it always did, but applications and
17327 the library itself, which was buggy for des_check_key == 1,
17328 have a cleaner way to pick the version they need.
17332 * New function PKCS12_newpass() which changes the password of a
17337 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17338 dynamic mix. In both cases the ids can be used as an index into the
17339 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17340 functions so they accept a list of the field values and the
17341 application doesn't need to directly manipulate the X509_TRUST
17346 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17351 * Modify the way the V3 extension code looks up extensions. This now
17352 works in a similar way to the object code: we have some "standard"
17353 extensions in a static table which is searched with OBJ_bsearch()
17354 and the application can add dynamic ones if needed. The file
17355 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17356 updated whenever a new extension is added to the core code and kept
17357 in ext_nid order. There is a simple program 'tabtest.c' which checks
17358 this. New extensions are not added too often so this file can readily
17359 be maintained manually.
17361 There are two big advantages in doing things this way. The extensions
17362 can be looked up immediately and no longer need to be "added" using
17363 X509V3_add_standard_extensions(): this function now does nothing.
17364 Side note: I get *lots* of email saying the extension code doesn't
17365 work because people forget to call this function.
17366 Also no dynamic allocation is done unless new extensions are added:
17367 so if we don't add custom extensions there is no need to call
17368 X509V3_EXT_cleanup().
17372 * Modify enc utility's salting as follows: make salting the default. Add a
17373 magic header, so unsalted files fail gracefully instead of just decrypting
17374 to garbage. This is because not salting is a big security hole, so people
17375 should be discouraged from doing it.
17379 * Fixes and enhancements to the 'x509' utility. It allowed a message
17380 digest to be passed on the command line but it only used this
17381 parameter when signing a certificate. Modified so all relevant
17382 operations are affected by the digest parameter including the
17383 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17384 DSA key was used because it didn't fix the digest.
17388 * Initial certificate chain verify code. Currently tests the untrusted
17389 certificates for consistency with the verify purpose (which is set
17390 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17392 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17393 this is because it will reject chains with invalid extensions whereas
17394 every previous version of OpenSSL and SSLeay made no checks at all.
17396 Trust code: checks the root CA for the relevant trust settings. Trust
17397 settings have an initial value consistent with the verify purpose: e.g.
17398 if the verify purpose is for SSL client use it expects the CA to be
17399 trusted for SSL client use. However the default value can be changed to
17400 permit custom trust settings: one example of this would be to only trust
17401 certificates from a specific "secure" set of CAs.
17403 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17404 which should be used for version portability: especially since the
17405 verify structure is likely to change more often now.
17407 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17408 to set them. If not set then assume SSL clients will verify SSL servers
17411 Two new options to the verify program: -untrusted allows a set of
17412 untrusted certificates to be passed in and -purpose which sets the
17413 intended purpose of the certificate. If a purpose is set then the
17414 new chain verify code is used to check extension consistency.
17418 * Support for the authority information access extension.
17422 * Modify RSA and DSA PEM read routines to transparently handle
17423 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17424 public keys in a format compatible with certificate
17425 SubjectPublicKeyInfo structures. Unfortunately there were already
17426 functions called *_PublicKey_* which used various odd formats so
17427 these are retained for compatibility: however the DSA variants were
17428 never in a public release so they have been deleted. Changed dsa/rsa
17429 utilities to handle the new format: note no releases ever handled public
17430 keys so we should be OK.
17432 The primary motivation for this change is to avoid the same fiasco
17433 that dogs private keys: there are several incompatible private key
17434 formats some of which are standard and some OpenSSL specific and
17435 require various evil hacks to allow partial transparent handling and
17436 even then it doesn't work with DER formats. Given the option anything
17437 other than PKCS#8 should be dumped: but the other formats have to
17438 stay in the name of compatibility.
17440 With public keys and the benefit of hindsight one standard format
17441 is used which works with EVP_PKEY, RSA or DSA structures: though
17442 it clearly returns an error if you try to read the wrong kind of key.
17444 Added a -pubkey option to the 'x509' utility to output the public key.
17445 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17446 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17447 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17448 that do the same as the `EVP_PKEY_assign_*()` except they up the
17449 reference count of the added key (they don't "swallow" the
17454 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17455 CRLs would fail if the file contained no certificates or no CRLs:
17456 added a new function to read in both types and return the number
17457 read: this means that if none are read it will be an error. The
17458 DER versions of the certificate and CRL reader would always fail
17459 because it isn't possible to mix certificates and CRLs in DER format
17460 without choking one or the other routine. Changed this to just read
17461 a certificate: this is the best we can do. Also modified the code
17462 in `apps/verify.c` to take notice of return codes: it was previously
17463 attempting to read in certificates from NULL pointers and ignoring
17464 any errors: this is one reason why the cert and CRL reader seemed
17465 to work. It doesn't check return codes from the default certificate
17466 routines: these may well fail if the certificates aren't installed.
17470 * Code to support otherName option in GeneralName.
17474 * First update to verify code. Change the verify utility
17475 so it warns if it is passed a self signed certificate:
17476 for consistency with the normal behaviour. X509_verify
17477 has been modified to it will now verify a self signed
17478 certificate if *exactly* the same certificate appears
17479 in the store: it was previously impossible to trust a
17480 single self signed certificate. This means that:
17481 openssl verify ss.pem
17482 now gives a warning about a self signed certificate but
17483 openssl verify -CAfile ss.pem ss.pem
17488 * For servers, store verify_result in SSL_SESSION data structure
17489 (and add it to external session representation).
17490 This is needed when client certificate verifications fails,
17491 but an application-provided verification callback (set by
17492 SSL_CTX_set_cert_verify_callback) allows accepting the session
17493 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17494 but returns 1): When the session is reused, we have to set
17495 ssl->verify_result to the appropriate error code to avoid
17498 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17500 * Fix a bug in the new PKCS#7 code: it didn't consider the
17501 case in PKCS7_dataInit() where the signed PKCS7 structure
17502 didn't contain any existing data because it was being created.
17504 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17506 * Add a salt to the key derivation routines in enc.c. This
17507 forms the first 8 bytes of the encrypted file. Also add a
17508 -S option to allow a salt to be input on the command line.
17512 * New function X509_cmp(). Oddly enough there wasn't a function
17513 to compare two certificates. We do this by working out the SHA1
17514 hash and comparing that. X509_cmp() will be needed by the trust
17519 * SSL_get1_session() is like SSL_get_session(), but increments
17520 the reference count in the SSL_SESSION returned.
17522 *Geoff Thorpe <geoff@eu.c2.net>*
17524 * Fix for 'req': it was adding a null to request attributes.
17525 Also change the X509_LOOKUP and X509_INFO code to handle
17526 certificate auxiliary information.
17530 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17535 * Add the possibility to add extra information to the memory leak
17536 detecting output, to form tracebacks, showing from where each
17537 allocation was originated: CRYPTO_push_info("constant string") adds
17538 the string plus current file name and line number to a per-thread
17539 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17540 is like calling CYRPTO_pop_info() until the stack is empty.
17541 Also updated memory leak detection code to be multi-thread-safe.
17545 * Add options -text and -noout to pkcs7 utility and delete the
17546 encryption options which never did anything. Update docs.
17550 * Add options to some of the utilities to allow the pass phrase
17551 to be included on either the command line (not recommended on
17552 OSes like Unix) or read from the environment. Update the
17553 manpages and fix a few bugs.
17557 * Add a few manpages for some of the openssl commands.
17561 * Fix the -revoke option in ca. It was freeing up memory twice,
17562 leaking and not finding already revoked certificates.
17566 * Extensive changes to support certificate auxiliary information.
17567 This involves the use of X509_CERT_AUX structure and X509_AUX
17568 functions. An X509_AUX function such as PEM_read_X509_AUX()
17569 can still read in a certificate file in the usual way but it
17570 will also read in any additional "auxiliary information". By
17571 doing things this way a fair degree of compatibility can be
17572 retained: existing certificates can have this information added
17573 using the new 'x509' options.
17575 Current auxiliary information includes an "alias" and some trust
17576 settings. The trust settings will ultimately be used in enhanced
17577 certificate chain verification routines: currently a certificate
17578 can only be trusted if it is self signed and then it is trusted
17583 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17584 The problem was that one of the replacement routines had not been working
17585 since SSLeay releases. For now the offending routine has been replaced
17586 with non-optimised assembler. Even so, this now gives around 95%
17587 performance improvement for 1024 bit RSA signs.
17591 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17592 handling. Most clients have the effective key size in bits equal to
17593 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17594 A few however don't do this and instead use the size of the decrypted key
17595 to determine the RC2 key length and the AlgorithmIdentifier to determine
17596 the effective key length. In this case the effective key length can still
17597 be 40 bits but the key length can be 168 bits for example. This is fixed
17598 by manually forcing an RC2 key into the EVP_PKEY structure because the
17599 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17600 the key length and effective key length are equal.
17604 * Add a bunch of functions that should simplify the creation of
17605 X509_NAME structures. Now you should be able to do:
17606 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17607 and have it automatically work out the correct field type and fill in
17608 the structures. The more adventurous can try:
17609 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17610 and it will (hopefully) work out the correct multibyte encoding.
17614 * Change the 'req' utility to use the new field handling and multibyte
17615 copy routines. Before the DN field creation was handled in an ad hoc
17616 way in req, ca, and x509 which was rather broken and didn't support
17617 BMPStrings or UTF8Strings. Since some software doesn't implement
17618 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17619 using the dirstring_type option. See the new comment in the default
17620 openssl.cnf for more info.
17624 * Make crypto/rand/md_rand.c more robust:
17625 - Assure unique random numbers after fork().
17626 - Make sure that concurrent threads access the global counter and
17627 md serializably so that we never lose entropy in them
17628 or use exactly the same state in multiple threads.
17629 Access to the large state is not always serializable because
17630 the additional locking could be a performance killer, and
17631 md should be large enough anyway.
17635 * New file `apps/app_rand.c` with commonly needed functionality
17636 for handling the random seed file.
17638 Use the random seed file in some applications that previously did not:
17640 dsaparam -genkey (which also ignored its '-rand' option),
17643 x509 (when signing).
17644 Except on systems with /dev/urandom, it is crucial to have a random
17645 seed file at least for key creation, DSA signing, and for DH exchanges;
17646 for RSA signatures we could do without one.
17648 gendh and gendsa (unlike genrsa) used to read only the first byte
17649 of each file listed in the '-rand' option. The function as previously
17650 found in genrsa is now in app_rand.c and is used by all programs
17651 that support '-rand'.
17655 * In RAND_write_file, use mode 0600 for creating files;
17656 don't just chmod when it may be too late.
17660 * Report an error from X509_STORE_load_locations
17661 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17665 * New function ASN1_mbstring_copy() this copies a string in either
17666 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17667 into an ASN1_STRING type. A mask of permissible types is passed
17668 and it chooses the "minimal" type to use or an error if not type
17673 * Add function equivalents to the various macros in asn1.h. The old
17674 macros are retained with an `M_` prefix. Code inside the library can
17675 use the `M_` macros. External code (including the openssl utility)
17676 should *NOT* in order to be "shared library friendly".
17680 * Add various functions that can check a certificate's extensions
17681 to see if it usable for various purposes such as SSL client,
17682 server or S/MIME and CAs of these types. This is currently
17683 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17684 verification. Also added a -purpose flag to x509 utility to
17685 print out all the purposes.
17689 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17694 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17695 for, obtain and decode and extension and obtain its critical flag.
17696 This allows all the necessary extension code to be handled in a
17697 single function call.
17701 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17702 platforms. See crypto/rc4/rc4_enc.c for further details.
17706 * New -noout option to asn1parse. This causes no output to be produced
17707 its main use is when combined with -strparse and -out to extract data
17708 from a file (which may not be in ASN.1 format).
17712 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17713 when producing the local key id.
17715 *Richard Levitte <levitte@stacken.kth.se>*
17717 * New option -dhparam in s_server. This allows a DH parameter file to be
17718 stated explicitly. If it is not stated then it tries the first server
17719 certificate file. The previous behaviour hard coded the filename
17724 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17725 a public key to be input or output. For example:
17726 openssl rsa -in key.pem -pubout -out pubkey.pem
17727 Also added necessary DSA public key functions to handle this.
17731 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17732 in the message. This was handled by allowing
17733 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17735 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17737 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17738 to the end of the strings whereas this didn't. This would cause problems
17739 if strings read with d2i_ASN1_bytes() were later modified.
17741 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17743 * Fix for base64 decode bug. When a base64 bio reads only one line of
17744 data and it contains EOF it will end up returning an error. This is
17745 caused by input 46 bytes long. The cause is due to the way base64
17746 BIOs find the start of base64 encoded data. They do this by trying a
17747 trial decode on each line until they find one that works. When they
17748 do a flag is set and it starts again knowing it can pass all the
17749 data directly through the decoder. Unfortunately it doesn't reset
17750 the context it uses. This means that if EOF is reached an attempt
17751 is made to pass two EOFs through the context and this causes the
17752 resulting error. This can also cause other problems as well. As is
17753 usual with these problems it takes *ages* to find and the fix is
17754 trivial: move one line.
17756 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17758 * Ugly workaround to get s_client and s_server working under Windows. The
17759 old code wouldn't work because it needed to select() on sockets and the
17760 tty (for keypresses and to see if data could be written). Win32 only
17761 supports select() on sockets so we select() with a 1s timeout on the
17762 sockets and then see if any characters are waiting to be read, if none
17763 are present then we retry, we also assume we can always write data to
17764 the tty. This isn't nice because the code then blocks until we've
17765 received a complete line of data and it is effectively polling the
17766 keyboard at 1s intervals: however it's quite a bit better than not
17767 working at all :-) A dedicated Windows application might handle this
17768 with an event loop for example.
17772 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17773 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17774 will be called when RSA_sign() and RSA_verify() are used. This is useful
17775 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17776 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17777 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17778 This necessitated the support of an extra signature type NID_md5_sha1
17779 for SSL signatures and modifications to the SSL library to use it instead
17780 of calling RSA_public_decrypt() and RSA_private_encrypt().
17784 * Add new -verify -CAfile and -CApath options to the crl program, these
17785 will lookup a CRL issuers certificate and verify the signature in a
17786 similar way to the verify program. Tidy up the crl program so it
17787 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17788 less strict. It will now permit CRL extensions even if it is not
17789 a V2 CRL: this will allow it to tolerate some broken CRLs.
17793 * Initialize all non-automatic variables each time one of the openssl
17794 sub-programs is started (this is necessary as they may be started
17795 multiple times from the "OpenSSL>" prompt).
17797 *Lennart Bang, Bodo Moeller*
17799 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17800 removing all other RSA functionality (this is what NO_RSA does). This
17801 is so (for example) those in the US can disable those operations covered
17802 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17807 * Non-copying interface to BIO pairs.
17808 (still largely untested)
17812 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17813 ASCII string. This was handled independently in various places before.
17817 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17818 UTF8 strings a character at a time.
17822 * Use client_version from client hello to select the protocol
17823 (s23_srvr.c) and for RSA client key exchange verification
17824 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17828 * Add various utility functions to handle SPKACs, these were previously
17829 handled by poking round in the structure internals. Added new function
17830 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17831 print, verify and generate SPKACs. Based on an original idea from
17832 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17836 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17840 * Allow the config file extension section to be overwritten on the
17841 command line. Based on an original idea from Massimiliano Pala
17842 <madwolf@comune.modena.it>. The new option is called -extensions
17843 and can be applied to ca, req and x509. Also -reqexts to override
17844 the request extensions in req and -crlexts to override the crl extensions
17849 * Add new feature to the SPKAC handling in ca. Now you can include
17850 the same field multiple times by preceding it by "XXXX." for example:
17853 this is the same syntax as used in the req config file.
17857 * Allow certificate extensions to be added to certificate requests. These
17858 are specified in a 'req_extensions' option of the req section of the
17859 config file. They can be printed out with the -text option to req but
17860 are otherwise ignored at present.
17864 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17865 data read consists of only the final block it would not decrypted because
17866 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17867 A misplaced 'break' also meant the decrypted final block might not be
17868 copied until the next read.
17872 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17873 a few extra parameters to the DH structure: these will be useful if
17874 for example we want the value of 'q' or implement X9.42 DH.
17878 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17879 provides hooks that allow the default DSA functions or functions on a
17880 "per key" basis to be replaced. This allows hardware acceleration and
17881 hardware key storage to be handled without major modification to the
17882 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17883 associated functions.
17887 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17888 as "read only": it can't be written to and the buffer it points to will
17889 not be freed. Reading from a read only BIO is much more efficient than
17890 a normal memory BIO. This was added because there are several times when
17891 an area of memory needs to be read from a BIO. The previous method was
17892 to create a memory BIO and write the data to it, this results in two
17893 copies of the data and an O(n^2) reading algorithm. There is a new
17894 function BIO_new_mem_buf() which creates a read only memory BIO from
17895 an area of memory. Also modified the PKCS#7 routines to use read only
17900 * Bugfix: ssl23_get_client_hello did not work properly when called in
17901 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17902 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17903 but a retry condition occurred while trying to read the rest.
17907 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17908 NID_pkcs7_encrypted by default: this was wrong since this should almost
17909 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17910 the encrypted data type: this is a more sensible place to put it and it
17911 allows the PKCS#12 code to be tidied up that duplicated this
17916 * Changed obj_dat.pl script so it takes its input and output files on
17917 the command line. This should avoid shell escape redirection problems
17922 * Initial support for certificate extension requests, these are included
17923 in things like Xenroll certificate requests. Included functions to allow
17924 extensions to be obtained and added.
17928 * -crlf option to s_client and s_server for sending newlines as
17929 CRLF (as required by many protocols).
17933 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17935 * Install libRSAglue.a when OpenSSL is built with RSAref.
17937 *Ralf S. Engelschall*
17939 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17941 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17943 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17948 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17949 DH parameters/keys (q is lost during that conversion, but the resulting
17950 DH parameters contain its length).
17952 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17953 much faster than DH_generate_parameters (which creates parameters
17954 where `p = 2*q + 1`), and also the smaller q makes DH computations
17955 much more efficient (160-bit exponentiation instead of 1024-bit
17956 exponentiation); so this provides a convenient way to support DHE
17957 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17958 utter importance to use
17959 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17961 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17962 when such DH parameters are used, because otherwise small subgroup
17963 attacks may become possible!
17967 * Avoid memory leak in i2d_DHparams.
17971 * Allow the -k option to be used more than once in the enc program:
17972 this allows the same encrypted message to be read by multiple recipients.
17976 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17977 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17978 it will always use the numerical form of the OID, even if it has a short
17983 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17984 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17985 otherwise bn_mod_exp was called. In the case of hardware keys for example
17986 no private key components need be present and it might store extra data
17987 in the RSA structure, which cannot be accessed from bn_mod_exp.
17988 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17989 private key operations.
17993 * Added support for SPARC Linux.
17997 * pem_password_cb function type incompatibly changed from
17998 typedef int pem_password_cb(char *buf, int size, int rwflag);
18000 ....(char *buf, int size, int rwflag, void *userdata);
18001 so that applications can pass data to their callbacks:
18002 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18003 additional void * argument, which is just handed through whenever
18004 the password callback is called.
18006 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18008 New function SSL_CTX_set_default_passwd_cb_userdata.
18010 Compatibility note: As many C implementations push function arguments
18011 onto the stack in reverse order, the new library version is likely to
18012 interoperate with programs that have been compiled with the old
18013 pem_password_cb definition (PEM_whatever takes some data that
18014 happens to be on the stack as its last argument, and the callback
18015 just ignores this garbage); but there is no guarantee whatsoever that
18018 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18019 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18020 problems not only on Windows, but also on some Unix platforms.
18021 To avoid problematic command lines, these definitions are now in an
18022 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18023 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18027 * MIPS III/IV assembler module is reimplemented.
18031 * More DES library cleanups: remove references to srand/rand and
18032 delete an unused file.
18036 * Add support for the free Netwide assembler (NASM) under Win32,
18037 since not many people have MASM (ml) and it can be hard to obtain.
18038 This is currently experimental but it seems to work OK and pass all
18039 the tests. Check out INSTALL.W32 for info.
18043 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18044 without temporary keys kept an extra copy of the server key,
18045 and connections with temporary keys did not free everything in case
18050 * New function RSA_check_key and new openssl rsa option -check
18051 for verifying the consistency of RSA keys.
18053 *Ulf Moeller, Bodo Moeller*
18055 * Various changes to make Win32 compile work:
18056 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18057 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18058 comparison" warnings.
18059 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18063 * Add a debugging option to PKCS#5 v2 key generation function: when
18064 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18065 derived keys are printed to stderr.
18069 * Copy the flags in ASN1_STRING_dup().
18071 *Roman E. Pavlov <pre@mo.msk.ru>*
18073 * The x509 application mishandled signing requests containing DSA
18074 keys when the signing key was also DSA and the parameters didn't match.
18076 It was supposed to omit the parameters when they matched the signing key:
18077 the verifying software was then supposed to automatically use the CA's
18078 parameters if they were absent from the end user certificate.
18080 Omitting parameters is no longer recommended. The test was also
18081 the wrong way round! This was probably due to unusual behaviour in
18082 EVP_cmp_parameters() which returns 1 if the parameters match.
18083 This meant that parameters were omitted when they *didn't* match and
18084 the certificate was useless. Certificates signed with 'ca' didn't have
18087 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18089 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18090 The interface is as follows:
18091 Applications can use
18092 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18093 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18094 "off" is now the default.
18095 The library internally uses
18096 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18097 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18098 to disable memory-checking temporarily.
18100 Some inconsistent states that previously were possible (and were
18101 even the default) are now avoided.
18103 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18104 with each memory chunk allocated; this is occasionally more helpful
18105 than just having a counter.
18107 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18109 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18114 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18115 which largely parallels "options", but is for changing API behaviour,
18116 whereas "options" are about protocol behaviour.
18117 Initial "mode" flags are:
18119 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18120 a single record has been written.
18121 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18122 retries use the same buffer location.
18123 (But all of the contents must be
18128 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18131 * Fix problems with no-hmac etc.
18133 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18135 * New functions RSA_get_default_method(), RSA_set_method() and
18136 RSA_get_method(). These allows replacement of RSA_METHODs without having
18137 to mess around with the internals of an RSA structure.
18141 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18142 Also really enable memory leak checks in openssl.c and in some
18145 *Chad C. Mulligan, Bodo Moeller*
18147 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18148 up the length of negative integers. This has now been simplified to just
18149 store the length when it is first determined and use it later, rather
18150 than trying to keep track of where data is copied and updating it to
18152 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18154 * Add a new function PKCS7_signatureVerify. This allows the verification
18155 of a PKCS#7 signature but with the signing certificate passed to the
18156 function itself. This contrasts with PKCS7_dataVerify which assumes the
18157 certificate is present in the PKCS#7 structure. This isn't always the
18158 case: certificates can be omitted from a PKCS#7 structure and be
18159 distributed by "out of band" means (such as a certificate database).
18163 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18164 function prototypes in pem.h, also change util/mkdef.pl to add the
18165 necessary function names.
18169 * mk1mf.pl (used by Windows builds) did not properly read the
18170 options set by Configure in the top level Makefile, and Configure
18171 was not even able to write more than one option correctly.
18172 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18176 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18177 file to be loaded from a BIO or FILE pointer. The BIO version will
18178 for example allow memory BIOs to contain config info.
18182 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18183 Whoever hopes to achieve shared-library compatibility across versions
18184 must use this, not the compile-time macro.
18185 (Exercise 0.9.4: Which is the minimum library version required by
18187 Note: All this applies only to multi-threaded programs, others don't
18192 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18193 through a BIO pair triggered the default case, i.e.
18194 SSLerr(...,SSL_R_UNKNOWN_STATE).
18198 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18199 can use the SSL library even if none of the specific BIOs is
18204 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18205 for the encoded length.
18207 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18209 * Add initial documentation of the X509V3 functions.
18213 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18214 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18215 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18216 secure PKCS#8 private key format with a high iteration count.
18220 * Fix determination of Perl interpreter: A perl or perl5
18221 *directory* in $PATH was also accepted as the interpreter.
18223 *Ralf S. Engelschall*
18225 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18226 wrong with it but it was very old and did things like calling
18227 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18228 unusual formatting.
18232 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18233 to use the new extension code.
18237 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18238 with macros. This should make it easier to change their form, add extra
18239 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18244 * Add to configuration table a new entry that can specify an alternative
18245 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18246 according to Mark Crispin <MRC@Panda.COM>.
18250 * DES CBC did not update the IV. Weird.
18254 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18255 Changing the behaviour of the former might break existing programs --
18256 where IV updating is needed, des_ncbc_encrypt can be used.
18259 * When bntest is run from "make test" it drives bc to check its
18260 calculations, as well as internally checking them. If an internal check
18261 fails, it needs to cause bc to give a non-zero result or make test carries
18262 on without noticing the failure. Fixed.
18266 * DES library cleanups.
18270 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18271 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18272 ciphers. NOTE: although the key derivation function has been verified
18273 against some published test vectors it has not been extensively tested
18274 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18279 * Instead of "mkdir -p", which is not fully portable, use new
18280 Perl script "util/mkdir-p.pl".
18284 * Rewrite the way password based encryption (PBE) is handled. It used to
18285 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18286 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18287 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18288 the 'parameter' field of the AlgorithmIdentifier is passed to the
18289 underlying key generation function so it must do its own ASN1 parsing.
18290 This has also changed the EVP_PBE_CipherInit() function which now has a
18291 'parameter' argument instead of literal salt and iteration count values
18292 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18296 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18297 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18298 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18299 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18300 value was just used as a "magic string" and not used directly its
18301 value doesn't matter.
18305 * Introduce some semblance of const correctness to BN. Shame C doesn't
18310 * "linux-sparc64" configuration (ultrapenguin).
18312 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18313 "linux-sparc" configuration.
18315 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18317 * config now generates no-xxx options for missing ciphers.
18321 * Support the EBCDIC character set (work in progress).
18322 File ebcdic.c not yet included because it has a different license.
18324 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18326 * Support BS2000/OSD-POSIX.
18328 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18330 * Make callbacks for key generation use `void *` instead of `char *`.
18334 * Make S/MIME samples compile (not yet tested).
18338 * Additional typesafe stacks.
18342 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18346 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18348 * New configuration variant "sco5-gcc".
18350 * Updated some demos.
18352 *Sean O Riordain, Wade Scholine*
18354 * Add missing BIO_free at exit of pkcs12 application.
18358 * Fix memory leak in conf.c.
18362 * Updates for Win32 to assembler version of MD5.
18366 * Set #! path to perl in `apps/der_chop` to where we found it
18367 instead of using a fixed path.
18371 * SHA library changes for irix64-mips4-cc.
18375 * Improvements for VMS support.
18379 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18381 * Bignum library bug fix. IRIX 6 passes "make test" now!
18382 This also avoids the problems with SC4.2 and unpatched SC5.
18384 *Andy Polyakov <appro@fy.chalmers.se>*
18386 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18387 These are required because of the typesafe stack would otherwise break
18388 existing code. If old code used a structure member which used to be STACK
18389 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18390 sk_num or sk_value it would produce an error because the num, data members
18391 are not present in STACK_OF. Now it just produces a warning. sk_set
18392 replaces the old method of assigning a value to sk_value
18393 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18394 that does this will no longer work (and should use sk_set instead) but
18395 this could be regarded as a "questionable" behaviour anyway.
18399 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18400 correctly handle encrypted S/MIME data.
18404 * Change type of various DES function arguments from des_cblock
18405 (which means, in function argument declarations, pointer to char)
18406 to des_cblock * (meaning pointer to array with 8 char elements),
18407 which allows the compiler to do more typechecking; it was like
18408 that back in SSLeay, but with lots of ugly casts.
18410 Introduce new type const_des_cblock.
18414 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18415 problems: find RecipientInfo structure that matches recipient certificate
18416 and initialise the ASN1 structures properly based on passed cipher.
18420 * Belatedly make the BN tests actually check the results.
18424 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18425 to and from BNs: it was completely broken. New compilation option
18426 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18427 key elements as negative integers.
18431 * Reorganize and speed up MD5.
18433 *Andy Polyakov <appro@fy.chalmers.se>*
18437 *Richard Levitte <richard@levitte.org>*
18439 * New option -out to asn1parse to allow the parsed structure to be
18440 output to a file. This is most useful when combined with the -strparse
18441 option to examine the output of things like OCTET STRINGS.
18445 * Make SSL library a little more fool-proof by not requiring any longer
18446 that `SSL_set_{accept,connect}_state` be called before
18447 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18448 in many applications because usually everything *appeared* to work as
18449 intended anyway -- now it really works as intended).
18453 * Move openssl.cnf out of lib/.
18457 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18458 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18459 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18461 *Ralf S. Engelschall*
18463 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18464 handle PKCS#7 enveloped data properly.
18466 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18468 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18469 copying pointers. The cert_st handling is changed by this in
18470 various ways (and thus what used to be known as ctx->default_cert
18471 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18472 any longer when s->cert does not give us what we need).
18473 ssl_cert_instantiate becomes obsolete by this change.
18474 As soon as we've got the new code right (possibly it already is?),
18475 we have solved a couple of bugs of the earlier code where s->cert
18476 was used as if it could not have been shared with other SSL structures.
18478 Note that using the SSL API in certain dirty ways now will result
18479 in different behaviour than observed with earlier library versions:
18480 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18481 does not influence s as it used to.
18483 In order to clean up things more thoroughly, inside SSL_SESSION
18484 we don't use CERT any longer, but a new structure SESS_CERT
18485 that holds per-session data (if available); currently, this is
18486 the peer's certificate chain and, for clients, the server's certificate
18487 and temporary key. CERT holds only those values that can have
18488 meaningful defaults in an SSL_CTX.
18492 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18493 from the internal representation. Various PKCS#7 fixes: remove some
18494 evil casts and set the enc_dig_alg field properly based on the signing
18499 * Allow PKCS#12 password to be set from the command line or the
18500 environment. Let 'ca' get its config file name from the environment
18501 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18506 * Allow certificate policies extension to use an IA5STRING for the
18507 organization field. This is contrary to the PKIX definition but
18508 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18513 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18514 without disallowing inline assembler and the like for non-pedantic builds.
18518 * Support Borland C++ builder.
18520 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18526 * SHA-1 cleanups and performance enhancements.
18528 *Andy Polyakov <appro@fy.chalmers.se>*
18530 * Sparc v8plus assembler for the bignum library.
18532 *Andy Polyakov <appro@fy.chalmers.se>*
18534 * Accept any -xxx and +xxx compiler options in Configure.
18538 * Update HPUX configuration.
18542 * Add missing `sk_<type>_unshift()` function to safestack.h
18544 *Ralf S. Engelschall*
18546 * New function SSL_CTX_use_certificate_chain_file that sets the
18547 "extra_cert"s in addition to the certificate. (This makes sense
18548 only for "PEM" format files, as chains as a whole are not
18553 * Support verify_depth from the SSL API.
18554 x509_vfy.c had what can be considered an off-by-one-error:
18555 Its depth (which was not part of the external interface)
18556 was actually counting the number of certificates in a chain;
18557 now it really counts the depth.
18561 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18562 instead of X509err, which often resulted in confusing error
18563 messages since the error codes are not globally unique
18564 (e.g. an alleged error in ssl3_accept when a certificate
18565 didn't match the private key).
18567 * New function SSL_CTX_set_session_id_context that allows to set a default
18568 value (so that you don't need SSL_set_session_id_context for each
18569 connection using the SSL_CTX).
18573 * OAEP decoding bug fix.
18577 * Support INSTALL_PREFIX for package builders, as proposed by
18582 * New Configure options "threads" and "no-threads". For systems
18583 where the proper compiler options are known (currently Solaris
18584 and Linux), "threads" is the default.
18588 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18592 * Install various scripts to $(OPENSSLDIR)/misc, not to
18593 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18594 such as /usr/local/bin.
18598 * "make linux-shared" to build shared libraries.
18600 *Niels Poppe <niels@netbox.org>*
18602 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18606 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18607 extension adding in x509 utility.
18611 * Remove NOPROTO sections and error code comments.
18615 * Partial rewrite of the DEF file generator to now parse the ANSI
18620 * New Configure options --prefix=DIR and --openssldir=DIR.
18624 * Complete rewrite of the error code script(s). It is all now handled
18625 by one script at the top level which handles error code gathering,
18626 header rewriting and C source file generation. It should be much better
18627 than the old method: it now uses a modified version of Ulf's parser to
18628 read the ANSI prototypes in all header files (thus the old K&R definitions
18629 aren't needed for error creation any more) and do a better job of
18630 translating function codes into names. The old 'ASN1 error code embedded
18631 in a comment' is no longer necessary and it doesn't use .err files which
18632 have now been deleted. Also the error code call doesn't have to appear all
18633 on one line (which resulted in some large lines...).
18637 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18641 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18642 0 (which usually indicates a closed connection), but continue reading.
18646 * Fix some race conditions.
18650 * Add support for CRL distribution points extension. Add Certificate
18651 Policies and CRL distribution points documentation.
18655 * Move the autogenerated header file parts to crypto/opensslconf.h.
18659 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18660 8 of keying material. Merlin has also confirmed interop with this fix
18661 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18663 *Merlin Hughes <merlin@baltimore.ie>*
18665 * Fix lots of warnings.
18667 *Richard Levitte <levitte@stacken.kth.se>*
18669 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18670 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18672 *Richard Levitte <levitte@stacken.kth.se>*
18674 * Fix problems with sizeof(long) == 8.
18676 *Andy Polyakov <appro@fy.chalmers.se>*
18678 * Change functions to ANSI C.
18682 * Fix typos in error codes.
18684 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18686 * Remove defunct assembler files from Configure.
18690 * SPARC v8 assembler BIGNUM implementation.
18692 *Andy Polyakov <appro@fy.chalmers.se>*
18694 * Support for Certificate Policies extension: both print and set.
18695 Various additions to support the r2i method this uses.
18699 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18700 return a const string when you are expecting an allocated buffer.
18704 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18705 types DirectoryString and DisplayText.
18709 * Add code to allow r2i extensions to access the configuration database,
18710 add an LHASH database driver and add several ctx helper functions.
18714 * Fix an evil bug in bn_expand2() which caused various BN functions to
18715 fail when they extended the size of a BIGNUM.
18719 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18720 support typesafe stack.
18724 * Fix typo in SSL_[gs]et_options().
18726 *Nils Frostberg <nils@medcom.se>*
18728 * Delete various functions and files that belonged to the (now obsolete)
18729 old X509V3 handling code.
18733 * New Configure option "rsaref".
18737 * Don't auto-generate pem.h.
18741 * Introduce type-safe ASN.1 SETs.
18745 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18747 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18749 * Introduce type-safe STACKs. This will almost certainly break lots of code
18750 that links with OpenSSL (well at least cause lots of warnings), but fear
18751 not: the conversion is trivial, and it eliminates loads of evil casts. A
18752 few STACKed things have been converted already. Feel free to convert more.
18753 In the fullness of time, I'll do away with the STACK type altogether.
18757 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18758 specified in `<certfile>` by updating the entry in the index.txt file.
18759 This way one no longer has to edit the index.txt file manually for
18760 revoking a certificate. The -revoke option does the gory details now.
18762 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18764 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18765 `-text` option at all and this way the `-noout -text` combination was
18766 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18768 *Ralf S. Engelschall*
18770 * Make sure a corresponding plain text error message exists for the
18771 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18772 verify callback function determined that a certificate was revoked.
18774 *Ralf S. Engelschall*
18776 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18777 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18778 all available ciphers including rc5, which was forgotten until now.
18779 In order to let the testing shell script know which algorithms
18780 are available, a new (up to now undocumented) command
18781 `openssl list-cipher-commands` is used.
18785 * Bugfix: s_client occasionally would sleep in select() when
18786 it should have checked SSL_pending() first.
18790 * New functions DSA_do_sign and DSA_do_verify to provide access to
18791 the raw DSA values prior to ASN.1 encoding.
18795 * Tweaks to Configure
18797 *Niels Poppe <niels@netbox.org>*
18799 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18804 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18808 * New config option to avoid instructions that are illegal on the 80386.
18809 The default code is faster, but requires at least a 486.
18813 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18814 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18815 same as SSL2_VERSION anyway.
18819 * New "-showcerts" option for s_client.
18823 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18824 application. Various cleanups and fixes.
18828 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18829 modify error routines to work internally. Add error codes and PBE init
18830 to library startup routines.
18834 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18835 packing functions to asn1 and evp. Changed function names and error
18836 codes along the way.
18840 * PKCS12 integration: and so it begins... First of several patches to
18841 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18842 objects to objects.h
18846 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18847 and display support for Thawte strong extranet extension.
18851 * Add LinuxPPC support.
18853 *Jeff Dubrule <igor@pobox.org>*
18855 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18856 bn_div_words in alpha.s.
18858 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18860 * Make sure the RSA OAEP test is skipped under -DRSAref because
18861 OAEP isn't supported when OpenSSL is built with RSAref.
18863 *Ulf Moeller <ulf@fitug.de>*
18865 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18866 so they no longer are missing under -DNOPROTO.
18868 *Soren S. Jorvang <soren@t.dk>*
18870 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18872 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18873 doesn't work when the session is reused. Coming soon!
18877 * Fix a security hole, that allows sessions to be reused in the wrong
18878 context thus bypassing client cert protection! All software that uses
18879 client certs and session caches in multiple contexts NEEDS PATCHING to
18880 allow session reuse! A fuller solution is in the works.
18882 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18884 * Some more source tree cleanups (removed obsolete files
18885 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18886 permission on "config" script to be executable) and a fix for the INSTALL
18889 *Ulf Moeller <ulf@fitug.de>*
18891 * Remove some legacy and erroneous uses of malloc, free instead of
18894 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18896 * Make rsa_oaep_test return non-zero on error.
18898 *Ulf Moeller <ulf@fitug.de>*
18900 * Add support for native Solaris shared libraries. Configure
18901 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18902 if someone would make that last step automatic.
18904 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18906 * ctx_size was not built with the right compiler during "make links". Fixed.
18910 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18911 except NULL ciphers". This means the default cipher list will no longer
18912 enable NULL ciphers. They need to be specifically enabled e.g. with
18913 the string "DEFAULT:eNULL".
18917 * Fix to RSA private encryption routines: if p < q then it would
18918 occasionally produce an invalid result. This will only happen with
18919 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18923 * Be less restrictive and allow also `perl util/perlpath.pl
18924 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18925 because this way one can also use an interpreter named `perl5` (which is
18926 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18927 installed as `perl`).
18929 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18931 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18933 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18935 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18936 advapi32.lib to Win32 build and change the pem test comparison
18937 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18938 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18939 and crypto/des/ede_cbcm_enc.c.
18943 * DES quad checksum was broken on big-endian architectures. Fixed.
18947 * Comment out two functions in bio.h that aren't implemented. Fix up the
18948 Win32 test batch file so it (might) work again. The Win32 test batch file
18949 is horrible: I feel ill....
18953 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18954 in e_os.h. Audit of header files to check ANSI and non ANSI
18955 sections: 10 functions were absent from non ANSI section and not exported
18956 from Windows DLLs. Fixed up libeay.num for new functions.
18960 * Make `openssl version` output lines consistent.
18962 *Ralf S. Engelschall*
18964 * Fix Win32 symbol export lists for BIO functions: Added
18965 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18966 to ms/libeay{16,32}.def.
18968 *Ralf S. Engelschall*
18970 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18971 fine under Unix and passes some trivial tests I've now added. But the
18972 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18973 added to make sure no one expects that this stuff really works in the
18974 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18975 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18978 *Ralf S. Engelschall*
18980 * Fix the generation of two part addresses in perl.
18982 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18984 * Add config entry for Linux on MIPS.
18986 *John Tobey <jtobey@channel1.com>*
18988 * Make links whenever Configure is run, unless we are on Windoze.
18992 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18993 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18998 * Add a useful kludge to allow package maintainers to specify compiler and
18999 other platforms details on the command line without having to patch the
19000 Configure script every time: One now can use
19001 `perl Configure <id>:<details>`,
19002 i.e. platform ids are allowed to have details appended
19003 to them (separated by colons). This is treated as there would be a static
19004 pre-configured entry in Configure's %table under key `<id>` with value
19005 `<details>` and `perl Configure <id>` is called. So, when you want to
19006 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19007 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19008 now, which overrides the FreeBSD-elf entry on-the-fly.
19010 *Ralf S. Engelschall*
19012 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19016 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19017 on the `perl Configure ...` command line. This way one can compile
19018 OpenSSL libraries with Position Independent Code (PIC) which is needed
19019 for linking it into DSOs.
19021 *Ralf S. Engelschall*
19023 * Remarkably, export ciphers were totally broken and no-one had noticed!
19028 * Cleaned up the LICENSE document: The official contact for any license
19029 questions now is the OpenSSL core team under openssl-core@openssl.org.
19030 And add a paragraph about the dual-license situation to make sure people
19031 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19032 to the OpenSSL toolkit.
19034 *Ralf S. Engelschall*
19036 * General source tree makefile cleanups: Made `making xxx in yyy...`
19037 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19038 Additionally cleaned up the `make links` target: Remove unnecessary
19039 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19040 to speed processing and no longer clutter the display with confusing
19041 stuff. Instead only the actually done links are displayed.
19043 *Ralf S. Engelschall*
19045 * Permit null encryption ciphersuites, used for authentication only. It used
19046 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19047 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19052 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19053 signed attributes when verifying signatures (this would break them),
19054 the detached data encoding was wrong and public keys obtained using
19055 X509_get_pubkey() weren't freed.
19059 * Add text documentation for the BUFFER functions. Also added a work around
19060 to a Win95 console bug. This was triggered by the password read stuff: the
19061 last character typed gets carried over to the next fread(). If you were
19062 generating a new cert request using 'req' for example then the last
19063 character of the passphrase would be CR which would then enter the first
19068 * Added the new 'Includes OpenSSL Cryptography Software' button as
19069 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19070 button and can be used by applications based on OpenSSL to show the
19071 relationship to the OpenSSL project.
19073 *Ralf S. Engelschall*
19075 * Remove confusing variables in function signatures in files
19076 ssl/ssl_lib.c and ssl/ssl.h.
19078 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19080 * Don't install bss_file.c under PREFIX/include/
19082 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19084 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19085 functions that return function pointers and has support for NT specific
19086 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19087 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19088 unsigned to signed types: this was killing the Win32 compile.
19092 * Add new certificate file to stack functions,
19093 SSL_add_dir_cert_subjects_to_stack() and
19094 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19095 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19096 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19097 This means that Apache-SSL and similar packages don't have to mess around
19098 to add as many CAs as they want to the preferred list.
19102 * Experiment with doxygen documentation. Currently only partially applied to
19104 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19105 openssl.doxy as the configuration file.
19109 * Get rid of remaining C++-style comments which strict C compilers hate.
19111 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19113 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19114 compiled in by default: it has problems with large keys.
19118 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19119 DH private keys and/or callback functions which directly correspond to
19120 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19121 is needed for applications which have to configure certificates on a
19122 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19124 For the RSA certificate situation is makes no difference, but
19125 for the DSA certificate situation this fixes the "no shared cipher"
19126 problem where the OpenSSL cipher selection procedure failed because the
19127 temporary keys were not overtaken from the context and the API provided
19128 no way to reconfigure them.
19129 The new functions now let applications reconfigure the stuff and they
19130 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19131 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19132 non-public-API function ssl_cert_instantiate() is used as a helper
19133 function and also to reduce code redundancy inside ssl_rsa.c.
19135 *Ralf S. Engelschall*
19137 * Move s_server -dcert and -dkey options out of the undocumented feature
19138 area because they are useful for the DSA situation and should be
19139 recognized by the users.
19141 *Ralf S. Engelschall*
19143 * Fix the cipher decision scheme for export ciphers: the export bits are
19144 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19145 SSL_EXP_MASK. So, the original variable has to be used instead of the
19146 already masked variable.
19148 *Richard Levitte <levitte@stacken.kth.se>*
19150 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19152 *Richard Levitte <levitte@stacken.kth.se>*
19154 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19155 from `int` to `unsigned int` because it is a length and initialized by
19156 EVP_DigestFinal() which expects an `unsigned int *`.
19158 *Richard Levitte <levitte@stacken.kth.se>*
19160 * Don't hard-code path to Perl interpreter on shebang line of Configure
19161 script. Instead use the usual Shell->Perl transition trick.
19163 *Ralf S. Engelschall*
19165 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19166 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19167 -noout -modulus` as it's already the case for `openssl rsa -noout
19168 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19169 currently the public key is printed (a decision which was already done by
19170 `openssl dsa -modulus` in the past) which serves a similar purpose.
19171 Additionally the NO_RSA no longer completely removes the whole -modulus
19172 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19175 *Ralf S. Engelschall*
19177 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19178 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19180 *Arne Ansper <arne@ats.cyber.ee>*
19182 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19183 to be added. Now both 'req' and 'ca' can use new objects defined in the
19188 * Add cool BIO that does syslog (or event log on NT).
19190 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19192 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19193 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19194 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19195 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19199 * Add preliminary config info for new extension code.
19203 * Make RSA_NO_PADDING really use no padding.
19205 *Ulf Moeller <ulf@fitug.de>*
19207 * Generate errors when private/public key check is done.
19211 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19212 for some CRL extensions and new objects added.
19216 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19217 key usage extension and fuller support for authority key id.
19221 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19222 padding method for RSA, which is recommended for new applications in PKCS
19223 #1 v2.0 (RFC 2437, October 1998).
19224 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19225 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19226 against Bleichbacher's attack on RSA.
19227 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19230 * Updates to the new SSL compression code
19232 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19234 * Fix so that the version number in the master secret, when passed
19235 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19236 (because the server will not accept higher), that the version number
19237 is 0x03,0x01, not 0x03,0x00
19239 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19241 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19242 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19243 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19247 * Support for RAW extensions where an arbitrary extension can be
19248 created by including its DER encoding. See `apps/openssl.cnf` for
19253 * Make sure latest Perl versions don't interpret some generated C array
19254 code as Perl array code in the crypto/err/err_genc.pl script.
19256 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19258 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19259 not many people have the assembler. Various Win32 compilation fixes and
19260 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19261 build instructions.
19265 * Modify configure script 'Configure' to automatically create crypto/date.h
19266 file under Win32 and also build pem.h from pem.org. New script
19267 util/mkfiles.pl to create the MINFO file on environments that can't do a
19268 'make files': perl util/mkfiles.pl >MINFO should work.
19272 * Major rework of DES function declarations, in the pursuit of correctness
19273 and purity. As a result, many evil casts evaporated, and some weirdness,
19274 too. You may find this causes warnings in your code. Zapping your evil
19275 casts will probably fix them. Mostly.
19279 * Fix for a typo in asn1.h. Bug fix to object creation script
19280 obj_dat.pl. It considered a zero in an object definition to mean
19281 "end of object": none of the objects in objects.h have any zeros
19282 so it wasn't spotted.
19284 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19286 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19287 Masking (CBCM). In the absence of test vectors, the best I have been able
19288 to do is check that the decrypt undoes the encrypt, so far. Send me test
19289 vectors if you have them.
19293 * Correct calculation of key length for export ciphers (too much space was
19294 allocated for null ciphers). This has not been tested!
19298 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19299 message is now correct (it understands "crypto" and "ssl" on its
19300 command line). There is also now an "update" option. This will update
19301 the util/ssleay.num and util/libeay.num files with any new functions.
19303 perl util/mkdef.pl crypto ssl update
19304 it will update them.
19308 * Overhauled the Perl interface:
19309 - ported BN stuff to OpenSSL's different BN library
19310 - made the perl/ source tree CVS-aware
19311 - renamed the package from SSLeay to OpenSSL (the files still contain
19312 their history because I've copied them in the repository)
19313 - removed obsolete files (the test scripts will be replaced
19314 by better Test::Harness variants in the future)
19316 *Ralf S. Engelschall*
19318 * First cut for a very conservative source tree cleanup:
19319 1. merge various obsolete readme texts into doc/ssleay.txt
19320 where we collect the old documents and readme texts.
19321 2. remove the first part of files where I'm already sure that we no
19322 longer need them because of three reasons: either they are just temporary
19323 files which were left by Eric or they are preserved original files where
19324 I've verified that the diff is also available in the CVS via "cvs diff
19325 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19326 the crypto/md/ stuff).
19328 *Ralf S. Engelschall*
19330 * More extension code. Incomplete support for subject and issuer alt
19331 name, issuer and authority key id. Change the i2v function parameters
19332 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19333 what that's for :-) Fix to ASN1 macro which messed up
19334 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19338 * Preliminary support for ENUMERATED type. This is largely copied from the
19343 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19345 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19347 * Make sure `make rehash` target really finds the `openssl` program.
19349 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19351 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19352 like to hear about it if this slows down other processors.
19356 * Add CygWin32 platform information to Configure script.
19358 *Alan Batie <batie@aahz.jf.intel.com>*
19360 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19362 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19364 * New program nseq to manipulate netscape certificate sequences
19368 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19373 * Fixes to BN code. Previously the default was to define BN_RECURSION
19374 but the BN code had some problems that would cause failures when
19375 doing certificate verification and some other functions.
19377 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19379 * Add ASN1 and PEM code to support netscape certificate sequences.
19383 * Add ASN1 and PEM code to support netscape certificate sequences.
19387 * Add several PKIX and private extended key usage OIDs.
19391 * Modify the 'ca' program to handle the new extension code. Modify
19392 openssl.cnf for new extension format, add comments.
19396 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19397 and add a sample to openssl.cnf so req -x509 now adds appropriate
19402 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19403 error code, add initial support to X509_print() and x509 application.
19407 * Takes a deep breath and start adding X509 V3 extension support code. Add
19408 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19409 stuff is currently isolated and isn't even compiled yet.
19413 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19414 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19415 Removed the versions check from X509 routines when loading extensions:
19416 this allows certain broken certificates that don't set the version
19417 properly to be processed.
19421 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19422 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19423 can still be regenerated with "make depend".
19427 * Spelling mistake in C version of CAST-128.
19429 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19431 * Changes to the error generation code. The perl script err-code.pl
19432 now reads in the old error codes and retains the old numbers, only
19433 adding new ones if necessary. It also only changes the .err files if new
19434 codes are added. The makefiles have been modified to only insert errors
19435 when needed (to avoid needlessly modifying header files). This is done
19436 by only inserting errors if the .err file is newer than the auto generated
19437 C file. To rebuild all the error codes from scratch (the old behaviour)
19438 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19439 or delete all the .err files.
19443 * CAST-128 was incorrectly implemented for short keys. The C version has
19444 been fixed, but is untested. The assembler versions are also fixed, but
19445 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19446 to regenerate it if needed.
19447 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19448 Hagino <itojun@kame.net>*
19450 * File was opened incorrectly in randfile.c.
19452 *Ulf Möller <ulf@fitug.de>*
19454 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19455 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19456 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19457 al: it's just almost always a UTCTime. Note this patch adds new error
19458 codes so do a "make errors" if there are problems.
19462 * Correct Linux 1 recognition in config.
19464 *Ulf Möller <ulf@fitug.de>*
19466 * Remove pointless MD5 hash when using DSA keys in ca.
19468 *Anonymous <nobody@replay.com>*
19470 * Generate an error if given an empty string as a cert directory. Also
19471 generate an error if handed NULL (previously returned 0 to indicate an
19472 error, but didn't set one).
19474 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19476 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19480 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19481 parameters. This was causing a warning which killed off the Win32 compile.
19485 * Remove C++ style comments from crypto/bn/bn_local.h.
19487 *Neil Costigan <neil.costigan@celocom.com>*
19489 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19490 based on a text string, looking up short and long names and finally
19491 "dot" format. The "dot" format stuff didn't work. Added new function
19492 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19493 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19494 OID is not part of the table.
19498 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19499 X509_LOOKUP_by_alias().
19503 * Sort openssl functions by name.
19507 * Get the `gendsa` command working and add it to the `list` command. Remove
19508 encryption from sample DSA keys (in case anyone is interested the password
19513 * Make *all* `*_free` functions accept a NULL pointer.
19515 *Frans Heymans <fheymans@isaserver.be>*
19517 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19520 *Anonymous <nobody@replay.com>*
19522 * s_server should send the CAfile as acceptable CAs, not its own cert.
19524 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19526 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19528 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19530 * Temp key "for export" tests were wrong in s3_srvr.c.
19532 *Anonymous <nobody@replay.com>*
19534 * Add prototype for temp key callback functions
19535 SSL_CTX_set_tmp_{rsa,dh}_callback().
19539 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19540 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19544 * X509_name_add_entry() freed the wrong thing after an error.
19546 *Arne Ansper <arne@ats.cyber.ee>*
19548 * rsa_eay.c would attempt to free a NULL context.
19550 *Arne Ansper <arne@ats.cyber.ee>*
19552 * BIO_s_socket() had a broken should_retry() on Windoze.
19554 *Arne Ansper <arne@ats.cyber.ee>*
19556 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19558 *Arne Ansper <arne@ats.cyber.ee>*
19560 * Make sure the already existing X509_STORE->depth variable is initialized
19561 in X509_STORE_new(), but document the fact that this variable is still
19562 unused in the certificate verification process.
19564 *Ralf S. Engelschall*
19566 * Fix the various library and `apps/` files to free up pkeys obtained from
19567 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19571 * Fix reference counting in X509_PUBKEY_get(). This makes
19572 demos/maurice/example2.c work, amongst others, probably.
19574 *Steve Henson and Ben Laurie*
19576 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19577 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19578 are no longer created. This way we have a single and consistent command
19579 line interface `openssl <command>`, similar to `cvs <command>`.
19581 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19583 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19584 BIT STRING wrapper always have zero unused bits.
19588 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19592 * Make the top-level INSTALL documentation easier to understand.
19596 * Makefiles updated to exit if an error occurs in a sub-directory
19597 make (including if user presses ^C) [Paul Sutton]
19599 * Make Montgomery context stuff explicit in RSA data structure.
19603 * Fix build order of pem and err to allow for generated pem.h.
19607 * Fix renumbering bug in X509_NAME_delete_entry().
19611 * Enhanced the err-ins.pl script so it makes the error library number
19612 global and can add a library name. This is needed for external ASN1 and
19613 other error libraries.
19617 * Fixed sk_insert which never worked properly.
19621 * Fix ASN1 macros so they can handle indefinite length constructed
19622 EXPLICIT tags. Some non standard certificates use these: they can now
19627 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19628 into a single doc/ssleay.txt bundle. This way the information is still
19629 preserved but no longer messes up this directory. Now it's new room for
19630 the new set of documentation files.
19632 *Ralf S. Engelschall*
19634 * SETs were incorrectly DER encoded. This was a major pain, because they
19635 shared code with SEQUENCEs, which aren't coded the same. This means that
19636 almost everything to do with SETs or SEQUENCEs has either changed name or
19637 number of arguments.
19639 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19641 * Fix test data to work with the above.
19645 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19646 was already fixed by Eric for 0.9.1 it seems.
19648 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19650 * Autodetect FreeBSD3.
19654 * Fix various bugs in Configure. This affects the following platforms:
19658 unixware-2.0-pentium
19663 * Eliminate generated files from CVS. Reorder tests to regenerate files
19664 before they are needed.
19668 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19672 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19674 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19675 changed SSLeay to OpenSSL in version strings.
19677 *Ralf S. Engelschall*
19679 * Some fixups to the top-level documents.
19683 * Fixed the nasty bug where rsaref.h was not found under compile-time
19684 because the symlink to include/ was missing.
19686 *Ralf S. Engelschall*
19688 * Incorporated the popular no-RSA/DSA-only patches
19689 which allow to compile an RSA-free SSLeay.
19691 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19693 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19694 when "ssleay" is still not found.
19696 *Ralf S. Engelschall*
19698 * Added more platforms to Configure: Cray T3E, HPUX 11,
19700 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19702 * Updated the README file.
19704 *Ralf S. Engelschall*
19706 * Added various .cvsignore files in the CVS repository subdirs
19707 to make a "cvs update" really silent.
19709 *Ralf S. Engelschall*
19711 * Recompiled the error-definition header files and added
19712 missing symbols to the Win32 linker tables.
19714 *Ralf S. Engelschall*
19716 * Cleaned up the top-level documents;
19717 o new files: CHANGES and LICENSE
19718 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19719 o merged COPYRIGHT into LICENSE
19720 o removed obsolete TODO file
19721 o renamed MICROSOFT to INSTALL.W32
19723 *Ralf S. Engelschall*
19725 * Removed dummy files from the 0.9.1b source tree:
19726 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19727 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19728 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19729 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19730 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19732 *Ralf S. Engelschall*
19734 * Added various platform portability fixes.
19738 * The Genesis of the OpenSSL rpject:
19739 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19740 Young and Tim J. Hudson created while they were working for C2Net until
19743 *The OpenSSL Project*
19745 ### Changes between 0.9.0b and 0.9.1b [not released]
19747 * Updated a few CA certificates under certs/
19751 * Changed some BIGNUM api stuff.
19755 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19756 DGUX x86, Linux Alpha, etc.
19760 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19761 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19766 * Add -strparse option to asn1pars program which parses nested
19769 *Dr Stephen Henson <shenson@bigfoot.com>*
19771 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19775 * DSA fix for "ca" program.
19779 * Added "-genkey" option to "dsaparam" program.
19783 * Added RIPE MD160 (rmd160) message digest.
19787 * Added -a (all) option to "ssleay version" command.
19791 * Added PLATFORM define which is the id given to Configure.
19795 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19799 * Extended the ASN.1 parser routines.
19803 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19807 * Added a BN_CTX to the BN library.
19811 * Fixed the weak key values in DES library
19815 * Changed API in EVP library for cipher aliases.
19819 * Added support for RC2/64bit cipher.
19823 * Converted the lhash library to the crypto/mem.c functions.
19827 * Added more recognized ASN.1 object ids.
19831 * Added more RSA padding checks for SSL/TLS.
19835 * Added BIO proxy/filter functionality.
19839 * Added extra_certs to SSL_CTX which can be used
19840 send extra CA certificates to the client in the CA cert chain sending
19841 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19845 * Now Fortezza is denied in the authentication phase because
19846 this is key exchange mechanism is not supported by SSLeay at all.
19850 * Additional PKCS1 checks.
19854 * Support the string "TLSv1" for all TLS v1 ciphers.
19858 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19859 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19863 * Fixed a few memory leaks.
19867 * Fixed various code and comment typos.
19871 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19872 bytes sent in the client random.
19874 *Edward Bishop <ebishop@spyglass.com>*
19878 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
19879 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
19880 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
19881 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
19882 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
19883 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
19884 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
19885 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
19886 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
19887 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19888 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
19889 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19890 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19891 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19892 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19893 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19894 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19895 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19896 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19897 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19898 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19899 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19900 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19901 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19902 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19903 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19904 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19905 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19906 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19907 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19908 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19909 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19910 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19911 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19912 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19913 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19914 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19915 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19916 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19917 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19918 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19919 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19920 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19921 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19922 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19923 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19924 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19925 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19926 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19927 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19928 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19929 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19930 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19931 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19932 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19933 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19934 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19935 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19936 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19937 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19938 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19939 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19940 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19941 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19942 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19943 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19944 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19945 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19946 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19947 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19948 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19949 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19950 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19951 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19952 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19953 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19954 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19955 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19956 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19957 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19958 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19959 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19960 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19961 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19962 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19963 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19964 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19965 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19966 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19967 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19968 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19969 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19970 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19971 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19972 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19973 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19974 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19975 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19976 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19977 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19978 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19979 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19980 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19981 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19982 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19983 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19984 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19985 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19986 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19987 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19988 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19989 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19990 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19991 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19992 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19993 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19994 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19995 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19996 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19997 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19998 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19999 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20000 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20001 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20002 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20003 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20004 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20005 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20006 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20007 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20008 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20009 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20010 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20011 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20012 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20013 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20014 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20015 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20016 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20017 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20018 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20019 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20020 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20021 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20022 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20023 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20024 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20025 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20026 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20027 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20028 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20029 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20030 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20031 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20032 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20033 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20034 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20035 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20036 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20037 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20038 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20039 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20040 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20041 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20042 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20043 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20044 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20045 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20046 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20047 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20048 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20049 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20050 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655