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 * The FIPS provider includes a few non-approved algorithms for
241 backward compatibility purposes and the "fips=yes" property query
242 must be used for all algorithm fetches to ensure FIPS compliance.
244 The algorithms that are included but not approved are Triple DES ECB,
245 Triple DES CBC and EdDSA.
249 * Added support for KMAC in KBKDF.
253 * RNDR and RNDRRS support in provider functions to provide
254 random number generation for Arm CPUs (aarch64).
258 * s_client and s_server apps now explicitly say when the TLS version
259 does not include the renegotiation mechanism. This avoids confusion
260 between that scenario versus when the TLS version includes secure
261 renegotiation but the peer lacks support for it.
265 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
267 *Tomasz Kantecki, Andrey Matyukov*
269 * The various OBJ_* functions have been made thread safe.
273 * Parallel dual-prime 1536/2048-bit modular exponentiation for
274 AVX512_IFMA capable processors.
276 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
278 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
279 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
280 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
281 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
282 `OPENSSL_NO_DEPRECATED_3_1`.
284 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
285 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
286 definitions for these functions regardless of whether
287 `OPENSSL_NO_DEPRECATED_3_1` is defined.
289 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
290 functions regardless of whether they are using them. It is recommended that
291 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
295 * When generating safe-prime DH parameters set the recommended private key
296 length equivalent to minimum key lengths as in RFC 7919.
300 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
301 maximum size that is smaller or equal to the digest length to comply with
302 FIPS 186-4 section 5. This is implemented by a new option
303 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
304 `rsa_pss_saltlen` parameter, which is now the default. Signature
305 verification is not affected by this change and continues to work as before.
312 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
313 listed here are only a brief description.
314 The migration guide contains more detailed information related to new features,
315 breaking changes, and mappings for the large list of deprecated functions.
317 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
319 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
321 * Fixed NULL dereference during PKCS7 data verification.
323 A NULL pointer can be dereferenced when signatures are being
324 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
325 algorithm used for the signature is known to the OpenSSL library but
326 the implementation of the hash algorithm is not available the digest
327 initialization will fail. There is a missing check for the return
328 value from the initialization function which later leads to invalid
329 usage of the digest API most likely leading to a crash.
332 PKCS7 data is processed by the SMIME library calls and also by the
333 time stamp (TS) library calls. The TLS implementation in OpenSSL does
334 not call these functions however third party applications would be
335 affected if they call these functions to verify signatures on untrusted
340 * Fixed X.400 address type confusion in X.509 GeneralName.
342 There is a type confusion vulnerability relating to X.400 address processing
343 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
344 but the public structure definition for GENERAL_NAME incorrectly specified
345 the type of the x400Address field as ASN1_TYPE. This field is subsequently
346 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
349 When CRL checking is enabled (i.e. the application sets the
350 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
351 pass arbitrary pointers to a memcmp call, enabling them to read memory
352 contents or enact a denial of service.
357 * Fixed NULL dereference validating DSA public key.
359 An invalid pointer dereference on read can be triggered when an
360 application tries to check a malformed DSA public key by the
361 EVP_PKEY_public_check() function. This will most likely lead
362 to an application crash. This function can be called on public
363 keys supplied from untrusted sources which could allow an attacker
364 to cause a denial of service attack.
366 The TLS implementation in OpenSSL does not call this function
367 but applications might call the function if there are additional
368 security requirements imposed by standards such as FIPS 140-3.
371 *Shane Lontis, Tomáš Mráz*
373 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
375 An invalid pointer dereference on read can be triggered when an
376 application tries to load malformed PKCS7 data with the
377 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
379 The result of the dereference is an application crash which could
380 lead to a denial of service attack. The TLS implementation in OpenSSL
381 does not call this function however third party applications might
382 call these functions on untrusted data.
387 * Fixed Use-after-free following BIO_new_NDEF.
389 The public API function BIO_new_NDEF is a helper function used for
390 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
391 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
392 be called directly by end user applications.
394 The function receives a BIO from the caller, prepends a new BIO_f_asn1
395 filter BIO onto the front of it to form a BIO chain, and then returns
396 the new head of the BIO chain to the caller. Under certain conditions,
397 for example if a CMS recipient public key is invalid, the new filter BIO
398 is freed and the function returns a NULL result indicating a failure.
399 However, in this case, the BIO chain is not properly cleaned up and the
400 BIO passed by the caller still retains internal pointers to the previously
401 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
402 then a use-after-free will occur. This will most likely result in a crash.
405 *Viktor Dukhovni, Matt Caswell*
407 * Fixed Double free after calling PEM_read_bio_ex.
409 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
410 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
411 data. If the function succeeds then the "name_out", "header" and "data"
412 arguments are populated with pointers to buffers containing the relevant
413 decoded data. The caller is responsible for freeing those buffers. It is
414 possible to construct a PEM file that results in 0 bytes of payload data.
415 In this case PEM_read_bio_ex() will return a failure code but will populate
416 the header argument with a pointer to a buffer that has already been freed.
417 If the caller also frees this buffer then a double free will occur. This
418 will most likely lead to a crash.
420 The functions PEM_read_bio() and PEM_read() are simple wrappers around
421 PEM_read_bio_ex() and therefore these functions are also directly affected.
423 These functions are also called indirectly by a number of other OpenSSL
424 functions including PEM_X509_INFO_read_bio_ex() and
425 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
426 internal uses of these functions are not vulnerable because the caller does
427 not free the header argument if PEM_read_bio_ex() returns a failure code.
430 *Kurt Roeckx, Matt Caswell*
432 * Fixed Timing Oracle in RSA Decryption.
434 A timing based side channel exists in the OpenSSL RSA Decryption
435 implementation which could be sufficient to recover a plaintext across
436 a network in a Bleichenbacher style attack. To achieve a successful
437 decryption an attacker would have to be able to send a very large number
438 of trial messages for decryption. The vulnerability affects all RSA padding
439 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
442 *Dmitry Belyavsky, Hubert Kario*
444 * Fixed X.509 Name Constraints Read Buffer Overflow.
446 A read buffer overrun can be triggered in X.509 certificate verification,
447 specifically in name constraint checking. The read buffer overrun might
448 result in a crash which could lead to a denial of service attack.
449 In a TLS client, this can be triggered by connecting to a malicious
450 server. In a TLS server, this can be triggered if the server requests
451 client authentication and a malicious client connects.
456 * Fixed X.509 Policy Constraints Double Locking security issue.
458 If an X.509 certificate contains a malformed policy constraint and
459 policy processing is enabled, then a write lock will be taken twice
460 recursively. On some operating systems (most widely: Windows) this
461 results in a denial of service when the affected process hangs. Policy
462 processing being enabled on a publicly facing server is not considered
463 to be a common setup.
468 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
469 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
470 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
471 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
472 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
473 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
474 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
475 for legacy EC and SM2 keys is also changed similarly to honor the
476 equivalent conversion format flag as specified in the underlying
477 `EC_KEY` object being exported to a provider, when this function is
478 called through `EVP_PKEY_export()`.
482 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
484 * Fixed two buffer overflows in punycode decoding functions.
486 A buffer overrun can be triggered in X.509 certificate verification,
487 specifically in name constraint checking. Note that this occurs after
488 certificate chain signature verification and requires either a CA to
489 have signed the malicious certificate or for the application to continue
490 certificate verification despite failure to construct a path to a trusted
493 In a TLS client, this can be triggered by connecting to a malicious
494 server. In a TLS server, this can be triggered if the server requests
495 client authentication and a malicious client connects.
497 An attacker can craft a malicious email address to overflow
498 an arbitrary number of bytes containing the `.` character (decimal 46)
499 on the stack. This buffer overflow could result in a crash (causing a
503 An attacker can craft a malicious email address to overflow four
504 attacker-controlled bytes on the stack. This buffer overflow could
505 result in a crash (causing a denial of service) or potentially remote code
506 execution depending on stack layout for any given platform/compiler.
511 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
512 parameters in OpenSSL code.
513 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
514 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
515 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
516 Using these invalid names may cause algorithms to use slower methods
517 that ignore the CRT parameters.
521 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
526 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
527 data to be signed before signing the certificate.
531 * Added RIPEMD160 to the default provider.
535 * Ensured that the key share group sent or accepted for the key exchange
536 is allowed for the protocol version.
540 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
542 * OpenSSL supports creating a custom cipher via the legacy
543 EVP_CIPHER_meth_new() function and associated function calls. This function
544 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
545 to use the new provider mechanism in order to implement custom ciphers.
547 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
548 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
549 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
550 and decryption initialisation functions). Instead of using the custom cipher
551 directly it incorrectly tries to fetch an equivalent cipher from the
552 available providers. An equivalent cipher is found based on the NID passed to
553 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
554 given cipher. However it is possible for an application to incorrectly pass
555 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
556 is used in this way the OpenSSL encryption/decryption initialisation function
557 will match the NULL cipher as being equivalent and will fetch this from the
558 available providers. This will succeed if the default provider has been
559 loaded (or if a third party provider has been loaded that offers this
560 cipher). Using the NULL cipher means that the plaintext is emitted as the
563 Applications are only affected by this issue if they call
564 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
565 encryption/decryption initialisation function. Applications that only use
566 SSL/TLS are not impacted by this issue.
571 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
576 * Fixed the linux-mips64 Configure target which was missing the
577 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
582 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
587 * Correctly handle a retransmitted ClientHello in DTLS
591 * Fixed detection of ktls support in cross-compile environment on Linux
595 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
600 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
601 report correct results in some cases
605 * Fix UWP builds by defining VirtualLock
609 * For known safe primes use the minimum key length according to RFC 7919.
610 Longer private key sizes unnecessarily raise the cycles needed to compute the
611 shared secret without any increase of the real security. This fixes a
612 regression from 1.1.1 where these shorter keys were generated for the known
617 * Added the loongarch64 target
621 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
622 only passed to the FIPS provider and not to the default or legacy provider.
626 * Fixed reported performance degradation on aarch64. Restored the
627 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
628 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
629 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
630 The new algorithm is still used for 32 bit targets.
634 * Added a missing header for memcmp that caused compilation failure on some
639 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
641 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
642 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
643 This issue makes the RSA implementation with 2048 bit private keys
644 incorrect on such machines and memory corruption will happen during
645 the computation. As a consequence of the memory corruption an attacker
646 may be able to trigger a remote code execution on the machine performing
649 SSL/TLS servers or other servers using 2048 bit RSA private keys running
650 on machines supporting AVX512IFMA instructions of the X86_64 architecture
651 are affected by this issue.
656 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
657 implementation would not encrypt the entirety of the data under some
658 circumstances. This could reveal sixteen bytes of data that was
659 preexisting in the memory that wasn't written. In the special case of
660 "in place" encryption, sixteen bytes of the plaintext would be revealed.
662 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
663 they are both unaffected.
666 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
668 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
670 * In addition to the c_rehash shell command injection identified in
671 CVE-2022-1292, further bugs where the c_rehash script does not
672 properly sanitise shell metacharacters to prevent command injection have been
675 When the CVE-2022-1292 was fixed it was not discovered that there
676 are other places in the script where the file names of certificates
677 being hashed were possibly passed to a command executed through the shell.
679 This script is distributed by some operating systems in a manner where
680 it is automatically executed. On such operating systems, an attacker
681 could execute arbitrary commands with the privileges of the script.
683 Use of the c_rehash script is considered obsolete and should be replaced
684 by the OpenSSL rehash command line tool.
687 *Daniel Fiala, Tomáš Mráz*
689 * Case insensitive string comparison no longer uses locales. It has instead
690 been directly implemented.
694 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
696 * Case insensitive string comparison is reimplemented via new locale-agnostic
697 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
698 comparison. The previous implementation had problems when the Turkish locale
703 * Fixed a bug in the c_rehash script which was not properly sanitising shell
704 metacharacters to prevent command injection. This script is distributed by
705 some operating systems in a manner where it is automatically executed. On
706 such operating systems, an attacker could execute arbitrary commands with the
707 privileges of the script.
709 Use of the c_rehash script is considered obsolete and should be replaced
710 by the OpenSSL rehash command line tool.
715 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
716 certificate on an OCSP response. The bug caused the function in the case
717 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
718 response (meaning a successful verification) even in the case where the
719 response signing certificate fails to verify.
721 It is anticipated that most users of `OCSP_basic_verify` will not use the
722 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
723 a negative value (indicating a fatal error) in the case of a certificate
724 verification failure. The normal expected return value in this case would be
727 This issue also impacts the command line OpenSSL "ocsp" application. When
728 verifying an ocsp response with the "-no_cert_checks" option the command line
729 application will report that the verification is successful even though it
730 has in fact failed. In this case the incorrect successful response will also
731 be accompanied by error messages showing the failure and contradicting the
732 apparently successful result.
737 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
738 AAD data as the MAC key. This made the MAC key trivially predictable.
740 An attacker could exploit this issue by performing a man-in-the-middle attack
741 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
742 that the modified data would still pass the MAC integrity check.
744 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
745 endpoint will always be rejected by the recipient and the connection will
746 fail at that point. Many application protocols require data to be sent from
747 the client to the server first. Therefore, in such a case, only an OpenSSL
748 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
750 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
751 sent in both directions. In this case both clients and servers could be
752 affected, regardless of the application protocol.
754 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
755 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
756 the handshake when using this ciphersuite.
758 The confidentiality of data is not impacted by this issue, i.e. an attacker
759 cannot decrypt data that has been encrypted using this ciphersuite - they can
762 In order for this attack to work both endpoints must legitimately negotiate
763 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
764 OpenSSL 3.0, and is not available within the default provider or the default
765 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
766 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
767 following must have occurred:
769 1) OpenSSL must have been compiled with the (non-default) compile time option
770 enable-weak-ssl-ciphers
772 2) OpenSSL must have had the legacy provider explicitly loaded (either
773 through application code or via configuration)
775 3) The ciphersuite must have been explicitly added to the ciphersuite list
777 4) The libssl security level must have been set to 0 (default is 1)
779 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
781 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
782 others that both endpoints have in common
787 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
788 occupied by the removed hash table entries.
790 This function is used when decoding certificates or keys. If a long lived
791 process periodically decodes certificates or keys its memory usage will
792 expand without bounds and the process might be terminated by the operating
793 system causing a denial of service. Also traversing the empty hash table
794 entries will take increasingly more time.
796 Typically such long lived processes might be TLS clients or TLS servers
797 configured to accept client certificate authentication.
800 *Hugo Landau, Aliaksei Levin*
802 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
803 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
804 statistics are no longer supported. For compatibility, these statistics are
805 still listed in the output but are now always reported as zero.
809 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
811 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
812 for non-prime moduli.
814 Internally this function is used when parsing certificates that contain
815 elliptic curve public keys in compressed form or explicit elliptic curve
816 parameters with a base point encoded in compressed form.
818 It is possible to trigger the infinite loop by crafting a certificate that
819 has invalid explicit curve parameters.
821 Since certificate parsing happens prior to verification of the certificate
822 signature, any process that parses an externally supplied certificate may thus
823 be subject to a denial of service attack. The infinite loop can also be
824 reached when parsing crafted private keys as they can contain explicit
825 elliptic curve parameters.
827 Thus vulnerable situations include:
829 - TLS clients consuming server certificates
830 - TLS servers consuming client certificates
831 - Hosting providers taking certificates or private keys from customers
832 - Certificate authorities parsing certification requests from subscribers
833 - Anything else which parses ASN.1 elliptic curve parameters
835 Also any other applications that use the BN_mod_sqrt() where the attacker
836 can control the parameter values are vulnerable to this DoS issue.
841 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
842 to the list of ciphersuites providing Perfect Forward Secrecy as
843 required by SECLEVEL >= 3.
845 *Dmitry Belyavskiy, Nicola Tuveri*
847 * Made the AES constant time code for no-asm configurations
848 optional due to the resulting 95% performance degradation.
849 The AES constant time code can be enabled, for no assembly
850 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
854 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
859 * The negative return value handling of the certificate verification callback
860 was reverted. The replacement is to set the verification retry state with
861 the SSL_set_retry_verify() function.
865 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
867 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
868 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
869 verify a certificate supplied by a server. That function may return a
870 negative return value to indicate an internal error (for example out of
871 memory). Such a negative return value is mishandled by OpenSSL and will cause
872 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
873 success and a subsequent call to SSL_get_error() to return the value
874 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
875 returned by OpenSSL if the application has previously called
876 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
877 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
878 totally unexpected and applications may not behave correctly as a result. The
879 exact behaviour will depend on the application but it could result in
880 crashes, infinite loops or other similar incorrect responses.
882 This issue is made more serious in combination with a separate bug in OpenSSL
883 3.0 that will cause X509_verify_cert() to indicate an internal error when
884 processing a certificate chain. This will occur where a certificate does not
885 include the Subject Alternative Name extension but where a Certificate
886 Authority has enforced name constraints. This issue can occur even with valid
892 * Corrected a few file name and file reference bugs in the build,
893 installation and setup scripts, which lead to installation verification
894 failures. Slightly enhanced the installation verification script.
898 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
903 * Fixed PVK encoder to properly query for the passphrase.
907 * Multiple fixes in the OSSL_HTTP API functions.
911 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
912 OSSL_PARAM_INTEGER data type and return error on negative numbers
913 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
914 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
918 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
922 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
926 * Multiple threading fixes.
930 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
934 * Allow fetching an operation from the provider that owns an unexportable key
935 as a fallback if that is still allowed by the property query.
939 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
941 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
946 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
947 S390X capability vector to zero. This simplifies testing of different code
948 paths on S390X architecture.
952 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
953 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
954 SP 800-38D". The communication will fail at this point.
958 * The EC_GROUP_clear_free() function is deprecated as there is nothing
959 confidential in EC_GROUP data.
963 * The byte order mark (BOM) character is ignored if encountered at the
964 beginning of a PEM-formatted file.
968 * Added CMS support for the Russian GOST algorithms.
972 * Due to move of the implementation of cryptographic operations
973 to the providers, validation of various operation parameters can
974 be postponed until the actual operation is executed where previously
975 it happened immediately when an operation parameter was set.
977 For example when setting an unsupported curve with
978 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
979 fail but later keygen operations with the EVP_PKEY_CTX will fail.
981 *OpenSSL team members and many third party contributors*
983 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
984 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
985 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
986 instead to retrieve these algorithms from a provider.
990 * On build targets where the multilib postfix is set in the build
991 configuration the libdir directory was changing based on whether
992 the lib directory with the multilib postfix exists on the system
993 or not. This unpredictable behavior was removed and eventual
994 multilib postfix is now always added to the default libdir. Use
995 `--libdir=lib` to override the libdir if adding the postfix is
1000 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1001 no longer interoperable with OpenSSL 1.1.1.
1005 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1006 function codes, this function can only cause problems for calling
1011 * Add a configurable flag to output date formats as ISO 8601. Does not
1012 change the default date format.
1016 * Version of MSVC earlier than 1300 could get link warnings, which could
1017 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1018 Support for this flag has been removed.
1022 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1023 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1024 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1025 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1026 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1030 * The signatures of the functions to get and set options on SSL and
1031 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1032 Some source code changes may be required.
1036 * The public definitions of conf_method_st and conf_st have been
1037 deprecated. They will be made opaque in a future release.
1039 *Rich Salz and Tomáš Mráz*
1041 * Client-initiated renegotiation is disabled by default. To allow it, use
1042 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1043 flag, or the "ClientRenegotiation" config parameter as appropriate.
1047 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1048 or modify relative pathname inclusion.
1052 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1053 validated. Please consult the README-FIPS and
1054 README-PROVIDERS files, as well as the migration guide.
1056 *OpenSSL team members and many third party contributors*
1058 * For the key types DH and DHX the allowed settable parameters are now different.
1062 * The openssl commands that read keys, certificates, and CRLs now
1063 automatically detect the PEM or DER format of the input files.
1065 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1067 * Added enhanced PKCS#12 APIs which accept a library context.
1071 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1075 * Added support for Kernel TLS (KTLS).
1077 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1079 * Support for RFC 5746 secure renegotiation is now required by default for
1080 SSL or TLS connections to succeed.
1084 * The signature of the `copy` functional parameter of the
1085 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1086 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1087 the signature of the `pub_decode` functional parameter of the
1088 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1089 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1093 * The error return values from some control calls (ctrl) have changed.
1097 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1101 * Many functions in the EVP_ namespace that are getters of values from
1102 implementations or contexts were renamed to include get or get0 in their
1103 names. Old names are provided as macro aliases for compatibility and
1108 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1109 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1110 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1115 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1118 * The output from the command line applications may have minor
1123 * The output from numerous "printing" may have minor changes.
1127 * Windows thread synchronization uses read/write primitives (SRWLock) when
1128 supported by the OS, otherwise CriticalSection continues to be used.
1132 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1133 work on read only BIO source/sinks that do not support these functions.
1134 This allows piping or redirection of a file BIO using stdin to be buffered
1135 into memory. This is used internally in OSSL_DECODER_from_bio().
1139 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1140 this function would return one of the values OSSL_STORE_INFO_NAME,
1141 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1142 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1143 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1144 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1145 using this function should be amended to handle the changed return value.
1149 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1150 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1151 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1152 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1153 contains more than one certificate identifier: This means that all
1154 certificates referenced there MUST be part of the validation chain.
1158 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1159 RC5, DESX and DES have been moved to the legacy provider.
1163 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1164 RIPEMD-160 have been moved to the legacy provider.
1168 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1173 * The deprecated functions EVP_PKEY_get0_RSA(),
1174 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1175 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1176 well as the similarly named "get1" functions behave differently in
1181 * A number of functions handling low-level keys or engines were deprecated
1182 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1183 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1184 EVP_PKEY_get0_siphash().
1188 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1189 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1190 will need to load the legacy crypto provider. This includes these PBE
1191 algorithms which use this KDF:
1192 - NID_pbeWithMD2AndDES_CBC
1193 - NID_pbeWithMD5AndDES_CBC
1194 - NID_pbeWithSHA1AndRC2_CBC
1195 - NID_pbeWithMD2AndRC2_CBC
1196 - NID_pbeWithMD5AndRC2_CBC
1197 - NID_pbeWithSHA1AndDES_CBC
1201 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1202 BIO_debug_callback() functions.
1206 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1207 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1211 * The RAND_METHOD APIs have been deprecated.
1215 * The SRP APIs have been deprecated.
1219 * Add a compile time option to prevent the caching of provider fetched
1220 algorithms. This is enabled by including the no-cached-fetch option
1221 at configuration time.
1225 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1226 count of PKCS12_DEFAULT_ITER.
1228 *Tomáš Mráz and Sahana Prasad*
1230 * The openssl speed command does not use low-level API calls anymore.
1234 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1237 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1239 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1243 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1244 providers may supply their own group implementations (using either the "key
1245 exchange" or the "key encapsulation" methods) which will automatically be
1246 detected and used by libssl.
1248 *Matt Caswell, Nicola Tuveri*
1250 * The undocumented function X509_certificate_type() has been deprecated;
1254 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1258 * Removed RSA padding mode for SSLv23 (which was only used for
1259 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1260 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1265 * Deprecated the obsolete X9.31 RSA key generation related functions.
1267 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1268 is not allowed to return a value > 1, this is no more taken as failure.
1270 *Viktor Dukhovni and David von Oheimb*
1272 * Deprecated the obsolete X9.31 RSA key generation related functions
1273 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1274 BN_X931_generate_prime_ex().
1278 * The default key generation method for the regular 2-prime RSA keys was
1279 changed to the FIPS 186-4 B.3.6 method.
1283 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1287 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1291 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1292 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1294 *Rich Salz, Richard Levitte, and David von Oheimb*
1296 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1300 * Deprecated `OCSP_parse_url()`.
1304 * Validation of SM2 keys has been separated from the validation of regular EC
1309 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1310 switches: a validation failure triggers an early exit, returning a failure
1311 exit status to the parent process.
1315 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1316 to ignore unknown ciphers.
1320 * The `-cipher-commands` and `-digest-commands` options
1321 of the command line utility `list` have been deprecated.
1322 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1326 * Added convenience functions for generating asymmetric key pairs:
1327 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1328 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1332 * All of the low-level EC_KEY functions have been deprecated.
1334 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1336 * Deprecated all the libcrypto and libssl error string loading
1341 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1342 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1347 * The `-crypt` option to the `passwd` command line tool has been removed.
1351 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1356 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1360 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1361 EVP_PKEY_get1_tls_encodedpoint().
1365 * The security callback, which can be customised by application code, supports
1366 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1367 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1371 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1372 interface. Their functionality remains unchanged.
1376 * Added new option for 'openssl list', '-providers', which will display the
1377 list of loaded providers, their names, version and status. It optionally
1378 displays their gettable parameters.
1382 * Removed EVP_PKEY_set_alias_type().
1386 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1387 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1391 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1392 parameters are still checked for type safety at compile time via helper
1397 * Remove the RAND_DRBG API
1399 *Paul Dale and Matthias St. Pierre*
1401 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1402 as well as actual hostnames.
1406 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1407 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1408 conversely, silently ignore DTLS protocol version bounds when configuring
1409 TLS-based contexts. The commands can be repeated to set bounds of both
1410 types. The same applies with the corresponding "min_protocol" and
1411 "max_protocol" command-line switches, in case some application uses both TLS
1414 SSL_CTX instances that are created for a fixed protocol version (e.g.
1415 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1416 attempts to apply bounds to these protocol versions would result in an
1417 error. Now only the "version-flexible" SSL_CTX instances are subject to
1418 limits in configuration files in command-line options.
1422 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1427 * Reworked the recorded ERR codes to make better space for system errors.
1428 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1429 given code is a system error (true) or an OpenSSL error (false).
1433 * Reworked the test perl framework to better allow parallel testing.
1435 *Nicola Tuveri and David von Oheimb*
1437 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1438 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1442 * 'Configure' has been changed to figure out the configuration target if
1443 none is given on the command line. Consequently, the 'config' script is
1444 now only a mere wrapper. All documentation is changed to only mention
1447 *Rich Salz and Richard Levitte*
1449 * Added a library context `OSSL_LIB_CTX` that applications as well as
1450 other libraries can use to form a separate context within which
1451 libcrypto operations are performed.
1455 * Added various `_ex` functions to the OpenSSL API that support using
1456 a non-default `OSSL_LIB_CTX`.
1460 * Handshake now fails if Extended Master Secret extension is dropped
1465 * Dropped interactive mode from the `openssl` program.
1469 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1471 *David von Oheimb and Shane Lontis*
1473 * Deprecated `EC_METHOD_get_field_type()`.
1477 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1478 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1479 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1483 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1487 * Add CAdES-BES signature verification support, mostly derived
1488 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1490 *Filipe Raimundo da Silva*
1492 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1496 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1497 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1501 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1505 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1506 EC_KEY_precompute_mult().
1510 * Deprecated EC_POINTs_mul().
1514 * Removed FIPS_mode() and FIPS_mode_set().
1518 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1522 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1523 EC_POINT_get_Jprojective_coordinates_GFp().
1527 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1528 arrays to be more easily constructed via a series of utility functions.
1529 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1530 the various push functions and finally convert to a passable OSSL_PARAM
1531 array using OSSL_PARAM_BLD_to_param().
1535 * The security strength of SHA1 and MD5 based signatures in TLS has been
1540 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1541 contain a provider side internal key.
1545 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1549 * Project text documents not yet having a proper file name extension
1550 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1551 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1555 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1556 have been converted to Markdown with the goal to produce documents
1557 which not only look pretty when viewed online in the browser, but
1558 remain well readable inside a plain text editor.
1560 To achieve this goal, a 'minimalistic' Markdown style has been applied
1561 which avoids formatting elements that interfere too much with the
1562 reading flow in the text file. For example, it
1564 * avoids [ATX headings][] and uses [setext headings][] instead
1565 (which works for `<h1>` and `<h2>` headings only).
1566 * avoids [inline links][] and uses [reference links][] instead.
1567 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1569 [ATX headings]: https://github.github.com/gfm/#atx-headings
1570 [setext headings]: https://github.github.com/gfm/#setext-headings
1571 [inline links]: https://github.github.com/gfm/#inline-link
1572 [reference links]: https://github.github.com/gfm/#reference-link
1573 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1574 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1576 *Matthias St. Pierre*
1578 * The test suite is changed to preserve results of each test recipe.
1579 A new directory test-runs/ with subdirectories named like the
1580 test recipes are created in the build tree for this purpose.
1584 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1585 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1586 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1588 *David von Oheimb, Martin Peylo*
1590 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1591 It supports arbitrary request and response content types, GET redirection,
1592 TLS, connections via HTTP(S) proxies, connections and exchange via
1593 user-defined BIOs (allowing implicit connections), persistent connections,
1594 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1595 The legacy OCSP-focused (and only partly documented) API
1596 is retained for backward compatibility, while most of it is deprecated.
1600 * Added `util/check-format.pl`, a tool for checking adherence to the
1601 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1602 The checks performed are incomplete and yield some false positives.
1603 Still the tool should be useful for detecting most typical glitches.
1607 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1608 If domain name resolution yields multiple IP addresses all of them are tried
1609 after `connect()` failures.
1613 * All of the low-level RSA functions have been deprecated.
1617 * X509 certificates signed using SHA1 are no longer allowed at security
1622 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1623 modified to use PKEY APIs. These commands are now in maintenance mode
1624 and no new features will be added to them.
1628 * The command line utility rsautl has been deprecated.
1632 * The command line utilities genrsa and rsa have been modified to use PKEY
1633 APIs. They now write PKCS#8 keys by default. These commands are now in
1634 maintenance mode and no new features will be added to them.
1638 * All of the low-level DH functions have been deprecated.
1640 *Paul Dale and Matt Caswell*
1642 * All of the low-level DSA functions have been deprecated.
1646 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1647 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1651 * Deprecated low-level ECDH and ECDSA functions.
1655 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1659 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1660 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1661 a new formulation to include all the things it can be used for,
1662 as well as words of caution.
1666 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1670 * All of the low-level HMAC functions have been deprecated.
1672 *Paul Dale and David von Oheimb*
1674 * Over two thousand fixes were made to the documentation, including:
1675 - Common options (such as -rand/-writerand, TLS version control, etc)
1676 were refactored and point to newly-enhanced descriptions in openssl.pod.
1677 - Added style conformance for all options (with help from Richard Levitte),
1678 documented all reported missing options, added a CI build to check
1679 that all options are documented and that no unimplemented options
1681 - Documented some internals, such as all use of environment variables.
1682 - Addressed all internal broken L<> references.
1686 * All of the low-level CMAC functions have been deprecated.
1690 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1691 functions have been deprecated.
1693 *Paul Dale and David von Oheimb*
1695 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1696 set of functions. The documentation mentioned negative values for some
1697 errors, but this was never the case, so the mention of negative values
1700 Code that followed the documentation and thereby check with something
1701 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1705 * All of the low-level cipher functions have been deprecated.
1707 *Matt Caswell and Paul Dale*
1709 * Removed include/openssl/opensslconf.h.in and replaced it with
1710 include/openssl/configuration.h.in, which differs in not including
1711 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1712 was added to include both.
1714 This allows internal hacks where one might need to modify the set
1715 of configured macros, for example this if deprecated symbols are
1716 still supposed to be available internally:
1718 #include <openssl/configuration.h>
1720 #undef OPENSSL_NO_DEPRECATED
1721 #define OPENSSL_SUPPRESS_DEPRECATED
1723 #include <openssl/macros.h>
1725 This should not be used by applications that use the exported
1726 symbols, as that will lead to linking errors.
1730 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1731 used in exponentiation with 512-bit moduli. No EC algorithms are
1732 affected. Analysis suggests that attacks against 2-prime RSA1024,
1733 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1734 difficult to perform and are not believed likely. Attacks against DH512
1735 are considered just feasible. However, for an attack the target would
1736 have to re-use the DH512 private key, which is not recommended anyway.
1737 Also applications directly using the low-level API BN_mod_exp may be
1738 affected if they use BN_FLG_CONSTTIME.
1743 * Most memory-debug features have been deprecated, and the functionality
1744 replaced with no-ops.
1748 * Added documentation for the STACK API.
1752 * Introduced a new method type and API, OSSL_ENCODER, to represent
1753 generic encoders. These do the same sort of job that PEM writers
1754 and d2i functions do, but with support for methods supplied by
1755 providers, and the possibility for providers to support other
1760 * Introduced a new method type and API, OSSL_DECODER, to represent
1761 generic decoders. These do the same sort of job that PEM readers
1762 and i2d functions do, but with support for methods supplied by
1763 providers, and the possibility for providers to support other
1768 * Added a .pragma directive to the syntax of configuration files, to
1769 allow varying behavior in a supported and predictable manner.
1770 Currently added pragma:
1774 This allows dollar signs to be a keyword character unless it's
1775 followed by a opening brace or parenthesis. This is useful for
1776 platforms where dollar signs are commonly used in names, such as
1777 volume names and system directory names on VMS.
1781 * Added functionality to create an EVP_PKEY from user data.
1785 * Change the interpretation of the '--api' configuration option to
1786 mean that this is a desired API compatibility level with no
1787 further meaning. The previous interpretation, that this would
1788 also mean to remove all deprecated symbols up to and including
1789 the given version, no requires that 'no-deprecated' is also used
1790 in the configuration.
1792 When building applications, the desired API compatibility level
1793 can be set with the OPENSSL_API_COMPAT macro like before. For
1794 API compatibility version below 3.0, the old style numerical
1795 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1796 For version 3.0 and on, the value is expected to be the decimal
1797 value calculated from the major and minor version like this:
1799 MAJOR * 10000 + MINOR * 100
1803 -DOPENSSL_API_COMPAT=30000 For 3.0
1804 -DOPENSSL_API_COMPAT=30200 For 3.2
1806 To hide declarations that are deprecated up to and including the
1807 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1808 given when building the application as well.
1812 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1813 access to certificate and CRL stores via URIs and OSSL_STORE
1816 This adds the following functions:
1818 - X509_LOOKUP_store()
1819 - X509_STORE_load_file()
1820 - X509_STORE_load_path()
1821 - X509_STORE_load_store()
1822 - SSL_add_store_cert_subjects_to_stack()
1823 - SSL_CTX_set_default_verify_store()
1824 - SSL_CTX_load_verify_file()
1825 - SSL_CTX_load_verify_dir()
1826 - SSL_CTX_load_verify_store()
1830 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1831 The presence of this system service is determined at run-time.
1835 * Added functionality to create an EVP_PKEY context based on data
1836 for methods from providers. This takes an algorithm name and a
1837 property query string and simply stores them, with the intent
1838 that any operation that uses this context will use those strings
1839 to fetch the needed methods implicitly, thereby making the port
1840 of application written for pre-3.0 OpenSSL easier.
1844 * The undocumented function NCONF_WIN32() has been deprecated; for
1845 conversion details see the HISTORY section of doc/man5/config.pod
1849 * Introduced the new functions EVP_DigestSignInit_ex() and
1850 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1851 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1852 pages for further details.
1856 * Over two thousand fixes were made to the documentation, including:
1857 adding missing command flags, better style conformance, documentation
1860 *Rich Salz, Richard Levitte*
1862 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1863 X25519, X448, Ed25519 and Ed448.
1867 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1872 * Deprecated the public definition of `ERR_STATE` as well as the function
1873 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1878 * Added ERR functionality to give callers access to the stored function
1879 names that have replaced the older function code based functions.
1881 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1882 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1883 ERR_peek_error_all() and ERR_peek_last_error_all().
1885 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1886 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1887 ERR_func_error_string().
1891 * Extended testing to be verbose for failing tests only. The make variables
1892 VERBOSE_FAILURE or VF can be used to enable this:
1894 $ make VF=1 test # Unix
1895 $ mms /macro=(VF=1) test ! OpenVMS
1896 $ nmake VF=1 test # Windows
1900 * Added the `-copy_extensions` option to the `x509` command for use with
1901 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1902 all extensions in the request are copied to the certificate or vice versa.
1904 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1906 * Added the `-copy_extensions` option to the `req` command for use with
1907 `-x509`. When given with the `copy` or `copyall` argument,
1908 all extensions in the certification request are copied to the certificate.
1912 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1913 they generate are by default RFC 5280 compliant in the following sense:
1914 There is a subjectKeyIdentifier extension with a hash value of the public key
1915 and for not self-signed certs there is an authorityKeyIdentifier extension
1916 with a keyIdentifier field or issuer information identifying the signing key.
1917 This is done unless some configuration overrides the new default behavior,
1918 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1922 * Added several checks to `X509_verify_cert()` according to requirements in
1923 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1924 (which may be done by using the CLI option `-x509_strict`):
1925 * The basicConstraints of CA certificates must be marked critical.
1926 * CA certificates must explicitly include the keyUsage extension.
1927 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1928 * The issuer name of any certificate must not be empty.
1929 * The subject name of CA certs, certs with keyUsage crlSign,
1930 and certs without subjectAlternativeName must not be empty.
1931 * If a subjectAlternativeName extension is given it must not be empty.
1932 * The signatureAlgorithm field and the cert signature must be consistent.
1933 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1934 must not be marked critical.
1935 * The authorityKeyIdentifier must be given for X.509v3 certs
1936 unless they are self-signed.
1937 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1941 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1942 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1946 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1947 used even when parsing explicit parameters, when loading a encoded key
1948 or calling `EC_GROUP_new_from_ecpkparameters()`/
1949 `EC_GROUP_new_from_ecparameters()`.
1950 This prevents bypass of security hardening and performance gains,
1951 especially for curves with specialized EC_METHODs.
1952 By default, if a key encoded with explicit parameters is loaded and later
1953 encoded, the output is still encoded with explicit parameters, even if
1954 internally a "named" EC_GROUP is used for computation.
1958 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1959 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1960 NULL. After this change, only the cofactor parameter can be NULL. It also
1961 does some minimal sanity checks on the passed order.
1966 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1967 An attack is simple, if the first CMS_recipientInfo is valid but the
1968 second CMS_recipientInfo is chosen ciphertext. If the second
1969 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1970 encryption key will be replaced by garbage, and the message cannot be
1971 decoded, but if the RSA decryption fails, the correct encryption key is
1972 used and the recipient will not notice the attack.
1973 As a work around for this potential attack the length of the decrypted
1974 key must be equal to the cipher default key length, in case the
1975 certificate is not given and all recipientInfo are tried out.
1976 The old behaviour can be re-enabled in the CMS code by setting the
1977 CMS_DEBUG_DECRYPT flag.
1981 * Early start up entropy quality from the DEVRANDOM seed source has been
1982 improved for older Linux systems. The RAND subsystem will wait for
1983 /dev/random to be producing output before seeding from /dev/urandom.
1984 The seeded state is stored for future library initialisations using
1985 a system global shared memory segment. The shared memory identifier
1986 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1987 the desired value. The default identifier is 114.
1991 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1992 when primes for RSA keys are computed.
1993 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1994 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1995 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
1996 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1997 This avoids possible fingerprinting of newly generated RSA modules.
2001 * Correct the extended master secret constant on EBCDIC systems. Without this
2002 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2003 negotiate EMS will fail. Unfortunately this also means that TLS connections
2004 between EBCDIC systems with this fix, and EBCDIC systems without this
2005 fix will fail if they negotiate EMS.
2009 * Changed the library initialisation so that the config file is now loaded
2010 by default. This was already the case for libssl. It now occurs for both
2011 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2012 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2016 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2017 where the former acts as a replacement for `ERR_put_error()`, and the
2018 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2019 `ERR_raise_data()` adds more flexibility by taking a format string and
2020 an arbitrary number of arguments following it, to be processed with
2025 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2026 to check if a named provider is loaded and available. When called, it
2027 will also activate all fallback providers if such are still present.
2031 * Enforce a minimum DH modulus size of 512 bits.
2035 * Changed DH parameters to generate the order q subgroup instead of 2q.
2036 Previously generated DH parameters are still accepted by DH_check
2037 but DH_generate_key works around that by clearing bit 0 of the
2038 private key for those. This avoids leaking bit 0 of the private key.
2042 * Significantly reduce secure memory usage by the randomness pools.
2046 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2051 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2052 algorithms. An implementation of a key exchange algorithm can be obtained
2053 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2054 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2055 the older EVP_PKEY_derive_init() function. See the man pages for the new
2056 functions for further details.
2060 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2064 * Removed the function names from error messages and deprecated the
2069 * Removed NextStep support and the macro OPENSSL_UNISTD
2073 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2074 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2075 Also removed "export var as function" capability; we do not export
2076 variables, only functions.
2080 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2081 an error and 1 indicating success. In previous versions of OpenSSL this
2082 was a void type. If a key was set longer than the maximum possible this
2087 * Support SM2 signing and verification schemes with X509 certificate.
2091 * Use SHA256 as the default digest for TS query in the `ts` app.
2095 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2099 * Default cipher lists/suites are now available via a function, the
2100 #defines are deprecated.
2104 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2105 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2106 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2110 * Join the directories crypto/x509 and crypto/x509v3
2114 * Added command 'openssl kdf' that uses the EVP_KDF API.
2118 * Added command 'openssl mac' that uses the EVP_MAC API.
2122 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2123 as default directories. Also added the command 'openssl info'
2124 for scripting purposes.
2128 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2133 * Add prediction resistance to the DRBG reseeding process.
2137 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2138 mandated by IEEE Std 1619-2018.
2142 * Added newline escaping functionality to a filename when using openssl dgst.
2143 This output format is to replicate the output format found in the `*sum`
2144 checksum programs. This aims to preserve backward compatibility.
2146 *Matt Eaton, Richard Levitte, and Paul Dale*
2148 * Removed the heartbeat message in DTLS feature, as it has very
2149 little usage and doesn't seem to fulfill a valuable purpose.
2150 The configuration option is now deprecated.
2154 * Changed the output of 'openssl {digestname} < file' to display the
2155 digest name in its output.
2159 * Added a new generic trace API which provides support for enabling
2160 instrumentation through trace output.
2162 *Richard Levitte & Matthias St. Pierre*
2164 * Added build tests for C++. These are generated files that only do one
2165 thing, to include one public OpenSSL head file each. This tests that
2166 the public header files can be usefully included in a C++ application.
2168 This test isn't enabled by default. It can be enabled with the option
2169 'enable-buildtest-c++'.
2173 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2177 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2181 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2185 * Added KMAC to EVP_MAC.
2189 * Added property based algorithm implementation selection framework to
2194 * Added SCA hardening for modular field inversion in EC_GROUP through
2195 a new dedicated field_inv() pointer in EC_METHOD.
2196 This also addresses a leakage affecting conversions from projective
2197 to affine coordinates.
2199 *Billy Bob Brumley, Nicola Tuveri*
2201 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2202 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2203 those algorithms that were already supported through the EVP_PKEY API
2204 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2205 and scrypt are now wrappers that call EVP_KDF.
2209 * Build devcrypto engine as a dynamic engine.
2211 *Eneas U de Queiroz*
2213 * Add keyed BLAKE2 to EVP_MAC.
2217 * Fix a bug in the computation of the endpoint-pair shared secret used
2218 by DTLS over SCTP. This breaks interoperability with older versions
2219 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2220 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2221 interoperability with such broken implementations. However, enabling
2222 this switch breaks interoperability with correct implementations.
2224 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2225 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2229 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2233 * Changed the license to the Apache License v2.0.
2237 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2239 - Major releases (indicated by incrementing the MAJOR release number)
2240 may introduce incompatible API/ABI changes.
2241 - Minor releases (indicated by incrementing the MINOR release number)
2242 may introduce new features but retain API/ABI compatibility.
2243 - Patch releases (indicated by incrementing the PATCH number)
2244 are intended for bug fixes and other improvements of existing
2245 features only (like improving performance or adding documentation)
2246 and retain API/ABI compatibility.
2250 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2254 * Remove the 'dist' target and add a tarball building script. The
2255 'dist' target has fallen out of use, and it shouldn't be
2256 necessary to configure just to create a source distribution.
2260 * Recreate the OS390-Unix config target. It no longer relies on a
2261 special script like it did for OpenSSL pre-1.1.0.
2265 * Instead of having the source directories listed in Configure, add
2266 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2271 * Add GMAC to EVP_MAC.
2275 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2279 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2280 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2281 to facilitate the continued use of MACs through raw private keys in
2282 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2286 * Deprecate ECDH_KDF_X9_62().
2290 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2291 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2292 are retained for backwards compatibility.
2296 * AES-XTS mode now enforces that its two keys are different to mitigate
2297 the attacked described in "Efficient Instantiations of Tweakable
2298 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2299 Details of this attack can be obtained from:
2300 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2304 * Rename the object files, i.e. give them other names than in previous
2305 versions. Their names now include the name of the final product, as
2306 well as its type mnemonic (bin, lib, shlib).
2310 * Added new option for 'openssl list', '-objects', which will display the
2311 list of built in objects, i.e. OIDs with names.
2315 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2316 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2321 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2322 improves application performance by removing data copies and providing
2323 applications with zero-copy system calls such as sendfile and splice.
2327 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2331 * `PKCS12_parse` now maintains the order of the parsed certificates
2332 when outputting them via `*ca` (rather than reversing it).
2336 * Deprecated pthread fork support methods.
2340 * Added support for FFDHE key exchange in TLS 1.3.
2344 * Added a new concept for OpenSSL plugability: providers. This
2345 functionality is designed to replace the ENGINE API and ENGINE
2346 implementations, and to be much more dynamic, allowing provider
2347 authors to introduce new algorithms among other things, as long as
2348 there's an API that supports the algorithm type.
2350 With this concept comes a new core API for interaction between
2351 libcrypto and provider implementations. Public libcrypto functions
2352 that want to use providers do so through this core API.
2354 The main documentation for this core API is found in
2355 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2356 refer to other manuals describing the API specific for supported
2357 algorithm types (also called operations).
2364 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2366 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2368 * Avoid loading of a dynamic engine twice.
2372 * Prioritise DANE TLSA issuer certs over peer certs
2376 * Fixed random API for MacOS prior to 10.12
2378 These MacOS versions don't support the CommonCrypto APIs
2382 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2384 * Fixed an SM2 Decryption Buffer Overflow.
2386 In order to decrypt SM2 encrypted data an application is expected to
2387 call the API function EVP_PKEY_decrypt(). Typically an application will
2388 call this function twice. The first time, on entry, the "out" parameter
2389 can be NULL and, on exit, the "outlen" parameter is populated with the
2390 buffer size required to hold the decrypted plaintext. The application
2391 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2392 again, but this time passing a non-NULL value for the "out" parameter.
2394 A bug in the implementation of the SM2 decryption code means that the
2395 calculation of the buffer size required to hold the plaintext returned
2396 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2397 size required by the second call. This can lead to a buffer overflow
2398 when EVP_PKEY_decrypt() is called by the application a second time with
2399 a buffer that is too small.
2401 A malicious attacker who is able present SM2 content for decryption to
2402 an application could cause attacker chosen data to overflow the buffer
2403 by up to a maximum of 62 bytes altering the contents of other data held
2404 after the buffer, possibly changing application behaviour or causing
2405 the application to crash. The location of the buffer is application
2406 dependent but is typically heap allocated.
2411 * Fixed various read buffer overruns processing ASN.1 strings
2413 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2414 structure which contains a buffer holding the string data and a field
2415 holding the buffer length. This contrasts with normal C strings which
2416 are represented as a buffer for the string data which is terminated
2417 with a NUL (0) byte.
2419 Although not a strict requirement, ASN.1 strings that are parsed using
2420 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2421 well as any string whose value has been set with the ASN1_STRING_set()
2422 function will additionally NUL terminate the byte array in the
2423 ASN1_STRING structure.
2425 However, it is possible for applications to directly construct valid
2426 ASN1_STRING structures which do not NUL terminate the byte array by
2427 directly setting the "data" and "length" fields in the ASN1_STRING
2428 array. This can also happen by using the ASN1_STRING_set0() function.
2430 Numerous OpenSSL functions that print ASN.1 data have been found to
2431 assume that the ASN1_STRING byte array will be NUL terminated, even
2432 though this is not guaranteed for strings that have been directly
2433 constructed. Where an application requests an ASN.1 structure to be
2434 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2435 been directly constructed by the application without NUL terminating
2436 the "data" field, then a read buffer overrun can occur.
2438 The same thing can also occur during name constraints processing
2439 of certificates (for example if a certificate has been directly
2440 constructed by the application instead of loading it via the OpenSSL
2441 parsing functions, and the certificate contains non NUL terminated
2442 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2443 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2445 If a malicious actor can cause an application to directly construct an
2446 ASN1_STRING and then process it through one of the affected OpenSSL
2447 functions then this issue could be hit. This might result in a crash
2448 (causing a Denial of Service attack). It could also result in the
2449 disclosure of private memory contents (such as private keys, or
2450 sensitive plaintext).
2455 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2457 * Fixed a problem with verifying a certificate chain when using the
2458 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2459 the certificates present in a certificate chain. It is not set by default.
2461 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2462 the chain that have explicitly encoded elliptic curve parameters was added
2463 as an additional strict check.
2465 An error in the implementation of this check meant that the result of a
2466 previous check to confirm that certificates in the chain are valid CA
2467 certificates was overwritten. This effectively bypasses the check
2468 that non-CA certificates must not be able to issue other certificates.
2470 If a "purpose" has been configured then there is a subsequent opportunity
2471 for checks that the certificate is a valid CA. All of the named "purpose"
2472 values implemented in libcrypto perform this check. Therefore, where
2473 a purpose is set the certificate chain will still be rejected even when the
2474 strict flag has been used. A purpose is set by default in libssl client and
2475 server certificate verification routines, but it can be overridden or
2476 removed by an application.
2478 In order to be affected, an application must explicitly set the
2479 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2480 for the certificate verification or, in the case of TLS client or server
2481 applications, override the default purpose.
2486 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2487 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2488 renegotiation ClientHello omits the signature_algorithms extension (where it
2489 was present in the initial ClientHello), but includes a
2490 signature_algorithms_cert extension then a NULL pointer dereference will
2491 result, leading to a crash and a denial of service attack.
2493 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2494 (which is the default configuration). OpenSSL TLS clients are not impacted by
2498 *Peter Kästle and Samuel Sapalski*
2500 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2502 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2503 create a unique hash value based on the issuer and serial number data
2504 contained within an X509 certificate. However, it was failing to correctly
2505 handle any errors that may occur while parsing the issuer field (which might
2506 occur if the issuer field is maliciously constructed). This may subsequently
2507 result in a NULL pointer deref and a crash leading to a potential denial of
2513 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2514 padding mode to correctly check for rollback attacks. This is considered a
2515 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2520 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2521 functions. Previously they could overflow the output length argument in some
2522 cases where the input length is close to the maximum permissible length for
2523 an integer on the platform. In such cases the return value from the function
2524 call would be 1 (indicating success), but the output length value would be
2525 negative. This could cause applications to behave incorrectly or crash.
2530 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2531 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2532 could be exploited in a side channel attack to recover the password. Since
2533 the attack is local host only this is outside of the current OpenSSL
2534 threat model and therefore no CVE is assigned.
2536 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2541 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2543 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2544 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2545 If an attacker can control both items being compared then this could lead
2546 to a possible denial of service attack. OpenSSL itself uses the
2547 GENERAL_NAME_cmp function for two purposes:
2548 1) Comparing CRL distribution point names between an available CRL and a
2549 CRL distribution point embedded in an X509 certificate
2550 2) When verifying that a timestamp response token signer matches the
2551 timestamp authority name (exposed via the API functions
2552 TS_RESP_verify_response and TS_RESP_verify_token)
2557 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2559 * Certificates with explicit curve parameters are now disallowed in
2560 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2564 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2565 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2566 conversely, silently ignore DTLS protocol version bounds when configuring
2567 TLS-based contexts. The commands can be repeated to set bounds of both
2568 types. The same applies with the corresponding "min_protocol" and
2569 "max_protocol" command-line switches, in case some application uses both TLS
2572 SSL_CTX instances that are created for a fixed protocol version (e.g.
2573 TLSv1_server_method()) also silently ignore version bounds. Previously
2574 attempts to apply bounds to these protocol versions would result in an
2575 error. Now only the "version-flexible" SSL_CTX instances are subject to
2576 limits in configuration files in command-line options.
2580 * Handshake now fails if Extended Master Secret extension is dropped
2585 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2587 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2589 * Fixed segmentation fault in SSL_check_chain()
2590 Server or client applications that call the SSL_check_chain() function
2591 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2592 dereference as a result of incorrect handling of the
2593 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2594 or unrecognised signature algorithm is received from the peer. This could
2595 be exploited by a malicious peer in a Denial of Service attack.
2600 * Added AES consttime code for no-asm configurations
2601 an optional constant time support for AES was added
2602 when building openssl for no-asm.
2603 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2604 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2605 At this time this feature is by default disabled.
2606 It will be enabled by default in 3.0.
2610 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2612 * Revert the change of EOF detection while reading in libssl to avoid
2613 regressions in applications depending on the current way of reporting
2614 the EOF. As the existing method is not fully accurate the change to
2615 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2616 branch and will be present in the 3.0 release.
2620 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2621 when primes for RSA keys are computed.
2622 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2623 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2624 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2625 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2626 This avoids possible fingerprinting of newly generated RSA modules.
2630 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2632 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2633 while reading in libssl then we would report an error back to the
2634 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2635 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2636 therefore give a hint as to what went wrong.
2640 * Check that ed25519 and ed448 are allowed by the security level. Previously
2641 signature algorithms not using an MD were not being checked that they were
2642 allowed by the security level.
2646 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2647 was not quite right. The behaviour was not consistent between resumption
2648 and normal handshakes, and also not quite consistent with historical
2649 behaviour. The behaviour in various scenarios has been clarified and
2650 it has been updated to make it match historical behaviour as closely as
2655 * *[VMS only]* The header files that the VMS compilers include automatically,
2656 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2657 that the C++ compiler doesn't understand. This is a shortcoming in the
2658 compiler, but can be worked around with `__cplusplus` guards.
2660 C++ applications that use OpenSSL libraries must be compiled using the
2661 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2662 functions. Otherwise, only functions with symbols of less than 31
2663 characters can be used, as the linker will not be able to successfully
2664 resolve symbols with longer names.
2668 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2669 The presence of this system service is determined at run-time.
2673 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2678 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2680 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2681 number generator (RNG). This was intended to include protection in the
2682 event of a fork() system call in order to ensure that the parent and child
2683 processes did not share the same RNG state. However, this protection was not
2684 being used in the default case.
2686 A partial mitigation for this issue is that the output from a high
2687 precision timer is mixed into the RNG state so the likelihood of a parent
2688 and child process sharing state is significantly reduced.
2690 If an application already calls OPENSSL_init_crypto() explicitly using
2691 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2694 *Matthias St. Pierre*
2696 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2697 used even when parsing explicit parameters, when loading a encoded key
2698 or calling `EC_GROUP_new_from_ecpkparameters()`/
2699 `EC_GROUP_new_from_ecparameters()`.
2700 This prevents bypass of security hardening and performance gains,
2701 especially for curves with specialized EC_METHODs.
2702 By default, if a key encoded with explicit parameters is loaded and later
2703 encoded, the output is still encoded with explicit parameters, even if
2704 internally a "named" EC_GROUP is used for computation.
2708 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2709 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2710 NULL. After this change, only the cofactor parameter can be NULL. It also
2711 does some minimal sanity checks on the passed order.
2716 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2717 An attack is simple, if the first CMS_recipientInfo is valid but the
2718 second CMS_recipientInfo is chosen ciphertext. If the second
2719 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2720 encryption key will be replaced by garbage, and the message cannot be
2721 decoded, but if the RSA decryption fails, the correct encryption key is
2722 used and the recipient will not notice the attack.
2723 As a work around for this potential attack the length of the decrypted
2724 key must be equal to the cipher default key length, in case the
2725 certificate is not given and all recipientInfo are tried out.
2726 The old behaviour can be re-enabled in the CMS code by setting the
2727 CMS_DEBUG_DECRYPT flag.
2732 * Early start up entropy quality from the DEVRANDOM seed source has been
2733 improved for older Linux systems. The RAND subsystem will wait for
2734 /dev/random to be producing output before seeding from /dev/urandom.
2735 The seeded state is stored for future library initialisations using
2736 a system global shared memory segment. The shared memory identifier
2737 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2738 the desired value. The default identifier is 114.
2742 * Correct the extended master secret constant on EBCDIC systems. Without this
2743 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2744 negotiate EMS will fail. Unfortunately this also means that TLS connections
2745 between EBCDIC systems with this fix, and EBCDIC systems without this
2746 fix will fail if they negotiate EMS.
2750 * Use Windows installation paths in the mingw builds
2752 Mingw isn't a POSIX environment per se, which means that Windows
2753 paths should be used for installation.
2758 * Changed DH_check to accept parameters with order q and 2q subgroups.
2759 With order 2q subgroups the bit 0 of the private key is not secret
2760 but DH_generate_key works around that by clearing bit 0 of the
2761 private key for those. This avoids leaking bit 0 of the private key.
2765 * Significantly reduce secure memory usage by the randomness pools.
2769 * Revert the DEVRANDOM_WAIT feature for Linux systems
2771 The DEVRANDOM_WAIT feature added a select() call to wait for the
2772 /dev/random device to become readable before reading from the
2773 /dev/urandom device.
2775 It turned out that this change had negative side effects on
2776 performance which were not acceptable. After some discussion it
2777 was decided to revert this feature and leave it up to the OS
2778 resp. the platform maintainer to ensure a proper initialization
2779 during early boot time.
2781 *Matthias St. Pierre*
2783 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2785 * Add build tests for C++. These are generated files that only do one
2786 thing, to include one public OpenSSL head file each. This tests that
2787 the public header files can be usefully included in a C++ application.
2789 This test isn't enabled by default. It can be enabled with the option
2790 'enable-buildtest-c++'.
2794 * Enable SHA3 pre-hashing for ECDSA and DSA.
2798 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2799 This changes the size when using the `genpkey` command when no size is given.
2800 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2801 generation commands to use 2048 bits by default.
2805 * Reorganize the manual pages to consistently have RETURN VALUES,
2806 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2807 util/fix-doc-nits accordingly.
2809 *Paul Yang, Joshua Lock*
2811 * Add the missing accessor EVP_PKEY_get0_engine()
2815 * Have commands like `s_client` and `s_server` output the signature scheme
2816 along with other cipher suite parameters when debugging.
2820 * Make OPENSSL_config() error agnostic again.
2824 * Do the error handling in RSA decryption constant time.
2828 * Prevent over long nonces in ChaCha20-Poly1305.
2830 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2831 for every encryption operation. RFC 7539 specifies that the nonce value
2832 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2833 and front pads the nonce with 0 bytes if it is less than 12
2834 bytes. However it also incorrectly allows a nonce to be set of up to 16
2835 bytes. In this case only the last 12 bytes are significant and any
2836 additional leading bytes are ignored.
2838 It is a requirement of using this cipher that nonce values are
2839 unique. Messages encrypted using a reused nonce value are susceptible to
2840 serious confidentiality and integrity attacks. If an application changes
2841 the default nonce length to be longer than 12 bytes and then makes a
2842 change to the leading bytes of the nonce expecting the new value to be a
2843 new unique nonce then such an application could inadvertently encrypt
2844 messages with a reused nonce.
2846 Additionally the ignored bytes in a long nonce are not covered by the
2847 integrity guarantee of this cipher. Any application that relies on the
2848 integrity of these ignored leading bytes of a long nonce may be further
2849 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2850 is safe because no such use sets such a long nonce value. However user
2851 applications that use this cipher directly and set a non-default nonce
2852 length to be longer than 12 bytes may be vulnerable.
2854 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2860 * Add DEVRANDOM_WAIT feature for Linux systems
2862 On older Linux systems where the getrandom() system call is not available,
2863 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2864 Contrary to getrandom(), the /dev/urandom device will not block during
2865 early boot when the kernel CSPRNG has not been seeded yet.
2867 To mitigate this known weakness, use select() to wait for /dev/random to
2868 become readable before reading from /dev/urandom.
2870 * Ensure that SM2 only uses SM3 as digest algorithm
2874 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2876 * Change the info callback signals for the start and end of a post-handshake
2877 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2878 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2879 confused by this and assume that a TLSv1.2 renegotiation has started. This
2880 can break KeyUpdate handling. Instead we no longer signal the start and end
2881 of a post handshake message exchange (although the messages themselves are
2882 still signalled). This could break some applications that were expecting
2883 the old signals. However without this KeyUpdate is not usable for many
2888 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2890 * Timing vulnerability in DSA signature generation
2892 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2893 timing side channel attack. An attacker could use variations in the signing
2894 algorithm to recover the private key.
2896 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2901 * Timing vulnerability in ECDSA signature generation
2903 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2904 timing side channel attack. An attacker could use variations in the signing
2905 algorithm to recover the private key.
2907 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2912 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2913 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2914 of two gigabytes and the error handling improved.
2916 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2917 categorized as a normal bug, not a security issue, because the DRBG reseeds
2918 automatically and is fully functional even without additional randomness
2919 provided by the application.
2921 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2923 * Add a new ClientHello callback. Provides a callback interface that gives
2924 the application the ability to adjust the nascent SSL object at the
2925 earliest stage of ClientHello processing, immediately after extensions have
2926 been collected but before they have been processed. In particular, this
2927 callback can adjust the supported TLS versions in response to the contents
2932 * Add SM2 base algorithm support.
2936 * s390x assembly pack: add (improved) hardware-support for the following
2937 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2938 aes-cfb/cfb8, aes-ecb.
2942 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2943 parameter is no longer accepted, as it leads to a corrupt table. NULL
2944 pem_str is reserved for alias entries only.
2948 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2949 step for prime curves. The new implementation is based on formulae from
2950 differential addition-and-doubling in homogeneous projective coordinates
2951 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2952 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2953 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2954 to work in projective coordinates.
2956 *Billy Bob Brumley, Nicola Tuveri*
2958 * Change generating and checking of primes so that the error rate of not
2959 being prime depends on the intended use based on the size of the input.
2960 For larger primes this will result in more rounds of Miller-Rabin.
2961 The maximal error rate for primes with more than 1080 bits is lowered
2964 *Kurt Roeckx, Annie Yousar*
2966 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2970 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2971 moving between systems, and to avoid confusion when a Windows build is
2972 done with mingw vs with MSVC. For POSIX installs, there's still a
2973 symlink or copy named 'tsget' to avoid that confusion as well.
2977 * Revert blinding in ECDSA sign and instead make problematic addition
2978 length-invariant. Switch even to fixed-length Montgomery multiplication.
2982 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2983 step for binary curves. The new implementation is based on formulae from
2984 differential addition-and-doubling in mixed Lopez-Dahab projective
2985 coordinates, modified to independently blind the operands.
2987 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2989 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2990 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2991 EC_METHODs to implement their own specialized "ladder step", to take
2992 advantage of more favorable coordinate systems or more efficient
2993 differential addition-and-doubling algorithms.
2995 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2997 * Modified the random device based seed sources to keep the relevant
2998 file descriptors open rather than reopening them on each access.
2999 This allows such sources to operate in a chroot() jail without
3000 the associated device nodes being available. This behaviour can be
3001 controlled using RAND_keep_random_devices_open().
3005 * Numerous side-channel attack mitigations have been applied. This may have
3006 performance impacts for some algorithms for the benefit of improved
3007 security. Specific changes are noted in this change log by their respective
3012 * AIX shared library support overhaul. Switch to AIX "natural" way of
3013 handling shared libraries, which means collecting shared objects of
3014 different versions and bitnesses in one common archive. This allows to
3015 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3016 doesn't affect the way 3rd party applications are linked, only how
3017 multi-version installation is managed.
3021 * Make ec_group_do_inverse_ord() more robust and available to other
3022 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3023 mitigations are applied to the fallback BN_mod_inverse().
3024 When using this function rather than BN_mod_inverse() directly, new
3025 EC cryptosystem implementations are then safer-by-default.
3029 * Add coordinate blinding for EC_POINT and implement projective
3030 coordinate blinding for generic prime curves as a countermeasure to
3031 chosen point SCA attacks.
3033 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3035 * Add blinding to ECDSA and DSA signatures to protect against side channel
3036 attacks discovered by Keegan Ryan (NCC Group).
3040 * Enforce checking in the `pkeyutl` command to ensure that the input
3041 length does not exceed the maximum supported digest length when performing
3042 a sign, verify or verifyrecover operation.
3046 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3047 I/O in combination with something like select() or poll() will hang. This
3048 can be turned off again using SSL_CTX_clear_mode().
3049 Many applications do not properly handle non-application data records, and
3050 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3051 around the problems in those applications, but can also break some.
3052 It's recommended to read the manpages about SSL_read(), SSL_write(),
3053 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3054 SSL_CTX_set_read_ahead() again.
3058 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3059 now allow empty (zero character) pass phrases.
3063 * Apply blinding to binary field modular inversion and remove patent
3064 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3068 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3069 binary and prime elliptic curves.
3073 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3074 constant time fixed point multiplication.
3078 * Revise elliptic curve scalar multiplication with timing attack
3079 defenses: ec_wNAF_mul redirects to a constant time implementation
3080 when computing fixed point and variable point multiplication (which
3081 in OpenSSL are mostly used with secret scalars in keygen, sign,
3082 ECDH derive operations).
3083 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3086 * Updated CONTRIBUTING
3090 * Updated DRBG / RAND to request nonce and additional low entropy
3091 randomness from the system.
3093 *Matthias St. Pierre*
3095 * Updated 'openssl rehash' to use OpenSSL consistent default.
3099 * Moved the load of the ssl_conf module to libcrypto, which helps
3100 loading engines that libssl uses before libssl is initialised.
3104 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3108 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3110 *Ingo Schwarze, Rich Salz*
3112 * Added output of accepting IP address and port for 'openssl s_server'
3116 * Added a new API for TLSv1.3 ciphersuites:
3117 SSL_CTX_set_ciphersuites()
3118 SSL_set_ciphersuites()
3122 * Memory allocation failures consistently add an error to the error
3127 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3128 in libcrypto when run as setuid/setgid.
3132 * Load any config file by default when libssl is used.
3136 * Added new public header file <openssl/rand_drbg.h> and documentation
3137 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3139 *Matthias St. Pierre*
3141 * QNX support removed (cannot find contributors to get their approval
3142 for the license change).
3146 * TLSv1.3 replay protection for early data has been implemented. See the
3147 SSL_read_early_data() man page for further details.
3151 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3152 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3153 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3154 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3155 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3156 configuration has been separated out. See the ciphers man page or the
3157 SSL_CTX_set_ciphersuites() man page for more information.
3161 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3162 in responder mode now supports the new "-multi" option, which
3163 spawns the specified number of child processes to handle OCSP
3164 requests. The "-timeout" option now also limits the OCSP
3165 responder's patience to wait to receive the full client request
3166 on a newly accepted connection. Child processes are respawned
3167 as needed, and the CA index file is automatically reloaded
3168 when changed. This makes it possible to run the "ocsp" responder
3169 as a long-running service, making the OpenSSL CA somewhat more
3170 feature-complete. In this mode, most diagnostic messages logged
3171 after entering the event loop are logged via syslog(3) rather than
3176 * Added support for X448 and Ed448. Heavily based on original work by
3181 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3182 objects loaded. This adds the functions OSSL_STORE_expect() and
3183 OSSL_STORE_find() as well as needed tools to construct searches and
3184 get the search data out of them.
3188 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3189 version of OpenSSL should review their configuration settings to ensure
3190 that they are still appropriate for TLSv1.3. For further information see:
3191 <https://wiki.openssl.org/index.php/TLS1.3>
3195 * Grand redesign of the OpenSSL random generator
3197 The default RAND method now utilizes an AES-CTR DRBG according to
3198 NIST standard SP 800-90Ar1. The new random generator is essentially
3199 a port of the default random generator from the OpenSSL FIPS 2.0
3200 object module. It is a hybrid deterministic random bit generator
3201 using an AES-CTR bit stream and which seeds and reseeds itself
3202 automatically using trusted system entropy sources.
3204 Some of its new features are:
3205 - Support for multiple DRBG instances with seed chaining.
3206 - The default RAND method makes use of a DRBG.
3207 - There is a public and private DRBG instance.
3208 - The DRBG instances are fork-safe.
3209 - Keep all global DRBG instances on the secure heap if it is enabled.
3210 - The public and private DRBG instance are per thread for lock free
3213 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3215 * Changed Configure so it only says what it does and doesn't dump
3216 so much data. Instead, ./configdata.pm should be used as a script
3217 to display all sorts of configuration data.
3221 * Added processing of "make variables" to Configure.
3225 * Added SHA512/224 and SHA512/256 algorithm support.
3229 * The last traces of Netware support, first removed in 1.1.0, have
3234 * Get rid of Makefile.shared, and in the process, make the processing
3235 of certain files (rc.obj, or the .def/.map/.opt files produced from
3236 the ordinal files) more visible and hopefully easier to trace and
3237 debug (or make silent).
3241 * Make it possible to have environment variable assignments as
3242 arguments to config / Configure.
3246 * Add multi-prime RSA (RFC 8017) support.
3250 * Add SM3 implemented according to GB/T 32905-2016
3251 *Jack Lloyd <jack.lloyd@ribose.com>,*
3252 *Ronald Tse <ronald.tse@ribose.com>,*
3253 *Erick Borsboom <erick.borsboom@ribose.com>*
3255 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3256 as documented in RFC6066.
3257 Based on a patch from Tomasz Moń
3259 *Filipe Raimundo da Silva*
3261 * Add SM4 implemented according to GB/T 32907-2016.
3262 *Jack Lloyd <jack.lloyd@ribose.com>,*
3263 *Ronald Tse <ronald.tse@ribose.com>,*
3264 *Erick Borsboom <erick.borsboom@ribose.com>*
3266 * Reimplement -newreq-nodes and ERR_error_string_n; the
3267 original author does not agree with the license change.
3271 * Add ARIA AEAD TLS support.
3275 * Some macro definitions to support VS6 have been removed. Visual
3276 Studio 6 has not worked since 1.1.0
3280 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3281 without clearing the errors.
3285 * Add "atfork" functions. If building on a system that without
3286 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3287 requirements. The RAND facility now uses/requires this.
3295 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3296 not possible to disable entirely. However, it's still possible to
3297 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3300 To disable, configure with 'no-ui-console'. 'no-ui' is still
3301 possible to use as an alias. Check at compile time with the
3302 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3303 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3307 * Add a STORE module, which implements a uniform and URI based reader of
3308 stores that can contain keys, certificates, CRLs and numerous other
3309 objects. The main API is loosely based on a few stdio functions,
3310 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3311 OSSL_STORE_error and OSSL_STORE_close.
3312 The implementation uses backends called "loaders" to implement arbitrary
3313 URI schemes. There is one built in "loader" for the 'file' scheme.
3317 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3318 then adjusted to work on FreeBSD 8.4 as well.
3319 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3320 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3324 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3325 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3326 error code calls like this:
3328 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3330 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3331 that can be encoded in C. For the foreseeable future, this will only
3334 *Richard Levitte and Tim Hudson*
3336 * Removed BSD cryptodev engine.
3340 * Add a build target 'build_all_generated', to build all generated files
3341 and only that. This can be used to prepare everything that requires
3342 things like perl for a system that lacks perl and then move everything
3343 to that system and do the rest of the build there.
3347 * In the UI interface, make it possible to duplicate the user data. This
3348 can be used by engines that need to retain the data for a longer time
3349 than just the call where this user data is passed.
3353 * Ignore the '-named_curve auto' value for compatibility of applications
3356 *Tomáš Mráz <tmraz@fedoraproject.org>*
3358 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3359 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3360 alerts across multiple records (some of which could be empty). In practice
3361 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3362 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3363 support this at all. Supporting it adds significant complexity to the
3364 record layer, and its removal is unlikely to cause interoperability
3369 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3370 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3371 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3376 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3377 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3379 *Richard Levitte, Andy Polyakov*
3381 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3386 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3387 platform rather than 'mingw'.
3391 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3392 success if they are asked to add an object which already exists
3393 in the store. This change cascades to other functions which load
3394 certificates and CRLs.
3398 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3399 facilitate stack unwinding even from assembly subroutines.
3403 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3404 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3408 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3409 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3410 which is the minimum version we support.
3414 * Certificate time validation (X509_cmp_time) enforces stricter
3415 compliance with RFC 5280. Fractional seconds and timezone offsets
3416 are no longer allowed.
3420 * Add support for ARIA
3424 * s_client will now send the Server Name Indication (SNI) extension by
3425 default unless the new "-noservername" option is used. The server name is
3426 based on the host provided to the "-connect" option unless overridden by
3427 using "-servername".
3431 * Add support for SipHash
3435 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3436 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3437 prevent issues where no progress is being made and the peer continually
3438 sends unrecognised record types, using up resources processing them.
3442 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3443 using the algorithm defined in
3444 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3448 * Heartbeat support has been removed; the ABI is changed for now.
3450 *Richard Levitte, Rich Salz*
3452 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3456 * The RSA "null" method, which was partially supported to avoid patent
3457 issues, has been replaced to always returns NULL.
3464 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3466 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3467 used even when parsing explicit parameters, when loading a encoded key
3468 or calling `EC_GROUP_new_from_ecpkparameters()`/
3469 `EC_GROUP_new_from_ecparameters()`.
3470 This prevents bypass of security hardening and performance gains,
3471 especially for curves with specialized EC_METHODs.
3472 By default, if a key encoded with explicit parameters is loaded and later
3473 encoded, the output is still encoded with explicit parameters, even if
3474 internally a "named" EC_GROUP is used for computation.
3478 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3479 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3480 NULL. After this change, only the cofactor parameter can be NULL. It also
3481 does some minimal sanity checks on the passed order.
3486 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3487 An attack is simple, if the first CMS_recipientInfo is valid but the
3488 second CMS_recipientInfo is chosen ciphertext. If the second
3489 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3490 encryption key will be replaced by garbage, and the message cannot be
3491 decoded, but if the RSA decryption fails, the correct encryption key is
3492 used and the recipient will not notice the attack.
3493 As a work around for this potential attack the length of the decrypted
3494 key must be equal to the cipher default key length, in case the
3495 certificate is not given and all recipientInfo are tried out.
3496 The old behaviour can be re-enabled in the CMS code by setting the
3497 CMS_DEBUG_DECRYPT flag.
3502 * Use Windows installation paths in the mingw builds
3504 Mingw isn't a POSIX environment per se, which means that Windows
3505 paths should be used for installation.
3510 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3512 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3513 This changes the size when using the `genpkey` command when no size is given.
3514 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3515 generation commands to use 2048 bits by default.
3519 * Prevent over long nonces in ChaCha20-Poly1305.
3521 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3522 for every encryption operation. RFC 7539 specifies that the nonce value
3523 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3524 and front pads the nonce with 0 bytes if it is less than 12
3525 bytes. However it also incorrectly allows a nonce to be set of up to 16
3526 bytes. In this case only the last 12 bytes are significant and any
3527 additional leading bytes are ignored.
3529 It is a requirement of using this cipher that nonce values are
3530 unique. Messages encrypted using a reused nonce value are susceptible to
3531 serious confidentiality and integrity attacks. If an application changes
3532 the default nonce length to be longer than 12 bytes and then makes a
3533 change to the leading bytes of the nonce expecting the new value to be a
3534 new unique nonce then such an application could inadvertently encrypt
3535 messages with a reused nonce.
3537 Additionally the ignored bytes in a long nonce are not covered by the
3538 integrity guarantee of this cipher. Any application that relies on the
3539 integrity of these ignored leading bytes of a long nonce may be further
3540 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3541 is safe because no such use sets such a long nonce value. However user
3542 applications that use this cipher directly and set a non-default nonce
3543 length to be longer than 12 bytes may be vulnerable.
3545 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3551 * Added SCA hardening for modular field inversion in EC_GROUP through
3552 a new dedicated field_inv() pointer in EC_METHOD.
3553 This also addresses a leakage affecting conversions from projective
3554 to affine coordinates.
3556 *Billy Bob Brumley, Nicola Tuveri*
3558 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3559 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3563 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3567 * Remove the 'dist' target and add a tarball building script. The
3568 'dist' target has fallen out of use, and it shouldn't be
3569 necessary to configure just to create a source distribution.
3573 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3575 * Timing vulnerability in DSA signature generation
3577 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3578 timing side channel attack. An attacker could use variations in the signing
3579 algorithm to recover the private key.
3581 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3586 * Timing vulnerability in ECDSA signature generation
3588 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3589 timing side channel attack. An attacker could use variations in the signing
3590 algorithm to recover the private key.
3592 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3597 * Add coordinate blinding for EC_POINT and implement projective
3598 coordinate blinding for generic prime curves as a countermeasure to
3599 chosen point SCA attacks.
3601 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3603 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3605 * Client DoS due to large DH parameter
3607 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3608 malicious server can send a very large prime value to the client. This will
3609 cause the client to spend an unreasonably long period of time generating a
3610 key for this prime resulting in a hang until the client has finished. This
3611 could be exploited in a Denial Of Service attack.
3613 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3618 * Cache timing vulnerability in RSA Key Generation
3620 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3621 a cache timing side channel attack. An attacker with sufficient access to
3622 mount cache timing attacks during the RSA key generation process could
3623 recover the private key.
3625 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3626 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3631 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3632 parameter is no longer accepted, as it leads to a corrupt table. NULL
3633 pem_str is reserved for alias entries only.
3637 * Revert blinding in ECDSA sign and instead make problematic addition
3638 length-invariant. Switch even to fixed-length Montgomery multiplication.
3642 * Change generating and checking of primes so that the error rate of not
3643 being prime depends on the intended use based on the size of the input.
3644 For larger primes this will result in more rounds of Miller-Rabin.
3645 The maximal error rate for primes with more than 1080 bits is lowered
3648 *Kurt Roeckx, Annie Yousar*
3650 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3654 * Add blinding to ECDSA and DSA signatures to protect against side channel
3655 attacks discovered by Keegan Ryan (NCC Group).
3659 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3660 now allow empty (zero character) pass phrases.
3664 * Certificate time validation (X509_cmp_time) enforces stricter
3665 compliance with RFC 5280. Fractional seconds and timezone offsets
3666 are no longer allowed.
3670 * Fixed a text canonicalisation bug in CMS
3672 Where a CMS detached signature is used with text content the text goes
3673 through a canonicalisation process first prior to signing or verifying a
3674 signature. This process strips trailing space at the end of lines, converts
3675 line terminators to CRLF and removes additional trailing line terminators
3676 at the end of a file. A bug in the canonicalisation process meant that
3677 some characters, such as form-feed, were incorrectly treated as whitespace
3678 and removed. This is contrary to the specification (RFC5485). This fix
3679 could mean that detached text data signed with an earlier version of
3680 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3681 signed with a fixed OpenSSL may fail to verify with an earlier version of
3682 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3683 and use the "-binary" flag (for the "cms" command line application) or set
3684 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3688 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3690 * Constructed ASN.1 types with a recursive definition could exceed the stack
3692 Constructed ASN.1 types with a recursive definition (such as can be found
3693 in PKCS7) could eventually exceed the stack given malicious input with
3694 excessive recursion. This could result in a Denial Of Service attack. There
3695 are no such structures used within SSL/TLS that come from untrusted sources
3696 so this is considered safe.
3698 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3704 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3706 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3707 effectively reduced to only comparing the least significant bit of each
3708 byte. This allows an attacker to forge messages that would be considered as
3709 authenticated in an amount of tries lower than that guaranteed by the
3710 security claims of the scheme. The module can only be compiled by the
3711 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3713 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3719 * Add a build target 'build_all_generated', to build all generated files
3720 and only that. This can be used to prepare everything that requires
3721 things like perl for a system that lacks perl and then move everything
3722 to that system and do the rest of the build there.
3726 * Backport SSL_OP_NO_RENGOTIATION
3728 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3729 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3730 changes this is no longer possible in 1.1.0. Therefore, the new
3731 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3732 1.1.0 to provide equivalent functionality.
3734 Note that if an application built against 1.1.0h headers (or above) is run
3735 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3736 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3740 * Removed the OS390-Unix config target. It relied on a script that doesn't
3745 * rsaz_1024_mul_avx2 overflow bug on x86_64
3747 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3748 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3749 Analysis suggests that attacks against RSA and DSA as a result of this
3750 defect would be very difficult to perform and are not believed likely.
3751 Attacks against DH1024 are considered just feasible, because most of the
3752 work necessary to deduce information about a private key may be performed
3753 offline. The amount of resources required for such an attack would be
3754 significant. However, for an attack on TLS to be meaningful, the server
3755 would have to share the DH1024 private key among multiple clients, which is
3756 no longer an option since CVE-2016-0701.
3758 This only affects processors that support the AVX2 but not ADX extensions
3759 like Intel Haswell (4th generation).
3761 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3762 was originally found via the OSS-Fuzz project.
3767 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3769 * bn_sqrx8x_internal carry bug on x86_64
3771 There is a carry propagating bug in the x86_64 Montgomery squaring
3772 procedure. No EC algorithms are affected. Analysis suggests that attacks
3773 against RSA and DSA as a result of this defect would be very difficult to
3774 perform and are not believed likely. Attacks against DH are considered just
3775 feasible (although very difficult) because most of the work necessary to
3776 deduce information about a private key may be performed offline. The amount
3777 of resources required for such an attack would be very significant and
3778 likely only accessible to a limited number of attackers. An attacker would
3779 additionally need online access to an unpatched system using the target
3780 private key in a scenario with persistent DH parameters and a private
3781 key that is shared between multiple clients.
3783 This only affects processors that support the BMI1, BMI2 and ADX extensions
3784 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3786 This issue was reported to OpenSSL by the OSS-Fuzz project.
3791 * Malformed X.509 IPAddressFamily could cause OOB read
3793 If an X.509 certificate has a malformed IPAddressFamily extension,
3794 OpenSSL could do a one-byte buffer overread. The most likely result
3795 would be an erroneous display of the certificate in text format.
3797 This issue was reported to OpenSSL by the OSS-Fuzz project.
3802 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3804 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3805 platform rather than 'mingw'.
3809 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3810 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3811 which is the minimum version we support.
3815 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3817 * Encrypt-Then-Mac renegotiation crash
3819 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3820 negotiated where it was not in the original handshake (or vice-versa) then
3821 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3822 and servers are affected.
3824 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3829 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3831 * Truncated packet could crash via OOB read
3833 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3834 cipher is being used, then a truncated packet can cause that host to
3835 perform an out-of-bounds read, usually resulting in a crash.
3837 This issue was reported to OpenSSL by Robert Święcki of Google.
3842 * Bad (EC)DHE parameters cause a client crash
3844 If a malicious server supplies bad parameters for a DHE or ECDHE key
3845 exchange then this can result in the client attempting to dereference a
3846 NULL pointer leading to a client crash. This could be exploited in a Denial
3849 This issue was reported to OpenSSL by Guido Vranken.
3854 * BN_mod_exp may produce incorrect results on x86_64
3856 There is a carry propagating bug in the x86_64 Montgomery squaring
3857 procedure. No EC algorithms are affected. Analysis suggests that attacks
3858 against RSA and DSA as a result of this defect would be very difficult to
3859 perform and are not believed likely. Attacks against DH are considered just
3860 feasible (although very difficult) because most of the work necessary to
3861 deduce information about a private key may be performed offline. The amount
3862 of resources required for such an attack would be very significant and
3863 likely only accessible to a limited number of attackers. An attacker would
3864 additionally need online access to an unpatched system using the target
3865 private key in a scenario with persistent DH parameters and a private
3866 key that is shared between multiple clients. For example this can occur by
3867 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3868 similar to CVE-2015-3193 but must be treated as a separate problem.
3870 This issue was reported to OpenSSL by the OSS-Fuzz project.
3875 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3877 * ChaCha20/Poly1305 heap-buffer-overflow
3879 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3880 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3881 crash. This issue is not considered to be exploitable beyond a DoS.
3883 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3888 * CMS Null dereference
3890 Applications parsing invalid CMS structures can crash with a NULL pointer
3891 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3892 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3893 structure callback if an attempt is made to free certain invalid encodings.
3894 Only CHOICE structures using a callback which do not handle NULL value are
3897 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3902 * Montgomery multiplication may produce incorrect results
3904 There is a carry propagating bug in the Broadwell-specific Montgomery
3905 multiplication procedure that handles input lengths divisible by, but
3906 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3907 and DH private keys are impossible. This is because the subroutine in
3908 question is not used in operations with the private key itself and an input
3909 of the attacker's direct choice. Otherwise the bug can manifest itself as
3910 transient authentication and key negotiation failures or reproducible
3911 erroneous outcome of public-key operations with specially crafted input.
3912 Among EC algorithms only Brainpool P-512 curves are affected and one
3913 presumably can attack ECDH key negotiation. Impact was not analyzed in
3914 detail, because pre-requisites for attack are considered unlikely. Namely
3915 multiple clients have to choose the curve in question and the server has to
3916 share the private key among them, neither of which is default behaviour.
3917 Even then only clients that chose the curve will be affected.
3919 This issue was publicly reported as transient failures and was not
3920 initially recognized as a security issue. Thanks to Richard Morgan for
3921 providing reproducible case.
3926 * Removed automatic addition of RPATH in shared libraries and executables,
3927 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3931 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3933 * Fix Use After Free for large message sizes
3935 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3936 message larger than approx 16k is received then the underlying buffer to
3937 store the incoming message is reallocated and moved. Unfortunately a
3938 dangling pointer to the old location is left which results in an attempt to
3939 write to the previously freed location. This is likely to result in a
3940 crash, however it could potentially lead to execution of arbitrary code.
3942 This issue only affects OpenSSL 1.1.0a.
3944 This issue was reported to OpenSSL by Robert Święcki.
3949 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3951 * OCSP Status Request extension unbounded memory growth
3953 A malicious client can send an excessively large OCSP Status Request
3954 extension. If that client continually requests renegotiation, sending a
3955 large OCSP Status Request extension each time, then there will be unbounded
3956 memory growth on the server. This will eventually lead to a Denial Of
3957 Service attack through memory exhaustion. Servers with a default
3958 configuration are vulnerable even if they do not support OCSP. Builds using
3959 the "no-ocsp" build time option are not affected.
3961 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3966 * SSL_peek() hang on empty record
3968 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3969 sends an empty record. This could be exploited by a malicious peer in a
3970 Denial Of Service attack.
3972 This issue was reported to OpenSSL by Alex Gaynor.
3977 * Excessive allocation of memory in tls_get_message_header() and
3978 dtls1_preprocess_fragment()
3980 A (D)TLS message includes 3 bytes for its length in the header for the
3981 message. This would allow for messages up to 16Mb in length. Messages of
3982 this length are excessive and OpenSSL includes a check to ensure that a
3983 peer is sending reasonably sized messages in order to avoid too much memory
3984 being consumed to service a connection. A flaw in the logic of version
3985 1.1.0 means that memory for the message is allocated too early, prior to
3986 the excessive message length check. Due to way memory is allocated in
3987 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3988 to service a connection. This could lead to a Denial of Service through
3989 memory exhaustion. However, the excessive message length check still takes
3990 place, and this would cause the connection to immediately fail. Assuming
3991 that the application calls SSL_free() on the failed connection in a timely
3992 manner then the 21Mb of allocated memory will then be immediately freed
3993 again. Therefore, the excessive memory allocation will be transitory in
3994 nature. This then means that there is only a security impact if:
3996 1) The application does not call SSL_free() in a timely manner in the event
3997 that the connection fails
3999 2) The application is working in a constrained environment where there is
4000 very little free memory
4002 3) The attacker initiates multiple connection attempts such that there are
4003 multiple connections in a state where memory has been allocated for the
4004 connection; SSL_free() has not yet been called; and there is insufficient
4005 memory to service the multiple requests.
4007 Except in the instance of (1) above any Denial Of Service is likely to be
4008 transitory because as soon as the connection fails the memory is
4009 subsequently freed again in the SSL_free() call. However there is an
4010 increased risk during this period of application crashes due to the lack of
4011 memory - which would then mean a more serious Denial of Service.
4013 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4014 (CVE-2016-6307 and CVE-2016-6308)
4018 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4019 had to be removed. Primary reason is that vendor assembler can't
4020 assemble our modules with -KPIC flag. As result it, assembly
4021 support, was not even available as option. But its lack means
4022 lack of side-channel resistant code, which is incompatible with
4023 security by todays standards. Fortunately gcc is readily available
4024 prepackaged option, which we firmly point at...
4028 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4030 * Windows command-line tool supports UTF-8 opt-in option for arguments
4031 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4032 (to any value) allows Windows user to access PKCS#12 file generated
4033 with Windows CryptoAPI and protected with non-ASCII password, as well
4034 as files generated under UTF-8 locale on Linux also protected with
4039 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4040 have been disabled by default and removed from DEFAULT, just like RC4.
4041 See the RC4 item below to re-enable both.
4045 * The method for finding the storage location for the Windows RAND seed file
4046 has changed. First we check %RANDFILE%. If that is not set then we check
4047 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4048 all else fails we fall back to C:\.
4052 * The EVP_EncryptUpdate() function has had its return type changed from void
4053 to int. A return of 0 indicates and error while a return of 1 indicates
4058 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4059 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4060 off the constant time implementation for RSA, DSA and DH have been made
4061 no-ops and deprecated.
4065 * Windows RAND implementation was simplified to only get entropy by
4066 calling CryptGenRandom(). Various other RAND-related tickets
4069 *Joseph Wylie Yandle, Rich Salz*
4071 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4072 and `OPENSSL_LH_`, respectively. The old names are available
4073 with API compatibility. They new names are now completely documented.
4077 * Unify TYPE_up_ref(obj) methods signature.
4078 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4079 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4080 int (instead of void) like all others TYPE_up_ref() methods.
4081 So now these methods also check the return value of CRYPTO_atomic_add(),
4082 and the validity of object reference counter.
4084 *fdasilvayy@gmail.com*
4086 * With Windows Visual Studio builds, the .pdb files are installed
4087 alongside the installed libraries and executables. For a static
4088 library installation, ossl_static.pdb is the associate compiler
4089 generated .pdb file to be used when linking programs.
4093 * Remove openssl.spec. Packaging files belong with the packagers.
4097 * Automatic Darwin/OSX configuration has had a refresh, it will now
4098 recognise x86_64 architectures automatically. You can still decide
4099 to build for a different bitness with the environment variable
4100 KERNEL_BITS (can be 32 or 64), for example:
4102 KERNEL_BITS=32 ./config
4106 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4107 256 bit AES and HMAC with SHA256.
4111 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4115 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4119 * To enable users to have their own config files and build file templates,
4120 Configure looks in the directory indicated by the environment variable
4121 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4122 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4123 name and is used as is.
4127 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4128 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4129 X509_CERT_FILE_CTX was removed.
4133 * "shared" builds are now the default. To create only static libraries use
4134 the "no-shared" Configure option.
4138 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4139 All of these option have not worked for some while and are fundamental
4144 * Make various cleanup routines no-ops and mark them as deprecated. Most
4145 global cleanup functions are no longer required because they are handled
4146 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4147 Explicitly de-initing can cause problems (e.g. where a library that uses
4148 OpenSSL de-inits, but an application is still using it). The affected
4149 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4150 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4151 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4152 COMP_zlib_cleanup().
4156 * --strict-warnings no longer enables runtime debugging options
4157 such as REF_DEBUG. Instead, debug options are automatically
4158 enabled with '--debug' builds.
4160 *Andy Polyakov, Emilia Käsper*
4162 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4163 have been moved out of the public header files. New functions for managing
4164 these have been added.
4168 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4169 objects have been moved out of the public header files. New
4170 functions for managing these have been added.
4174 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4175 have been moved out of the public header files. New functions for managing
4176 these have been added.
4180 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4181 moved out of the public header files. New functions for managing these
4186 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4190 * Removed the mk1mf build scripts.
4194 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4195 it is always safe to #include a header now.
4199 * Removed the aged BC-32 config and all its supporting scripts
4203 * Removed support for Ultrix, Netware, and OS/2.
4207 * Add support for HKDF.
4209 *Alessandro Ghedini*
4211 * Add support for blake2b and blake2s
4215 * Added support for "pipelining". Ciphers that have the
4216 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4217 encryptions/decryptions simultaneously. There are currently no built-in
4218 ciphers with this property but the expectation is that engines will be able
4219 to offer it to significantly improve throughput. Support has been extended
4220 into libssl so that multiple records for a single connection can be
4221 processed in one go (for >=TLS 1.1).
4225 * Added the AFALG engine. This is an async capable engine which is able to
4226 offload work to the Linux kernel. In this initial version it only supports
4227 AES128-CBC. The kernel must be version 4.1.0 or greater.
4231 * OpenSSL now uses a new threading API. It is no longer necessary to
4232 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4233 are two supported threading models: pthreads and windows threads. It is
4234 also possible to configure OpenSSL at compile time for "no-threads". The
4235 old threading API should no longer be used. The functions have been
4236 replaced with "no-op" compatibility macros.
4238 *Alessandro Ghedini, Matt Caswell*
4240 * Modify behavior of ALPN to invoke callback after SNI/servername
4241 callback, such that updates to the SSL_CTX affect ALPN.
4245 * Add SSL_CIPHER queries for authentication and key-exchange.
4249 * Changes to the DEFAULT cipherlist:
4250 - Prefer (EC)DHE handshakes over plain RSA.
4251 - Prefer AEAD ciphers over legacy ciphers.
4252 - Prefer ECDSA over RSA when both certificates are available.
4253 - Prefer TLSv1.2 ciphers/PRF.
4254 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4259 * Change the ECC default curve list to be this, in order: x25519,
4260 secp256r1, secp521r1, secp384r1.
4264 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4265 disabled by default. They can be re-enabled using the
4266 enable-weak-ssl-ciphers option to Configure.
4270 * If the server has ALPN configured, but supports no protocols that the
4271 client advertises, send a fatal "no_application_protocol" alert.
4272 This behaviour is SHALL in RFC 7301, though it isn't universally
4273 implemented by other servers.
4277 * Add X25519 support.
4278 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4279 for public and private key encoding using the format documented in
4280 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4281 key generation and key derivation.
4283 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4288 * Deprecate SRP_VBASE_get_by_user.
4289 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4290 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4291 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4292 seed, even if the seed is configured.
4294 Users should use SRP_VBASE_get1_by_user instead. Note that in
4295 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4296 also that even though configuring the SRP seed attempts to hide
4297 invalid usernames by continuing the handshake with fake
4298 credentials, this behaviour is not constant time and no strong
4299 guarantees are made that the handshake is indistinguishable from
4300 that of a valid user.
4304 * Configuration change; it's now possible to build dynamic engines
4305 without having to build shared libraries and vice versa. This
4306 only applies to the engines in `engines/`, those in `crypto/engine/`
4307 will always be built into libcrypto (i.e. "static").
4309 Building dynamic engines is enabled by default; to disable, use
4310 the configuration option "disable-dynamic-engine".
4312 The only requirements for building dynamic engines are the
4313 presence of the DSO module and building with position independent
4314 code, so they will also automatically be disabled if configuring
4315 with "disable-dso" or "disable-pic".
4317 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4318 are also taken away from openssl/opensslconf.h, as they are
4323 * Configuration change; if there is a known flag to compile
4324 position independent code, it will always be applied on the
4325 libcrypto and libssl object files, and never on the application
4326 object files. This means other libraries that use routines from
4327 libcrypto / libssl can be made into shared libraries regardless
4328 of how OpenSSL was configured.
4330 If this isn't desirable, the configuration options "disable-pic"
4331 or "no-pic" can be used to disable the use of PIC. This will
4332 also disable building shared libraries and dynamic engines.
4336 * Removed JPAKE code. It was experimental and has no wide use.
4340 * The INSTALL_PREFIX Makefile variable has been renamed to
4341 DESTDIR. That makes for less confusion on what this variable
4342 is for. Also, the configuration option --install_prefix is
4347 * Heartbeat for TLS has been removed and is disabled by default
4348 for DTLS; configure with enable-heartbeats. Code that uses the
4349 old #define's might need to be updated.
4351 *Emilia Käsper, Rich Salz*
4353 * Rename REF_CHECK to REF_DEBUG.
4357 * New "unified" build system
4359 The "unified" build system is aimed to be a common system for all
4360 platforms we support. With it comes new support for VMS.
4362 This system builds supports building in a different directory tree
4363 than the source tree. It produces one Makefile (for unix family
4364 or lookalikes), or one descrip.mms (for VMS).
4366 The source of information to make the Makefile / descrip.mms is
4367 small files called 'build.info', holding the necessary
4368 information for each directory with source to compile, and a
4369 template in Configurations, like unix-Makefile.tmpl or
4372 With this change, the library names were also renamed on Windows
4373 and on VMS. They now have names that are closer to the standard
4374 on Unix, and include the major version number, and in certain
4375 cases, the architecture they are built for. See "Notes on shared
4376 libraries" in INSTALL.
4378 We rely heavily on the perl module Text::Template.
4382 * Added support for auto-initialisation and de-initialisation of the library.
4383 OpenSSL no longer requires explicit init or deinit routines to be called,
4384 except in certain circumstances. See the OPENSSL_init_crypto() and
4385 OPENSSL_init_ssl() man pages for further information.
4389 * The arguments to the DTLSv1_listen function have changed. Specifically the
4390 "peer" argument is now expected to be a BIO_ADDR object.
4392 * Rewrite of BIO networking library. The BIO library lacked consistent
4393 support of IPv6, and adding it required some more extensive
4394 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4395 which hold all types of addresses and chains of address information.
4396 It also introduces a new API, with functions like BIO_socket,
4397 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4398 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4399 have been adapted accordingly.
4403 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4408 * CRIME protection: disable compression by default, even if OpenSSL is
4409 compiled with zlib enabled. Applications can still enable compression
4410 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4411 using the SSL_CONF library to configure compression.
4415 * The signature of the session callback configured with
4416 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4417 was explicitly marked as `const unsigned char*` instead of
4422 * Always DPURIFY. Remove the use of uninitialized memory in the
4423 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4427 * Removed many obsolete configuration items, including
4428 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4429 MD2_CHAR, MD2_INT, MD2_LONG
4431 IDEA_SHORT, IDEA_LONG
4432 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4434 *Rich Salz, with advice from Andy Polyakov*
4436 * Many BN internals have been moved to an internal header file.
4438 *Rich Salz with help from Andy Polyakov*
4440 * Configuration and writing out the results from it has changed.
4441 Files such as Makefile include/openssl/opensslconf.h and are now
4442 produced through general templates, such as Makefile.in and
4443 crypto/opensslconf.h.in and some help from the perl module
4446 Also, the center of configuration information is no longer
4447 Makefile. Instead, Configure produces a perl module in
4448 configdata.pm which holds most of the config data (in the hash
4449 table %config), the target data that comes from the target
4450 configuration in one of the `Configurations/*.conf` files (in
4455 * To clarify their intended purposes, the Configure options
4456 --prefix and --openssldir change their semantics, and become more
4457 straightforward and less interdependent.
4459 --prefix shall be used exclusively to give the location INSTALLTOP
4460 where programs, scripts, libraries, include files and manuals are
4461 going to be installed. The default is now /usr/local.
4463 --openssldir shall be used exclusively to give the default
4464 location OPENSSLDIR where certificates, private keys, CRLs are
4465 managed. This is also where the default openssl.cnf gets
4467 If the directory given with this option is a relative path, the
4468 values of both the --prefix value and the --openssldir value will
4469 be combined to become OPENSSLDIR.
4470 The default for --openssldir is INSTALLTOP/ssl.
4472 Anyone who uses --openssldir to specify where OpenSSL is to be
4473 installed MUST change to use --prefix instead.
4477 * The GOST engine was out of date and therefore it has been removed. An up
4478 to date GOST engine is now being maintained in an external repository.
4479 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4480 support for GOST ciphersuites (these are only activated if a GOST engine
4485 * EGD is no longer supported by default; use enable-egd when
4488 *Ben Kaduk and Rich Salz*
4490 * The distribution now has Makefile.in files, which are used to
4491 create Makefile's when Configure is run. *Configure must be run
4492 before trying to build now.*
4496 * The return value for SSL_CIPHER_description() for error conditions
4501 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4503 Obtaining and performing DNSSEC validation of TLSA records is
4504 the application's responsibility. The application provides
4505 the TLSA records of its choice to OpenSSL, and these are then
4506 used to authenticate the peer.
4508 The TLSA records need not even come from DNS. They can, for
4509 example, be used to implement local end-entity certificate or
4510 trust-anchor "pinning", where the "pin" data takes the form
4511 of TLSA records, which can augment or replace verification
4512 based on the usual WebPKI public certification authorities.
4516 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4517 continues to support deprecated interfaces in default builds.
4518 However, applications are strongly advised to compile their
4519 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4520 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4521 or the 1.1.0 releases.
4523 In environments in which all applications have been ported to
4524 not use any deprecated interfaces OpenSSL's Configure script
4525 should be used with the --api=1.1.0 option to entirely remove
4526 support for the deprecated features from the library and
4527 unconditionally disable them in the installed headers.
4528 Essentially the same effect can be achieved with the "no-deprecated"
4529 argument to Configure, except that this will always restrict
4530 the build to just the latest API, rather than a fixed API
4533 As applications are ported to future revisions of the API,
4534 they should update their compile-time OPENSSL_API_COMPAT define
4535 accordingly, but in most cases should be able to continue to
4536 compile with later releases.
4538 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4539 0x10000000L and 0x00908000L, respectively. However those
4540 versions did not support the OPENSSL_API_COMPAT feature, and
4541 so applications are not typically tested for explicit support
4542 of just the undeprecated features of either release.
4546 * Add support for setting the minimum and maximum supported protocol.
4547 It can bet set via the SSL_set_min_proto_version() and
4548 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4549 MaxProtocol. It's recommended to use the new APIs to disable
4550 protocols instead of disabling individual protocols using
4551 SSL_set_options() or SSL_CONF's Protocol. This change also
4552 removes support for disabling TLS 1.2 in the OpenSSL TLS
4553 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4557 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4561 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4562 and integrates ECDSA and ECDH functionality into EC. Implementations can
4563 now redirect key generation and no longer need to convert to or from
4566 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4567 include the ec.h header file instead.
4571 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4572 ciphers who are no longer supported and drops support the ephemeral RSA key
4573 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4577 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4578 opaque. For HMAC_CTX, the following constructors and destructors
4581 HMAC_CTX *HMAC_CTX_new(void);
4582 void HMAC_CTX_free(HMAC_CTX *ctx);
4584 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4585 destroy such methods has been added. See EVP_MD_meth_new(3) and
4586 EVP_CIPHER_meth_new(3) for documentation.
4589 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4590 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4591 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4592 an already created structure.
4593 2) For consistency with the majority of our object creators and
4594 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4595 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4596 for deprecated builds.
4600 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4601 cryptographic operations to be performed asynchronously as long as an
4602 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4603 further details. Libssl has also had this capability integrated with the
4604 introduction of the new mode SSL_MODE_ASYNC and associated error
4605 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4606 pages. This work was developed in partnership with Intel Corp.
4610 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4611 always enabled now. If you want to disable the support you should
4612 exclude it using the list of supported ciphers. This also means that the
4613 "-no_ecdhe" option has been removed from s_server.
4617 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4618 SSL_{CTX_}set1_curves() which can set a list.
4622 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4623 curve you want to support using SSL_{CTX_}set1_curves().
4627 * State machine rewrite. The state machine code has been significantly
4628 refactored in order to remove much duplication of code and solve issues
4629 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4630 further details). This change does have some associated API changes.
4631 Notably the SSL_state() function has been removed and replaced by
4632 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4633 SSL_set_state() has been removed altogether. The previous handshake states
4634 defined in ssl.h and ssl3.h have also been removed.
4638 * All instances of the string "ssleay" in the public API were replaced
4639 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4640 Some error codes related to internal RSA_eay API's were renamed.
4644 * The demo files in crypto/threads were moved to demo/threads.
4648 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4651 *Matt Caswell, Rich Salz*
4653 * New ASN.1 embed macro.
4655 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4656 structure is not allocated: it is part of the parent. That is instead of
4664 This reduces memory fragmentation and make it impossible to accidentally
4665 set a mandatory field to NULL.
4667 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4668 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4669 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4674 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4678 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4679 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4680 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4681 DES and RC4 ciphersuites.
4685 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4686 This changes the decoding behaviour for some invalid messages,
4687 though the change is mostly in the more lenient direction, and
4688 legacy behaviour is preserved as much as possible.
4692 * Fix no-stdio build.
4693 *David Woodhouse <David.Woodhouse@intel.com> and also*
4694 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4696 * New testing framework
4697 The testing framework has been largely rewritten and is now using
4698 perl and the perl modules Test::Harness and an extended variant of
4699 Test::More called OpenSSL::Test to do its work. All test scripts in
4700 test/ have been rewritten into test recipes, and all direct calls to
4701 executables in test/Makefile have become individual recipes using the
4702 simplified testing OpenSSL::Test::Simple.
4704 For documentation on our testing modules, do:
4706 perldoc test/testlib/OpenSSL/Test/Simple.pm
4707 perldoc test/testlib/OpenSSL/Test.pm
4711 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4712 are used; the latter aborts on memory leaks (usually checked on exit).
4713 Some undocumented "set malloc, etc., hooks" functions were removed
4714 and others were changed. All are now documented.
4718 * In DSA_generate_parameters_ex, if the provided seed is too short,
4721 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4723 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4724 from RFC4279, RFC4785, RFC5487, RFC5489.
4726 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4727 original RSA_PSK patch.
4731 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4732 era flag was never set throughout the codebase (only read). Also removed
4733 SSL3_FLAGS_POP_BUFFER which was only used if
4734 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4738 * Changed the default name options in the "ca", "crl", "req" and "x509"
4739 to be "oneline" instead of "compat".
4743 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4744 not aware of clients that still exhibit this bug, and the workaround
4745 hasn't been working properly for a while.
4749 * The return type of BIO_number_read() and BIO_number_written() as well as
4750 the corresponding num_read and num_write members in the BIO structure has
4751 changed from unsigned long to uint64_t. On platforms where an unsigned
4752 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4757 * Given the pervasive nature of TLS extensions it is inadvisable to run
4758 OpenSSL without support for them. It also means that maintaining
4759 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4760 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4764 * Removed support for the two export grade static DH ciphersuites
4765 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4766 were newly added (along with a number of other static DH ciphersuites) to
4767 1.0.2. However the two export ones have *never* worked since they were
4768 introduced. It seems strange in any case to be adding new export
4769 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4773 * Version negotiation has been rewritten. In particular SSLv23_method(),
4774 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4775 and turned into macros which simply call the new preferred function names
4776 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4777 should use the new names instead. Also as part of this change the ssl23.h
4778 header file has been removed.
4782 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4783 code and the associated standard is no longer considered fit-for-purpose.
4787 * RT2547 was closed. When generating a private key, try to make the
4788 output file readable only by the owner. This behavior change might
4789 be noticeable when interacting with other software.
4791 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4796 * Added HTTP GET support to the ocsp command.
4800 * Changed default digest for the dgst and enc commands from MD5 to
4805 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4809 * Added support for TLS extended master secret from
4810 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4811 initial patch which was a great help during development.
4815 * All libssl internal structures have been removed from the public header
4816 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4817 now redundant). Users should not attempt to access internal structures
4818 directly. Instead they should use the provided API functions.
4822 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4823 Access to deprecated functions can be re-enabled by running config with
4824 "enable-deprecated". In addition applications wishing to use deprecated
4825 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4826 will, by default, disable some transitive includes that previously existed
4827 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4831 * Added support for OCB mode. OpenSSL has been granted a patent license
4832 compatible with the OpenSSL license for use of OCB. Details are available
4833 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4834 for OCB can be removed by calling config with no-ocb.
4838 * SSLv2 support has been removed. It still supports receiving an SSLv2
4839 compatible client hello.
4843 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4844 done while fixing the error code for the key-too-small case.
4846 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4848 * CA.sh has been removed; use CA.pl instead.
4852 * Removed old DES API.
4856 * Remove various unsupported platforms:
4862 Sinix/ReliantUNIX RM400
4867 16-bit platforms such as WIN16
4871 * Clean up OPENSSL_NO_xxx #define's
4872 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4873 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4874 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4875 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4876 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4877 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4878 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4879 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4880 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4881 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4885 * Cleaned up dead code
4886 Remove all but one '#ifdef undef' which is to be looked at.
4890 * Clean up calling of xxx_free routines.
4891 Just like free(), fix most of the xxx_free routines to accept
4892 NULL. Remove the non-null checks from callers. Save much code.
4896 * Add secure heap for storage of private keys (when possible).
4897 Add BIO_s_secmem(), CBIGNUM, etc.
4898 Contributed by Akamai Technologies under our Corporate CLA.
4902 * Experimental support for a new, fast, unbiased prime candidate generator,
4903 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4905 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4907 * New output format NSS in the sess_id command line tool. This allows
4908 exporting the session id and the master key in NSS keylog format.
4910 *Martin Kaiser <martin@kaiser.cx>*
4912 * Harmonize version and its documentation. -f flag is used to display
4915 *mancha <mancha1@zoho.com>*
4917 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4918 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4920 *mancha <mancha1@zoho.com>*
4922 * Fix some double frees. These are not thought to be exploitable.
4924 *mancha <mancha1@zoho.com>*
4926 * A missing bounds check in the handling of the TLS heartbeat extension
4927 can be used to reveal up to 64k of memory to a connected client or
4930 Thanks for Neel Mehta of Google Security for discovering this bug and to
4931 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4932 preparing the fix ([CVE-2014-0160])
4934 *Adam Langley, Bodo Moeller*
4936 * Fix for the attack described in the paper "Recovering OpenSSL
4937 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4938 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4939 <http://eprint.iacr.org/2014/140>
4941 Thanks to Yuval Yarom and Naomi Benger for discovering this
4942 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4944 *Yuval Yarom and Naomi Benger*
4946 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4947 this fixes a limitation in previous versions of OpenSSL.
4951 * Experimental encrypt-then-mac support.
4953 Experimental support for encrypt then mac from
4954 draft-gutmann-tls-encrypt-then-mac-02.txt
4956 To enable it set the appropriate extension number (0x42 for the test
4957 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4959 For non-compliant peers (i.e. just about everything) this should have no
4962 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4966 * Add EVP support for key wrapping algorithms, to avoid problems with
4967 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4968 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4969 algorithms and include tests cases.
4973 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4978 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4979 MGF1 digest and OAEP label.
4983 * Make openssl verify return errors.
4985 *Chris Palmer <palmer@google.com> and Ben Laurie*
4987 * New function ASN1_TIME_diff to calculate the difference between two
4988 ASN1_TIME structures or one structure and the current time.
4992 * Update fips_test_suite to support multiple command line options. New
4993 test to induce all self test errors in sequence and check expected
4998 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4999 sign or verify all in one operation.
5003 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5004 test programs and fips_test_suite. Includes functionality to parse
5005 the minimal script output of fipsalgest.pl directly.
5009 * Add authorisation parameter to FIPS_module_mode_set().
5013 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5017 * Use separate DRBG fields for internal and external flags. New function
5018 FIPS_drbg_health_check() to perform on demand health checking. Add
5019 generation tests to fips_test_suite with reduced health check interval to
5020 demonstrate periodic health checking. Add "nodh" option to
5021 fips_test_suite to skip very slow DH test.
5025 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5030 * More extensive health check for DRBG checking many more failure modes.
5031 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5032 combination: call this in fips_test_suite.
5036 * Add support for canonical generation of DSA parameter 'g'. See
5039 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5040 POST to handle HMAC cases.
5044 * Add functions FIPS_module_version() and FIPS_module_version_text()
5045 to return numerical and string versions of the FIPS module number.
5049 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5050 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5051 outside the validated module in the FIPS capable OpenSSL.
5055 * Minor change to DRBG entropy callback semantics. In some cases
5056 there is no multiple of the block length between min_len and
5057 max_len. Allow the callback to return more than max_len bytes
5058 of entropy but discard any extra: it is the callback's responsibility
5059 to ensure that the extra data discarded does not impact the
5060 requested amount of entropy.
5064 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5065 information in FIPS186-3, SP800-57 and SP800-131A.
5069 * CCM support via EVP. Interface is very similar to GCM case except we
5070 must supply all data in one chunk (i.e. no update, final) and the
5071 message length must be supplied if AAD is used. Add algorithm test
5076 * Initial version of POST overhaul. Add POST callback to allow the status
5077 of POST to be monitored and/or failures induced. Modify fips_test_suite
5078 to use callback. Always run all selftests even if one fails.
5082 * XTS support including algorithm test driver in the fips_gcmtest program.
5083 Note: this does increase the maximum key length from 32 to 64 bytes but
5084 there should be no binary compatibility issues as existing applications
5085 will never use XTS mode.
5089 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5090 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5091 performs algorithm blocking for unapproved PRNG types. Also do not
5092 set PRNG type in FIPS_mode_set(): leave this to the application.
5093 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5094 the standard OpenSSL PRNG: set additional data to a date time vector.
5098 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5099 This shouldn't present any incompatibility problems because applications
5100 shouldn't be using these directly and any that are will need to rethink
5101 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5105 * Extensive self tests and health checking required by SP800-90 DRBG.
5106 Remove strength parameter from FIPS_drbg_instantiate and always
5107 instantiate at maximum supported strength.
5111 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5115 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5119 * New function DH_compute_key_padded() to compute a DH key and pad with
5120 leading zeroes if needed: this complies with SP800-56A et al.
5124 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5125 anything, incomplete, subject to change and largely untested at present.
5129 * Modify fipscanisteronly build option to only build the necessary object
5130 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5134 * Add experimental option FIPSSYMS to give all symbols in
5135 fipscanister.o and FIPS or fips prefix. This will avoid
5136 conflicts with future versions of OpenSSL. Add perl script
5137 util/fipsas.pl to preprocess assembly language source files
5138 and rename any affected symbols.
5142 * Add selftest checks and algorithm block of non-fips algorithms in
5143 FIPS mode. Remove DES2 from selftests.
5147 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5148 return internal method without any ENGINE dependencies. Add new
5149 tiny fips sign and verify functions.
5153 * New build option no-ec2m to disable characteristic 2 code.
5157 * New build option "fipscanisteronly". This only builds fipscanister.o
5158 and (currently) associated fips utilities. Uses the file Makefile.fips
5159 instead of Makefile.org as the prototype.
5163 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5164 Update fips_gcmtest to use IV generator.
5168 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5169 setting output buffer to NULL. The `*Final` function must be
5170 called although it will not retrieve any additional data. The tag
5171 can be set or retrieved with a ctrl. The IV length is by default 12
5172 bytes (96 bits) but can be set to an alternative value. If the IV
5173 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5178 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5179 underlying do_cipher function handles all cipher semantics itself
5180 including padding and finalisation. This is useful if (for example)
5181 an ENGINE cipher handles block padding itself. The behaviour of
5182 do_cipher is subtly changed if this flag is set: the return value
5183 is the number of characters written to the output buffer (zero is
5184 no longer an error code) or a negative error code. Also if the
5185 input buffer is NULL and length 0 finalisation should be performed.
5189 * If a candidate issuer certificate is already part of the constructed
5190 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5194 * Improve forward-security support: add functions
5196 void SSL_CTX_set_not_resumable_session_callback(
5197 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5198 void SSL_set_not_resumable_session_callback(
5199 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5201 for use by SSL/TLS servers; the callback function will be called whenever a
5202 new session is created, and gets to decide whether the session may be
5203 cached to make it resumable (return 0) or not (return 1). (As by the
5204 SSL/TLS protocol specifications, the session_id sent by the server will be
5205 empty to indicate that the session is not resumable; also, the server will
5206 not generate RFC 4507 (RFC 5077) session tickets.)
5208 A simple reasonable callback implementation is to return is_forward_secure.
5209 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5210 by the SSL/TLS server library, indicating whether it can provide forward
5213 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5215 * New -verify_name option in command line utilities to set verification
5220 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5221 Add CMAC pkey methods.
5225 * Experimental renegotiation in s_server -www mode. If the client
5226 browses /reneg connection is renegotiated. If /renegcert it is
5227 renegotiated requesting a certificate.
5231 * Add an "external" session cache for debugging purposes to s_server. This
5232 should help trace issues which normally are only apparent in deployed
5233 multi-process servers.
5237 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5238 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5239 BIO_set_cipher() and some obscure PEM functions were changed so they
5240 can now return an error. The RAND changes required a change to the
5241 RAND_METHOD structure.
5245 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5246 a gcc attribute to warn if the result of a function is ignored. This
5247 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5248 whose return value is often ignored.
5252 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5253 These allow SCTs (signed certificate timestamps) to be requested and
5254 validated when establishing a connection.
5256 *Rob Percival <robpercival@google.com>*
5261 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5263 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5264 used even when parsing explicit parameters, when loading a encoded key
5265 or calling `EC_GROUP_new_from_ecpkparameters()`/
5266 `EC_GROUP_new_from_ecparameters()`.
5267 This prevents bypass of security hardening and performance gains,
5268 especially for curves with specialized EC_METHODs.
5269 By default, if a key encoded with explicit parameters is loaded and later
5270 encoded, the output is still encoded with explicit parameters, even if
5271 internally a "named" EC_GROUP is used for computation.
5275 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5276 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5277 NULL. After this change, only the cofactor parameter can be NULL. It also
5278 does some minimal sanity checks on the passed order.
5283 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5284 An attack is simple, if the first CMS_recipientInfo is valid but the
5285 second CMS_recipientInfo is chosen ciphertext. If the second
5286 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5287 encryption key will be replaced by garbage, and the message cannot be
5288 decoded, but if the RSA decryption fails, the correct encryption key is
5289 used and the recipient will not notice the attack.
5290 As a work around for this potential attack the length of the decrypted
5291 key must be equal to the cipher default key length, in case the
5292 certificate is not given and all recipientInfo are tried out.
5293 The old behaviour can be re-enabled in the CMS code by setting the
5294 CMS_DEBUG_DECRYPT flag.
5299 * Document issue with installation paths in diverse Windows builds
5301 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5302 binaries and run-time config file.
5307 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5309 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5310 This changes the size when using the `genpkey` command when no size is given.
5311 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5312 generation commands to use 2048 bits by default.
5316 * Add FIPS support for Android Arm 64-bit
5318 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5319 Module in Version 2.0.10. For some reason, the corresponding target
5320 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5321 built with FIPS support on Android Arm 64-bit. This omission has been
5324 *Matthias St. Pierre*
5326 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5328 * 0-byte record padding oracle
5330 If an application encounters a fatal protocol error and then calls
5331 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5332 then OpenSSL can respond differently to the calling application if a 0 byte
5333 record is received with invalid padding compared to if a 0 byte record is
5334 received with an invalid MAC. If the application then behaves differently
5335 based on that in a way that is detectable to the remote peer, then this
5336 amounts to a padding oracle that could be used to decrypt data.
5338 In order for this to be exploitable "non-stitched" ciphersuites must be in
5339 use. Stitched ciphersuites are optimised implementations of certain
5340 commonly used ciphersuites. Also the application must call SSL_shutdown()
5341 twice even if a protocol error has occurred (applications should not do
5342 this but some do anyway).
5344 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5345 Aviram, with additional investigation by Steven Collison and Andrew
5346 Hourselt. It was reported to OpenSSL on 10th December 2018.
5351 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5355 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5357 * Microarchitecture timing vulnerability in ECC scalar multiplication
5359 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5360 shown to be vulnerable to a microarchitecture timing side channel attack.
5361 An attacker with sufficient access to mount local timing attacks during
5362 ECDSA signature generation could recover the private key.
5364 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5365 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5371 * Timing vulnerability in DSA signature generation
5373 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5374 timing side channel attack. An attacker could use variations in the signing
5375 algorithm to recover the private key.
5377 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5382 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5383 Module, accidentally introduced while backporting security fixes from the
5384 development branch and hindering the use of ECC in FIPS mode.
5388 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5390 * Client DoS due to large DH parameter
5392 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5393 malicious server can send a very large prime value to the client. This will
5394 cause the client to spend an unreasonably long period of time generating a
5395 key for this prime resulting in a hang until the client has finished. This
5396 could be exploited in a Denial Of Service attack.
5398 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5403 * Cache timing vulnerability in RSA Key Generation
5405 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5406 a cache timing side channel attack. An attacker with sufficient access to
5407 mount cache timing attacks during the RSA key generation process could
5408 recover the private key.
5410 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5411 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5416 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5417 parameter is no longer accepted, as it leads to a corrupt table. NULL
5418 pem_str is reserved for alias entries only.
5422 * Revert blinding in ECDSA sign and instead make problematic addition
5423 length-invariant. Switch even to fixed-length Montgomery multiplication.
5427 * Change generating and checking of primes so that the error rate of not
5428 being prime depends on the intended use based on the size of the input.
5429 For larger primes this will result in more rounds of Miller-Rabin.
5430 The maximal error rate for primes with more than 1080 bits is lowered
5433 *Kurt Roeckx, Annie Yousar*
5435 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5439 * Add blinding to ECDSA and DSA signatures to protect against side channel
5440 attacks discovered by Keegan Ryan (NCC Group).
5444 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5445 now allow empty (zero character) pass phrases.
5449 * Certificate time validation (X509_cmp_time) enforces stricter
5450 compliance with RFC 5280. Fractional seconds and timezone offsets
5451 are no longer allowed.
5455 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5457 * Constructed ASN.1 types with a recursive definition could exceed the stack
5459 Constructed ASN.1 types with a recursive definition (such as can be found
5460 in PKCS7) could eventually exceed the stack given malicious input with
5461 excessive recursion. This could result in a Denial Of Service attack. There
5462 are no such structures used within SSL/TLS that come from untrusted sources
5463 so this is considered safe.
5465 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5471 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5473 * Read/write after SSL object in error state
5475 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5476 mechanism. The intent was that if a fatal error occurred during a handshake
5477 then OpenSSL would move into the error state and would immediately fail if
5478 you attempted to continue the handshake. This works as designed for the
5479 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5480 SSL_connect()), however due to a bug it does not work correctly if
5481 SSL_read() or SSL_write() is called directly. In that scenario, if the
5482 handshake fails then a fatal error will be returned in the initial function
5483 call. If SSL_read()/SSL_write() is subsequently called by the application
5484 for the same SSL object then it will succeed and the data is passed without
5485 being decrypted/encrypted directly from the SSL/TLS record layer.
5487 In order to exploit this issue an application bug would have to be present
5488 that resulted in a call to SSL_read()/SSL_write() being issued after having
5489 already received a fatal error.
5491 This issue was reported to OpenSSL by David Benjamin (Google).
5496 * rsaz_1024_mul_avx2 overflow bug on x86_64
5498 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5499 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5500 Analysis suggests that attacks against RSA and DSA as a result of this
5501 defect would be very difficult to perform and are not believed likely.
5502 Attacks against DH1024 are considered just feasible, because most of the
5503 work necessary to deduce information about a private key may be performed
5504 offline. The amount of resources required for such an attack would be
5505 significant. However, for an attack on TLS to be meaningful, the server
5506 would have to share the DH1024 private key among multiple clients, which is
5507 no longer an option since CVE-2016-0701.
5509 This only affects processors that support the AVX2 but not ADX extensions
5510 like Intel Haswell (4th generation).
5512 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5513 was originally found via the OSS-Fuzz project.
5518 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5520 * bn_sqrx8x_internal carry bug on x86_64
5522 There is a carry propagating bug in the x86_64 Montgomery squaring
5523 procedure. No EC algorithms are affected. Analysis suggests that attacks
5524 against RSA and DSA as a result of this defect would be very difficult to
5525 perform and are not believed likely. Attacks against DH are considered just
5526 feasible (although very difficult) because most of the work necessary to
5527 deduce information about a private key may be performed offline. The amount
5528 of resources required for such an attack would be very significant and
5529 likely only accessible to a limited number of attackers. An attacker would
5530 additionally need online access to an unpatched system using the target
5531 private key in a scenario with persistent DH parameters and a private
5532 key that is shared between multiple clients.
5534 This only affects processors that support the BMI1, BMI2 and ADX extensions
5535 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5537 This issue was reported to OpenSSL by the OSS-Fuzz project.
5542 * Malformed X.509 IPAddressFamily could cause OOB read
5544 If an X.509 certificate has a malformed IPAddressFamily extension,
5545 OpenSSL could do a one-byte buffer overread. The most likely result
5546 would be an erroneous display of the certificate in text format.
5548 This issue was reported to OpenSSL by the OSS-Fuzz project.
5552 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5554 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5555 platform rather than 'mingw'.
5559 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5561 * Truncated packet could crash via OOB read
5563 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5564 cipher is being used, then a truncated packet can cause that host to
5565 perform an out-of-bounds read, usually resulting in a crash.
5567 This issue was reported to OpenSSL by Robert Święcki of Google.
5572 * BN_mod_exp may produce incorrect results on x86_64
5574 There is a carry propagating bug in the x86_64 Montgomery squaring
5575 procedure. No EC algorithms are affected. Analysis suggests that attacks
5576 against RSA and DSA as a result of this defect would be very difficult to
5577 perform and are not believed likely. Attacks against DH are considered just
5578 feasible (although very difficult) because most of the work necessary to
5579 deduce information about a private key may be performed offline. The amount
5580 of resources required for such an attack would be very significant and
5581 likely only accessible to a limited number of attackers. An attacker would
5582 additionally need online access to an unpatched system using the target
5583 private key in a scenario with persistent DH parameters and a private
5584 key that is shared between multiple clients. For example this can occur by
5585 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5586 similar to CVE-2015-3193 but must be treated as a separate problem.
5588 This issue was reported to OpenSSL by the OSS-Fuzz project.
5593 * Montgomery multiplication may produce incorrect results
5595 There is a carry propagating bug in the Broadwell-specific Montgomery
5596 multiplication procedure that handles input lengths divisible by, but
5597 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5598 and DH private keys are impossible. This is because the subroutine in
5599 question is not used in operations with the private key itself and an input
5600 of the attacker's direct choice. Otherwise the bug can manifest itself as
5601 transient authentication and key negotiation failures or reproducible
5602 erroneous outcome of public-key operations with specially crafted input.
5603 Among EC algorithms only Brainpool P-512 curves are affected and one
5604 presumably can attack ECDH key negotiation. Impact was not analyzed in
5605 detail, because pre-requisites for attack are considered unlikely. Namely
5606 multiple clients have to choose the curve in question and the server has to
5607 share the private key among them, neither of which is default behaviour.
5608 Even then only clients that chose the curve will be affected.
5610 This issue was publicly reported as transient failures and was not
5611 initially recognized as a security issue. Thanks to Richard Morgan for
5612 providing reproducible case.
5617 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5618 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5619 prevent issues where no progress is being made and the peer continually
5620 sends unrecognised record types, using up resources processing them.
5624 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5626 * Missing CRL sanity check
5628 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5629 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5630 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5632 This issue only affects the OpenSSL 1.0.2i
5637 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5639 * OCSP Status Request extension unbounded memory growth
5641 A malicious client can send an excessively large OCSP Status Request
5642 extension. If that client continually requests renegotiation, sending a
5643 large OCSP Status Request extension each time, then there will be unbounded
5644 memory growth on the server. This will eventually lead to a Denial Of
5645 Service attack through memory exhaustion. Servers with a default
5646 configuration are vulnerable even if they do not support OCSP. Builds using
5647 the "no-ocsp" build time option are not affected.
5649 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5654 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5657 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5663 * OOB write in MDC2_Update()
5665 An overflow can occur in MDC2_Update() either if called directly or
5666 through the EVP_DigestUpdate() function using MDC2. If an attacker
5667 is able to supply very large amounts of input data after a previous
5668 call to EVP_EncryptUpdate() with a partial block then a length check
5669 can overflow resulting in a heap corruption.
5671 The amount of data needed is comparable to SIZE_MAX which is impractical
5674 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5679 * Malformed SHA512 ticket DoS
5681 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5682 DoS attack where a malformed ticket will result in an OOB read which will
5685 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5686 a custom server callback and ticket lookup mechanism.
5688 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5693 * OOB write in BN_bn2dec()
5695 The function BN_bn2dec() does not check the return value of BN_div_word().
5696 This can cause an OOB write if an application uses this function with an
5697 overly large BIGNUM. This could be a problem if an overly large certificate
5698 or CRL is printed out from an untrusted source. TLS is not affected because
5699 record limits will reject an oversized certificate before it is parsed.
5701 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5706 * OOB read in TS_OBJ_print_bio()
5708 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5709 the total length the OID text representation would use and not the amount
5710 of data written. This will result in OOB reads when large OIDs are
5713 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5718 * Pointer arithmetic undefined behaviour
5720 Avoid some undefined pointer arithmetic
5722 A common idiom in the codebase is to check limits in the following manner:
5725 Where "p" points to some malloc'd data of SIZE bytes and
5728 "len" here could be from some externally supplied data (e.g. from a TLS
5731 The rules of C pointer arithmetic are such that "p + len" is only well
5732 defined where len <= SIZE. Therefore the above idiom is actually
5733 undefined behaviour.
5735 For example this could cause problems if some malloc implementation
5736 provides an address for "p" such that "p + len" actually overflows for
5737 values of len that are too big and therefore p + len < limit.
5739 This issue was reported to OpenSSL by Guido Vranken
5744 * Constant time flag not preserved in DSA signing
5746 Operations in the DSA signing algorithm should run in constant time in
5747 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5748 implementation means that a non-constant time codepath is followed for
5749 certain operations. This has been demonstrated through a cache-timing
5750 attack to be sufficient for an attacker to recover the private DSA key.
5752 This issue was reported by César Pereida (Aalto University), Billy Brumley
5753 (Tampere University of Technology), and Yuval Yarom (The University of
5754 Adelaide and NICTA).
5759 * DTLS buffered message DoS
5761 In a DTLS connection where handshake messages are delivered out-of-order
5762 those messages that OpenSSL is not yet ready to process will be buffered
5763 for later use. Under certain circumstances, a flaw in the logic means that
5764 those messages do not get removed from the buffer even though the handshake
5765 has been completed. An attacker could force up to approx. 15 messages to
5766 remain in the buffer when they are no longer required. These messages will
5767 be cleared when the DTLS connection is closed. The default maximum size for
5768 a message is 100k. Therefore, the attacker could force an additional 1500k
5769 to be consumed per connection. By opening many simultaneous connections an
5770 attacker could cause a DoS attack through memory exhaustion.
5772 This issue was reported to OpenSSL by Quan Luo.
5777 * DTLS replay protection DoS
5779 A flaw in the DTLS replay attack protection mechanism means that records
5780 that arrive for future epochs update the replay protection "window" before
5781 the MAC for the record has been validated. This could be exploited by an
5782 attacker by sending a record for the next epoch (which does not have to
5783 decrypt or have a valid MAC), with a very large sequence number. This means
5784 that all subsequent legitimate packets are dropped causing a denial of
5785 service for a specific DTLS connection.
5787 This issue was reported to OpenSSL by the OCAP audit team.
5792 * Certificate message OOB reads
5794 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5795 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5796 theoretical DoS risk but this has not been observed in practice on common
5799 The messages affected are client certificate, client certificate request
5800 and server certificate. As a result the attack can only be performed
5801 against a client or a server which enables client authentication.
5803 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5808 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5810 * Prevent padding oracle in AES-NI CBC MAC check
5812 A MITM attacker can use a padding oracle attack to decrypt traffic
5813 when the connection uses an AES CBC cipher and the server support
5816 This issue was introduced as part of the fix for Lucky 13 padding
5817 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5818 constant time by making sure that always the same bytes are read and
5819 compared against either the MAC or padding bytes. But it no longer
5820 checked that there was enough data to have both the MAC and padding
5823 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5827 * Fix EVP_EncodeUpdate overflow
5829 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5830 Base64 encoding of binary data. If an attacker is able to supply very large
5831 amounts of input data then a length check can overflow resulting in a heap
5834 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5835 the `PEM_write_bio*` family of functions. These are mainly used within the
5836 OpenSSL command line applications, so any application which processes data
5837 from an untrusted source and outputs it as a PEM file should be considered
5838 vulnerable to this issue. User applications that call these APIs directly
5839 with large amounts of untrusted data may also be vulnerable.
5841 This issue was reported by Guido Vranken.
5846 * Fix EVP_EncryptUpdate overflow
5848 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5849 is able to supply very large amounts of input data after a previous call to
5850 EVP_EncryptUpdate() with a partial block then a length check can overflow
5851 resulting in a heap corruption. Following an analysis of all OpenSSL
5852 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5853 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5854 the first called function after an EVP_EncryptInit(), and therefore that
5855 specific call must be safe. The second form is where the length passed to
5856 EVP_EncryptUpdate() can be seen from the code to be some small value and
5857 therefore there is no possibility of an overflow. Since all instances are
5858 one of these two forms, it is believed that there can be no overflows in
5859 internal code due to this problem. It should be noted that
5860 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5861 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5862 of these calls have also been analysed too and it is believed there are no
5863 instances in internal usage where an overflow could occur.
5865 This issue was reported by Guido Vranken.
5870 * Prevent ASN.1 BIO excessive memory allocation
5872 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5873 a short invalid encoding can cause allocation of large amounts of memory
5874 potentially consuming excessive resources or exhausting memory.
5876 Any application parsing untrusted data through d2i BIO functions is
5877 affected. The memory based functions such as d2i_X509() are *not* affected.
5878 Since the memory based functions are used by the TLS library, TLS
5879 applications are not affected.
5881 This issue was reported by Brian Carpenter.
5888 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5889 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5890 in arbitrary stack data being returned in the buffer.
5892 This issue was reported by Guido Vranken.
5897 * Modify behavior of ALPN to invoke callback after SNI/servername
5898 callback, such that updates to the SSL_CTX affect ALPN.
5902 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5907 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5908 methods are enabled and ssl2 is disabled the methods return NULL.
5912 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5914 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5915 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5916 provide any "EXPORT" or "LOW" strength ciphers.
5920 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5921 is by default disabled at build-time. Builds that are not configured with
5922 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5923 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5924 will need to explicitly call either of:
5926 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5928 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5930 as appropriate. Even if either of those is used, or the application
5931 explicitly uses the version-specific SSLv2_method() or its client and
5932 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5933 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5934 ciphers, and SSLv2 56-bit DES are no longer available.
5939 * Fix a double-free in DSA code
5941 A double free bug was discovered when OpenSSL parses malformed DSA private
5942 keys and could lead to a DoS attack or memory corruption for applications
5943 that receive DSA private keys from untrusted sources. This scenario is
5946 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5952 * Disable SRP fake user seed to address a server memory leak.
5954 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5956 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5957 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5958 was changed to ignore the "fake user" SRP seed, even if the seed
5961 Users should use SRP_VBASE_get1_by_user instead. Note that in
5962 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5963 also that even though configuring the SRP seed attempts to hide
5964 invalid usernames by continuing the handshake with fake
5965 credentials, this behaviour is not constant time and no strong
5966 guarantees are made that the handshake is indistinguishable from
5967 that of a valid user.
5972 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5974 In the BN_hex2bn function the number of hex digits is calculated using an
5975 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5976 large values of `i` this can result in `bn_expand` not allocating any
5977 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5978 field as NULL leading to a subsequent NULL ptr deref. For very large values
5979 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5980 In this case memory is allocated to the internal BIGNUM data field, but it
5981 is insufficiently sized leading to heap corruption. A similar issue exists
5982 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5983 is ever called by user applications with very large untrusted hex/dec data.
5984 This is anticipated to be a rare occurrence.
5986 All OpenSSL internal usage of these functions use data that is not expected
5987 to be untrusted, e.g. config file data or application command line
5988 arguments. If user developed applications generate config file data based
5989 on untrusted data then it is possible that this could also lead to security
5990 consequences. This is also anticipated to be rare.
5992 This issue was reported to OpenSSL by Guido Vranken.
5997 * Fix memory issues in `BIO_*printf` functions
5999 The internal `fmtstr` function used in processing a "%s" format string in
6000 the `BIO_*printf` functions could overflow while calculating the length of a
6001 string and cause an OOB read when printing very long strings.
6003 Additionally the internal `doapr_outch` function can attempt to write to an
6004 OOB memory location (at an offset from the NULL pointer) in the event of a
6005 memory allocation failure. In 1.0.2 and below this could be caused where
6006 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6007 could be in processing a very long "%s" format string. Memory leaks can
6010 The first issue may mask the second issue dependent on compiler behaviour.
6011 These problems could enable attacks where large amounts of untrusted data
6012 is passed to the `BIO_*printf` functions. If applications use these functions
6013 in this way then they could be vulnerable. OpenSSL itself uses these
6014 functions when printing out human-readable dumps of ASN.1 data. Therefore
6015 applications that print this data could be vulnerable if the data is from
6016 untrusted sources. OpenSSL command line applications could also be
6017 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6018 as command line arguments.
6020 Libssl is not considered directly vulnerable. Additionally certificates etc
6021 received via remote connections via libssl are also unlikely to be able to
6022 trigger these issues because of message size limits enforced within libssl.
6024 This issue was reported to OpenSSL Guido Vranken.
6029 * Side channel attack on modular exponentiation
6031 A side-channel attack was found which makes use of cache-bank conflicts on
6032 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6033 of RSA keys. The ability to exploit this issue is limited as it relies on
6034 an attacker who has control of code in a thread running on the same
6035 hyper-threaded core as the victim thread which is performing decryptions.
6037 This issue was reported to OpenSSL by Yuval Yarom, The University of
6038 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6039 Nadia Heninger, University of Pennsylvania with more information at
6040 <http://cachebleed.info>.
6045 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6046 if no keysize is specified with default_bits. This fixes an
6047 omission in an earlier change that changed all RSA/DSA key generation
6048 commands to use 2048 bits by default.
6052 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6054 * DH small subgroups
6056 Historically OpenSSL only ever generated DH parameters based on "safe"
6057 primes. More recently (in version 1.0.2) support was provided for
6058 generating X9.42 style parameter files such as those required for RFC 5114
6059 support. The primes used in such files may not be "safe". Where an
6060 application is using DH configured with parameters based on primes that are
6061 not "safe" then an attacker could use this fact to find a peer's private
6062 DH exponent. This attack requires that the attacker complete multiple
6063 handshakes in which the peer uses the same private DH exponent. For example
6064 this could be used to discover a TLS server's private DH exponent if it's
6065 reusing the private DH exponent or it's using a static DH ciphersuite.
6067 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6068 TLS. It is not on by default. If the option is not set then the server
6069 reuses the same private DH exponent for the life of the server process and
6070 would be vulnerable to this attack. It is believed that many popular
6071 applications do set this option and would therefore not be at risk.
6073 The fix for this issue adds an additional check where a "q" parameter is
6074 available (as is the case in X9.42 based parameters). This detects the
6075 only known attack, and is the only possible defense for static DH
6076 ciphersuites. This could have some performance impact.
6078 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6079 default and cannot be disabled. This could have some performance impact.
6081 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6086 * SSLv2 doesn't block disabled ciphers
6088 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6089 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6090 been disabled, provided that the SSLv2 protocol was not also disabled via
6093 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6094 and Sebastian Schinzel.
6099 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6101 * BN_mod_exp may produce incorrect results on x86_64
6103 There is a carry propagating bug in the x86_64 Montgomery squaring
6104 procedure. No EC algorithms are affected. Analysis suggests that attacks
6105 against RSA and DSA as a result of this defect would be very difficult to
6106 perform and are not believed likely. Attacks against DH are considered just
6107 feasible (although very difficult) because most of the work necessary to
6108 deduce information about a private key may be performed offline. The amount
6109 of resources required for such an attack would be very significant and
6110 likely only accessible to a limited number of attackers. An attacker would
6111 additionally need online access to an unpatched system using the target
6112 private key in a scenario with persistent DH parameters and a private
6113 key that is shared between multiple clients. For example this can occur by
6114 default in OpenSSL DHE based SSL/TLS ciphersuites.
6116 This issue was reported to OpenSSL by Hanno Böck.
6121 * Certificate verify crash with missing PSS parameter
6123 The signature verification routines will crash with a NULL pointer
6124 dereference if presented with an ASN.1 signature using the RSA PSS
6125 algorithm and absent mask generation function parameter. Since these
6126 routines are used to verify certificate signature algorithms this can be
6127 used to crash any certificate verification operation and exploited in a
6128 DoS attack. Any application which performs certificate verification is
6129 vulnerable including OpenSSL clients and servers which enable client
6132 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6137 * X509_ATTRIBUTE memory leak
6139 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6140 memory. This structure is used by the PKCS#7 and CMS routines so any
6141 application which reads PKCS#7 or CMS data from untrusted sources is
6142 affected. SSL/TLS is not affected.
6144 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6150 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6151 This changes the decoding behaviour for some invalid messages,
6152 though the change is mostly in the more lenient direction, and
6153 legacy behaviour is preserved as much as possible.
6157 * In DSA_generate_parameters_ex, if the provided seed is too short,
6160 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6162 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6164 * Alternate chains certificate forgery
6166 During certificate verification, OpenSSL will attempt to find an
6167 alternative certificate chain if the first attempt to build such a chain
6168 fails. An error in the implementation of this logic can mean that an
6169 attacker could cause certain checks on untrusted certificates to be
6170 bypassed, such as the CA flag, enabling them to use a valid leaf
6171 certificate to act as a CA and "issue" an invalid certificate.
6173 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6178 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6180 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6181 incompatibility in the handling of HMAC. The previous ABI has now been
6186 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6188 * Malformed ECParameters causes infinite loop
6190 When processing an ECParameters structure OpenSSL enters an infinite loop
6191 if the curve specified is over a specially malformed binary polynomial
6194 This can be used to perform denial of service against any
6195 system which processes public keys, certificate requests or
6196 certificates. This includes TLS clients and TLS servers with
6197 client authentication enabled.
6199 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6204 * Exploitable out-of-bounds read in X509_cmp_time
6206 X509_cmp_time does not properly check the length of the ASN1_TIME
6207 string and can read a few bytes out of bounds. In addition,
6208 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6211 An attacker can use this to craft malformed certificates and CRLs of
6212 various sizes and potentially cause a segmentation fault, resulting in
6213 a DoS on applications that verify certificates or CRLs. TLS clients
6214 that verify CRLs are affected. TLS clients and servers with client
6215 authentication enabled may be affected if they use custom verification
6218 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6219 independently by Hanno Böck.
6224 * PKCS7 crash with missing EnvelopedContent
6226 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6227 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6228 with missing content and trigger a NULL pointer dereference on parsing.
6230 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6231 structures from untrusted sources are affected. OpenSSL clients and
6232 servers are not affected.
6234 This issue was reported to OpenSSL by Michal Zalewski (Google).
6239 * CMS verify infinite loop with unknown hash function
6241 When verifying a signedData message the CMS code can enter an infinite loop
6242 if presented with an unknown hash function OID. This can be used to perform
6243 denial of service against any system which verifies signedData messages using
6245 This issue was reported to OpenSSL by Johannes Bauer.
6250 * Race condition handling NewSessionTicket
6252 If a NewSessionTicket is received by a multi-threaded client when attempting to
6253 reuse a previous ticket then a race condition can occur potentially leading to
6254 a double free of the ticket data.
6259 * Only support 256-bit or stronger elliptic curves with the
6260 'ecdh_auto' setting (server) or by default (client). Of supported
6261 curves, prefer P-256 (both).
6265 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6267 * ClientHello sigalgs DoS fix
6269 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6270 invalid signature algorithms extension a NULL pointer dereference will
6271 occur. This can be exploited in a DoS attack against the server.
6273 This issue was was reported to OpenSSL by David Ramos of Stanford
6277 *Stephen Henson and Matt Caswell*
6279 * Multiblock corrupted pointer fix
6281 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6282 feature only applies on 64 bit x86 architecture platforms that support AES
6283 NI instructions. A defect in the implementation of "multiblock" can cause
6284 OpenSSL's internal write buffer to become incorrectly set to NULL when
6285 using non-blocking IO. Typically, when the user application is using a
6286 socket BIO for writing, this will only result in a failed connection.
6287 However if some other BIO is used then it is likely that a segmentation
6288 fault will be triggered, thus enabling a potential DoS attack.
6290 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6295 * Segmentation fault in DTLSv1_listen fix
6297 The DTLSv1_listen function is intended to be stateless and processes the
6298 initial ClientHello from many peers. It is common for user code to loop
6299 over the call to DTLSv1_listen until a valid ClientHello is received with
6300 an associated cookie. A defect in the implementation of DTLSv1_listen means
6301 that state is preserved in the SSL object from one invocation to the next
6302 that can lead to a segmentation fault. Errors processing the initial
6303 ClientHello can trigger this scenario. An example of such an error could be
6304 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6307 This issue was reported to OpenSSL by Per Allansson.
6312 * Segmentation fault in ASN1_TYPE_cmp fix
6314 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6315 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6316 certificate signature algorithm consistency this can be used to crash any
6317 certificate verification operation and exploited in a DoS attack. Any
6318 application which performs certificate verification is vulnerable including
6319 OpenSSL clients and servers which enable client authentication.
6324 * Segmentation fault for invalid PSS parameters fix
6326 The signature verification routines will crash with a NULL pointer
6327 dereference if presented with an ASN.1 signature using the RSA PSS
6328 algorithm and invalid parameters. Since these routines are used to verify
6329 certificate signature algorithms this can be used to crash any
6330 certificate verification operation and exploited in a DoS attack. Any
6331 application which performs certificate verification is vulnerable including
6332 OpenSSL clients and servers which enable client authentication.
6334 This issue was was reported to OpenSSL by Brian Carpenter.
6339 * ASN.1 structure reuse memory corruption fix
6341 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6342 memory corruption via an invalid write. Such reuse is and has been
6343 strongly discouraged and is believed to be rare.
6345 Applications that parse structures containing CHOICE or ANY DEFINED BY
6346 components may be affected. Certificate parsing (d2i_X509 and related
6347 functions) are however not affected. OpenSSL clients and servers are
6353 * PKCS7 NULL pointer dereferences fix
6355 The PKCS#7 parsing code does not handle missing outer ContentInfo
6356 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6357 missing content and trigger a NULL pointer dereference on parsing.
6359 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6360 otherwise parse PKCS#7 structures from untrusted sources are
6361 affected. OpenSSL clients and servers are not affected.
6363 This issue was reported to OpenSSL by Michal Zalewski (Google).
6368 * DoS via reachable assert in SSLv2 servers fix
6370 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6371 servers that both support SSLv2 and enable export cipher suites by sending
6372 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6374 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6375 (OpenSSL development team).
6380 * Empty CKE with client auth and DHE fix
6382 If client auth is used then a server can seg fault in the event of a DHE
6383 ciphersuite being selected and a zero length ClientKeyExchange message
6384 being sent by the client. This could be exploited in a DoS attack.
6389 * Handshake with unseeded PRNG fix
6391 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6392 with an unseeded PRNG. The conditions are:
6393 - The client is on a platform where the PRNG has not been seeded
6394 automatically, and the user has not seeded manually
6395 - A protocol specific client method version has been used (i.e. not
6396 SSL_client_methodv23)
6397 - A ciphersuite is used that does not require additional random data from
6398 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6400 If the handshake succeeds then the client random that has been used will
6401 have been generated from a PRNG with insufficient entropy and therefore the
6402 output may be predictable.
6404 For example using the following command with an unseeded openssl will
6405 succeed on an unpatched platform:
6407 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6412 * Use After Free following d2i_ECPrivatekey error fix
6414 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6415 could cause a use after free condition. This, in turn, could cause a double
6416 free in several private key parsing functions (such as d2i_PrivateKey
6417 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6418 for applications that receive EC private keys from untrusted
6419 sources. This scenario is considered rare.
6421 This issue was discovered by the BoringSSL project and fixed in their
6427 * X509_to_X509_REQ NULL pointer deref fix
6429 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6430 the certificate key is invalid. This function is rarely used in practice.
6432 This issue was discovered by Brian Carpenter.
6437 * Removed the export ciphers from the DEFAULT ciphers
6441 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6443 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6444 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6445 So far those who have to target multiple platforms would compromise
6446 and argue that binary targeting say ARMv5 would still execute on
6447 ARMv8. "Universal" build resolves this compromise by providing
6448 near-optimal performance even on newer platforms.
6452 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6453 (other platforms pending).
6455 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6457 * Add support for the SignedCertificateTimestampList certificate and
6458 OCSP response extensions from RFC6962.
6462 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6463 for corner cases. (Certain input points at infinity could lead to
6464 bogus results, with non-infinity inputs mapped to infinity too.)
6468 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6469 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6470 common cases are optimized and there still is room for further
6471 improvements. Vector Permutation AES for Altivec is also added.
6475 * Add support for little-endian ppc64 Linux target.
6477 *Marcelo Cerri (IBM)*
6479 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6480 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6481 are optimized and there still is room for further improvements.
6482 Both 32- and 64-bit modes are supported.
6484 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6486 * Improved ARMv7 NEON support.
6490 * Support for SPARC Architecture 2011 crypto extensions, first
6491 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6492 SHA256/512, MD5, GHASH and modular exponentiation.
6494 *Andy Polyakov, David Miller*
6496 * Accelerated modular exponentiation for Intel processors, a.k.a.
6499 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6501 * Support for new and upcoming Intel processors, including AVX2,
6502 BMI and SHA ISA extensions. This includes additional "stitched"
6503 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6506 This work was sponsored by Intel Corp.
6510 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6511 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6512 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6516 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6517 this fixes a limitation in previous versions of OpenSSL.
6521 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6522 MGF1 digest and OAEP label.
6526 * Add EVP support for key wrapping algorithms, to avoid problems with
6527 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6528 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6529 algorithms and include tests cases.
6533 * Add functions to allocate and set the fields of an ECDSA_METHOD
6536 *Douglas E. Engert, Steve Henson*
6538 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6539 difference in days and seconds between two tm or ASN1_TIME structures.
6543 * Add -rev test option to s_server to just reverse order of characters
6544 received by client and send back to server. Also prints an abbreviated
6545 summary of the connection parameters.
6549 * New option -brief for s_client and s_server to print out a brief summary
6550 of connection parameters.
6554 * Add callbacks for arbitrary TLS extensions.
6556 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6558 * New option -crl_download in several openssl utilities to download CRLs
6559 from CRLDP extension in certificates.
6563 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6567 * New function X509_CRL_diff to generate a delta CRL from the difference
6568 of two full CRLs. Add support to "crl" utility.
6572 * New functions to set lookup_crls function and to retrieve
6573 X509_STORE from X509_STORE_CTX.
6577 * Print out deprecated issuer and subject unique ID fields in
6582 * Extend OCSP I/O functions so they can be used for simple general purpose
6583 HTTP as well as OCSP. New wrapper function which can be used to download
6584 CRLs using the OCSP API.
6588 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6592 * `SSL_CONF*` functions. These provide a common framework for application
6593 configuration using configuration files or command lines.
6597 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6598 message callback and prints the results. Needs compile time option
6599 "enable-ssl-trace". New options to s_client and s_server to enable
6604 * New ctrl and macro to retrieve supported points extensions.
6605 Print out extension in s_server and s_client.
6609 * New functions to retrieve certificate signature and signature
6614 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6619 * New Suite B modes for TLS code. These use and enforce the requirements
6620 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6621 only use Suite B curves. The Suite B modes can be set by using the
6622 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6626 * New chain verification flags for Suite B levels of security. Check
6627 algorithms are acceptable when flags are set in X509_verify_cert.
6631 * Make tls1_check_chain return a set of flags indicating checks passed
6632 by a certificate chain. Add additional tests to handle client
6633 certificates: checks for matching certificate type and issuer name
6638 * If an attempt is made to use a signature algorithm not in the peer
6639 preference list abort the handshake. If client has no suitable
6640 signature algorithms in response to a certificate request do not
6641 use the certificate.
6645 * If server EC tmp key is not in client preference list abort handshake.
6649 * Add support for certificate stores in CERT structure. This makes it
6650 possible to have different stores per SSL structure or one store in
6651 the parent SSL_CTX. Include distinct stores for certificate chain
6652 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6653 to build and store a certificate chain in CERT structure: returning
6654 an error if the chain cannot be built: this will allow applications
6655 to test if a chain is correctly configured.
6657 Note: if the CERT based stores are not set then the parent SSL_CTX
6658 store is used to retain compatibility with existing behaviour.
6662 * New function ssl_set_client_disabled to set a ciphersuite disabled
6663 mask based on the current session, check mask when sending client
6664 hello and checking the requested ciphersuite.
6668 * New ctrls to retrieve and set certificate types in a certificate
6669 request message. Print out received values in s_client. If certificate
6670 types is not set with custom values set sensible values based on
6671 supported signature algorithms.
6675 * Support for distinct client and server supported signature algorithms.
6679 * Add certificate callback. If set this is called whenever a certificate
6680 is required by client or server. An application can decide which
6681 certificate chain to present based on arbitrary criteria: for example
6682 supported signature algorithms. Add very simple example to s_server.
6683 This fixes many of the problems and restrictions of the existing client
6684 certificate callback: for example you can now clear an existing
6685 certificate and specify the whole chain.
6689 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6690 the certificate can be used for (if anything). Set valid_flags field
6691 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6692 to have similar checks in it.
6694 Add new "cert_flags" field to CERT structure and include a "strict mode".
6695 This enforces some TLS certificate requirements (such as only permitting
6696 certificate signature algorithms contained in the supported algorithms
6697 extension) which some implementations ignore: this option should be used
6698 with caution as it could cause interoperability issues.
6702 * Update and tidy signature algorithm extension processing. Work out
6703 shared signature algorithms based on preferences and peer algorithms
6704 and print them out in s_client and s_server. Abort handshake if no
6705 shared signature algorithms.
6709 * Add new functions to allow customised supported signature algorithms
6710 for SSL and SSL_CTX structures. Add options to s_client and s_server
6715 * New function SSL_certs_clear() to delete all references to certificates
6716 from an SSL structure. Before this once a certificate had been added
6717 it couldn't be removed.
6721 * Integrate hostname, email address and IP address checking with certificate
6722 verification. New verify options supporting checking in openssl utility.
6726 * Fixes and wildcard matching support to hostname and email checking
6727 functions. Add manual page.
6729 *Florian Weimer (Red Hat Product Security Team)*
6731 * New functions to check a hostname email or IP address against a
6732 certificate. Add options x509 utility to print results of checks against
6737 * Fix OCSP checking.
6739 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6741 * Initial experimental support for explicitly trusted non-root CAs.
6742 OpenSSL still tries to build a complete chain to a root but if an
6743 intermediate CA has a trust setting included that is used. The first
6744 setting is used: whether to trust (e.g., -addtrust option to the x509
6749 * Add -trusted_first option which attempts to find certificates in the
6750 trusted store even if an untrusted chain is also supplied.
6754 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6755 platform support for Linux and Android.
6759 * Support for linux-x32, ILP32 environment in x86_64 framework.
6763 * Experimental multi-implementation support for FIPS capable OpenSSL.
6764 When in FIPS mode the approved implementations are used as normal,
6765 when not in FIPS mode the internal unapproved versions are used instead.
6766 This means that the FIPS capable OpenSSL isn't forced to use the
6767 (often lower performance) FIPS implementations outside FIPS mode.
6771 * Transparently support X9.42 DH parameters when calling
6772 PEM_read_bio_DHparameters. This means existing applications can handle
6773 the new parameter format automatically.
6777 * Initial experimental support for X9.42 DH parameter format: mainly
6778 to support use of 'q' parameter for RFC5114 parameters.
6782 * Add DH parameters from RFC5114 including test data to dhtest.
6786 * Support for automatic EC temporary key parameter selection. If enabled
6787 the most preferred EC parameters are automatically used instead of
6788 hardcoded fixed parameters. Now a server just has to call:
6789 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6790 support ECDH and use the most appropriate parameters.
6794 * Enhance and tidy EC curve and point format TLS extension code. Use
6795 static structures instead of allocation if default values are used.
6796 New ctrls to set curves we wish to support and to retrieve shared curves.
6797 Print out shared curves in s_server. New options to s_server and s_client
6798 to set list of supported curves.
6802 * New ctrls to retrieve supported signature algorithms and
6803 supported curve values as an array of NIDs. Extend openssl utility
6804 to print out received values.
6808 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6809 between NIDs and the more common NIST names such as "P-256". Enhance
6810 ecparam utility and ECC method to recognise the NIST names for curves.
6814 * Enhance SSL/TLS certificate chain handling to support different
6815 chains for each certificate instead of one chain in the parent SSL_CTX.
6819 * Support for fixed DH ciphersuite client authentication: where both
6820 server and client use DH certificates with common parameters.
6824 * Support for fixed DH ciphersuites: those requiring DH server
6829 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6831 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6832 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6833 X509_CINF_get_signature were reverted post internal team review.
6838 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6840 * OCSP Status Request extension unbounded memory growth
6842 A malicious client can send an excessively large OCSP Status Request
6843 extension. If that client continually requests renegotiation, sending a
6844 large OCSP Status Request extension each time, then there will be unbounded
6845 memory growth on the server. This will eventually lead to a Denial Of
6846 Service attack through memory exhaustion. Servers with a default
6847 configuration are vulnerable even if they do not support OCSP. Builds using
6848 the "no-ocsp" build time option are not affected.
6850 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6855 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6858 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6864 * OOB write in MDC2_Update()
6866 An overflow can occur in MDC2_Update() either if called directly or
6867 through the EVP_DigestUpdate() function using MDC2. If an attacker
6868 is able to supply very large amounts of input data after a previous
6869 call to EVP_EncryptUpdate() with a partial block then a length check
6870 can overflow resulting in a heap corruption.
6872 The amount of data needed is comparable to SIZE_MAX which is impractical
6875 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6880 * Malformed SHA512 ticket DoS
6882 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6883 DoS attack where a malformed ticket will result in an OOB read which will
6886 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6887 a custom server callback and ticket lookup mechanism.
6889 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6894 * OOB write in BN_bn2dec()
6896 The function BN_bn2dec() does not check the return value of BN_div_word().
6897 This can cause an OOB write if an application uses this function with an
6898 overly large BIGNUM. This could be a problem if an overly large certificate
6899 or CRL is printed out from an untrusted source. TLS is not affected because
6900 record limits will reject an oversized certificate before it is parsed.
6902 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6907 * OOB read in TS_OBJ_print_bio()
6909 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6910 the total length the OID text representation would use and not the amount
6911 of data written. This will result in OOB reads when large OIDs are
6914 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6919 * Pointer arithmetic undefined behaviour
6921 Avoid some undefined pointer arithmetic
6923 A common idiom in the codebase is to check limits in the following manner:
6926 Where "p" points to some malloc'd data of SIZE bytes and
6929 "len" here could be from some externally supplied data (e.g. from a TLS
6932 The rules of C pointer arithmetic are such that "p + len" is only well
6933 defined where len <= SIZE. Therefore, the above idiom is actually
6934 undefined behaviour.
6936 For example this could cause problems if some malloc implementation
6937 provides an address for "p" such that "p + len" actually overflows for
6938 values of len that are too big and therefore p + len < limit.
6940 This issue was reported to OpenSSL by Guido Vranken
6945 * Constant time flag not preserved in DSA signing
6947 Operations in the DSA signing algorithm should run in constant time in
6948 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6949 implementation means that a non-constant time codepath is followed for
6950 certain operations. This has been demonstrated through a cache-timing
6951 attack to be sufficient for an attacker to recover the private DSA key.
6953 This issue was reported by César Pereida (Aalto University), Billy Brumley
6954 (Tampere University of Technology), and Yuval Yarom (The University of
6955 Adelaide and NICTA).
6960 * DTLS buffered message DoS
6962 In a DTLS connection where handshake messages are delivered out-of-order
6963 those messages that OpenSSL is not yet ready to process will be buffered
6964 for later use. Under certain circumstances, a flaw in the logic means that
6965 those messages do not get removed from the buffer even though the handshake
6966 has been completed. An attacker could force up to approx. 15 messages to
6967 remain in the buffer when they are no longer required. These messages will
6968 be cleared when the DTLS connection is closed. The default maximum size for
6969 a message is 100k. Therefore, the attacker could force an additional 1500k
6970 to be consumed per connection. By opening many simultaneous connections an
6971 attacker could cause a DoS attack through memory exhaustion.
6973 This issue was reported to OpenSSL by Quan Luo.
6978 * DTLS replay protection DoS
6980 A flaw in the DTLS replay attack protection mechanism means that records
6981 that arrive for future epochs update the replay protection "window" before
6982 the MAC for the record has been validated. This could be exploited by an
6983 attacker by sending a record for the next epoch (which does not have to
6984 decrypt or have a valid MAC), with a very large sequence number. This means
6985 that all subsequent legitimate packets are dropped causing a denial of
6986 service for a specific DTLS connection.
6988 This issue was reported to OpenSSL by the OCAP audit team.
6993 * Certificate message OOB reads
6995 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6996 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6997 theoretical DoS risk but this has not been observed in practice on common
7000 The messages affected are client certificate, client certificate request
7001 and server certificate. As a result the attack can only be performed
7002 against a client or a server which enables client authentication.
7004 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7009 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7011 * Prevent padding oracle in AES-NI CBC MAC check
7013 A MITM attacker can use a padding oracle attack to decrypt traffic
7014 when the connection uses an AES CBC cipher and the server support
7017 This issue was introduced as part of the fix for Lucky 13 padding
7018 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7019 constant time by making sure that always the same bytes are read and
7020 compared against either the MAC or padding bytes. But it no longer
7021 checked that there was enough data to have both the MAC and padding
7024 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7029 * Fix EVP_EncodeUpdate overflow
7031 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7032 Base64 encoding of binary data. If an attacker is able to supply very large
7033 amounts of input data then a length check can overflow resulting in a heap
7036 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7037 the `PEM_write_bio*` family of functions. These are mainly used within the
7038 OpenSSL command line applications, so any application which processes data
7039 from an untrusted source and outputs it as a PEM file should be considered
7040 vulnerable to this issue. User applications that call these APIs directly
7041 with large amounts of untrusted data may also be vulnerable.
7043 This issue was reported by Guido Vranken.
7048 * Fix EVP_EncryptUpdate overflow
7050 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7051 is able to supply very large amounts of input data after a previous call to
7052 EVP_EncryptUpdate() with a partial block then a length check can overflow
7053 resulting in a heap corruption. Following an analysis of all OpenSSL
7054 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7055 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7056 the first called function after an EVP_EncryptInit(), and therefore that
7057 specific call must be safe. The second form is where the length passed to
7058 EVP_EncryptUpdate() can be seen from the code to be some small value and
7059 therefore there is no possibility of an overflow. Since all instances are
7060 one of these two forms, it is believed that there can be no overflows in
7061 internal code due to this problem. It should be noted that
7062 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7063 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7064 of these calls have also been analysed too and it is believed there are no
7065 instances in internal usage where an overflow could occur.
7067 This issue was reported by Guido Vranken.
7072 * Prevent ASN.1 BIO excessive memory allocation
7074 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7075 a short invalid encoding can casuse allocation of large amounts of memory
7076 potentially consuming excessive resources or exhausting memory.
7078 Any application parsing untrusted data through d2i BIO functions is
7079 affected. The memory based functions such as d2i_X509() are *not* affected.
7080 Since the memory based functions are used by the TLS library, TLS
7081 applications are not affected.
7083 This issue was reported by Brian Carpenter.
7090 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7091 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7092 in arbitrary stack data being returned in the buffer.
7094 This issue was reported by Guido Vranken.
7099 * Modify behavior of ALPN to invoke callback after SNI/servername
7100 callback, such that updates to the SSL_CTX affect ALPN.
7104 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7109 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7110 methods are enabled and ssl2 is disabled the methods return NULL.
7114 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7116 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7117 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7118 provide any "EXPORT" or "LOW" strength ciphers.
7122 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7123 is by default disabled at build-time. Builds that are not configured with
7124 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7125 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7126 will need to explicitly call either of:
7128 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7130 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7132 as appropriate. Even if either of those is used, or the application
7133 explicitly uses the version-specific SSLv2_method() or its client and
7134 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7135 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7136 ciphers, and SSLv2 56-bit DES are no longer available.
7141 * Fix a double-free in DSA code
7143 A double free bug was discovered when OpenSSL parses malformed DSA private
7144 keys and could lead to a DoS attack or memory corruption for applications
7145 that receive DSA private keys from untrusted sources. This scenario is
7148 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7154 * Disable SRP fake user seed to address a server memory leak.
7156 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7158 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7159 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7160 was changed to ignore the "fake user" SRP seed, even if the seed
7163 Users should use SRP_VBASE_get1_by_user instead. Note that in
7164 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7165 also that even though configuring the SRP seed attempts to hide
7166 invalid usernames by continuing the handshake with fake
7167 credentials, this behaviour is not constant time and no strong
7168 guarantees are made that the handshake is indistinguishable from
7169 that of a valid user.
7174 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7176 In the BN_hex2bn function the number of hex digits is calculated using an
7177 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7178 large values of `i` this can result in `bn_expand` not allocating any
7179 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7180 field as NULL leading to a subsequent NULL ptr deref. For very large values
7181 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7182 In this case memory is allocated to the internal BIGNUM data field, but it
7183 is insufficiently sized leading to heap corruption. A similar issue exists
7184 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7185 is ever called by user applications with very large untrusted hex/dec data.
7186 This is anticipated to be a rare occurrence.
7188 All OpenSSL internal usage of these functions use data that is not expected
7189 to be untrusted, e.g. config file data or application command line
7190 arguments. If user developed applications generate config file data based
7191 on untrusted data then it is possible that this could also lead to security
7192 consequences. This is also anticipated to be rare.
7194 This issue was reported to OpenSSL by Guido Vranken.
7199 * Fix memory issues in `BIO_*printf` functions
7201 The internal `fmtstr` function used in processing a "%s" format string in
7202 the `BIO_*printf` functions could overflow while calculating the length of a
7203 string and cause an OOB read when printing very long strings.
7205 Additionally the internal `doapr_outch` function can attempt to write to an
7206 OOB memory location (at an offset from the NULL pointer) in the event of a
7207 memory allocation failure. In 1.0.2 and below this could be caused where
7208 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7209 could be in processing a very long "%s" format string. Memory leaks can
7212 The first issue may mask the second issue dependent on compiler behaviour.
7213 These problems could enable attacks where large amounts of untrusted data
7214 is passed to the `BIO_*printf` functions. If applications use these functions
7215 in this way then they could be vulnerable. OpenSSL itself uses these
7216 functions when printing out human-readable dumps of ASN.1 data. Therefore
7217 applications that print this data could be vulnerable if the data is from
7218 untrusted sources. OpenSSL command line applications could also be
7219 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7220 as command line arguments.
7222 Libssl is not considered directly vulnerable. Additionally certificates etc
7223 received via remote connections via libssl are also unlikely to be able to
7224 trigger these issues because of message size limits enforced within libssl.
7226 This issue was reported to OpenSSL Guido Vranken.
7231 * Side channel attack on modular exponentiation
7233 A side-channel attack was found which makes use of cache-bank conflicts on
7234 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7235 of RSA keys. The ability to exploit this issue is limited as it relies on
7236 an attacker who has control of code in a thread running on the same
7237 hyper-threaded core as the victim thread which is performing decryptions.
7239 This issue was reported to OpenSSL by Yuval Yarom, The University of
7240 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7241 Nadia Heninger, University of Pennsylvania with more information at
7242 <http://cachebleed.info>.
7247 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7248 if no keysize is specified with default_bits. This fixes an
7249 omission in an earlier change that changed all RSA/DSA key generation
7250 commands to use 2048 bits by default.
7254 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7256 * Protection for DH small subgroup attacks
7258 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7259 switched on by default and cannot be disabled. This could have some
7264 * SSLv2 doesn't block disabled ciphers
7266 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7267 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7268 been disabled, provided that the SSLv2 protocol was not also disabled via
7271 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7272 and Sebastian Schinzel.
7277 * Reject DH handshakes with parameters shorter than 1024 bits.
7281 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7283 * Certificate verify crash with missing PSS parameter
7285 The signature verification routines will crash with a NULL pointer
7286 dereference if presented with an ASN.1 signature using the RSA PSS
7287 algorithm and absent mask generation function parameter. Since these
7288 routines are used to verify certificate signature algorithms this can be
7289 used to crash any certificate verification operation and exploited in a
7290 DoS attack. Any application which performs certificate verification is
7291 vulnerable including OpenSSL clients and servers which enable client
7294 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7299 * X509_ATTRIBUTE memory leak
7301 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7302 memory. This structure is used by the PKCS#7 and CMS routines so any
7303 application which reads PKCS#7 or CMS data from untrusted sources is
7304 affected. SSL/TLS is not affected.
7306 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7312 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7313 This changes the decoding behaviour for some invalid messages,
7314 though the change is mostly in the more lenient direction, and
7315 legacy behaviour is preserved as much as possible.
7319 * In DSA_generate_parameters_ex, if the provided seed is too short,
7320 use a random seed, as already documented.
7322 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7324 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7326 * Alternate chains certificate forgery
7328 During certificate verfification, OpenSSL will attempt to find an
7329 alternative certificate chain if the first attempt to build such a chain
7330 fails. An error in the implementation of this logic can mean that an
7331 attacker could cause certain checks on untrusted certificates to be
7332 bypassed, such as the CA flag, enabling them to use a valid leaf
7333 certificate to act as a CA and "issue" an invalid certificate.
7335 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7341 * Race condition handling PSK identify hint
7343 If PSK identity hints are received by a multi-threaded client then
7344 the values are wrongly updated in the parent SSL_CTX structure. This can
7345 result in a race condition potentially leading to a double free of the
7351 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7353 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7354 incompatibility in the handling of HMAC. The previous ABI has now been
7357 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7359 * Malformed ECParameters causes infinite loop
7361 When processing an ECParameters structure OpenSSL enters an infinite loop
7362 if the curve specified is over a specially malformed binary polynomial
7365 This can be used to perform denial of service against any
7366 system which processes public keys, certificate requests or
7367 certificates. This includes TLS clients and TLS servers with
7368 client authentication enabled.
7370 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7375 * Exploitable out-of-bounds read in X509_cmp_time
7377 X509_cmp_time does not properly check the length of the ASN1_TIME
7378 string and can read a few bytes out of bounds. In addition,
7379 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7382 An attacker can use this to craft malformed certificates and CRLs of
7383 various sizes and potentially cause a segmentation fault, resulting in
7384 a DoS on applications that verify certificates or CRLs. TLS clients
7385 that verify CRLs are affected. TLS clients and servers with client
7386 authentication enabled may be affected if they use custom verification
7389 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7390 independently by Hanno Böck.
7395 * PKCS7 crash with missing EnvelopedContent
7397 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7398 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7399 with missing content and trigger a NULL pointer dereference on parsing.
7401 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7402 structures from untrusted sources are affected. OpenSSL clients and
7403 servers are not affected.
7405 This issue was reported to OpenSSL by Michal Zalewski (Google).
7410 * CMS verify infinite loop with unknown hash function
7412 When verifying a signedData message the CMS code can enter an infinite loop
7413 if presented with an unknown hash function OID. This can be used to perform
7414 denial of service against any system which verifies signedData messages using
7416 This issue was reported to OpenSSL by Johannes Bauer.
7421 * Race condition handling NewSessionTicket
7423 If a NewSessionTicket is received by a multi-threaded client when attempting to
7424 reuse a previous ticket then a race condition can occur potentially leading to
7425 a double free of the ticket data.
7430 * Reject DH handshakes with parameters shorter than 768 bits.
7432 *Kurt Roeckx and Emilia Kasper*
7434 * dhparam: generate 2048-bit parameters by default.
7436 *Kurt Roeckx and Emilia Kasper*
7438 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7440 * Segmentation fault in ASN1_TYPE_cmp fix
7442 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7443 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7444 certificate signature algorithm consistency this can be used to crash any
7445 certificate verification operation and exploited in a DoS attack. Any
7446 application which performs certificate verification is vulnerable including
7447 OpenSSL clients and servers which enable client authentication.
7452 * ASN.1 structure reuse memory corruption fix
7454 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7455 memory corruption via an invalid write. Such reuse is and has been
7456 strongly discouraged and is believed to be rare.
7458 Applications that parse structures containing CHOICE or ANY DEFINED BY
7459 components may be affected. Certificate parsing (d2i_X509 and related
7460 functions) are however not affected. OpenSSL clients and servers are
7466 * PKCS7 NULL pointer dereferences fix
7468 The PKCS#7 parsing code does not handle missing outer ContentInfo
7469 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7470 missing content and trigger a NULL pointer dereference on parsing.
7472 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7473 otherwise parse PKCS#7 structures from untrusted sources are
7474 affected. OpenSSL clients and servers are not affected.
7476 This issue was reported to OpenSSL by Michal Zalewski (Google).
7481 * DoS via reachable assert in SSLv2 servers fix
7483 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7484 servers that both support SSLv2 and enable export cipher suites by sending
7485 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7487 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7488 (OpenSSL development team).
7493 * Use After Free following d2i_ECPrivatekey error fix
7495 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7496 could cause a use after free condition. This, in turn, could cause a double
7497 free in several private key parsing functions (such as d2i_PrivateKey
7498 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7499 for applications that receive EC private keys from untrusted
7500 sources. This scenario is considered rare.
7502 This issue was discovered by the BoringSSL project and fixed in their
7508 * X509_to_X509_REQ NULL pointer deref fix
7510 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7511 the certificate key is invalid. This function is rarely used in practice.
7513 This issue was discovered by Brian Carpenter.
7518 * Removed the export ciphers from the DEFAULT ciphers
7522 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7524 * Build fixes for the Windows and OpenVMS platforms
7526 *Matt Caswell and Richard Levitte*
7528 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7530 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7531 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7532 dereference. This could lead to a Denial Of Service attack. Thanks to
7533 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7538 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7539 dtls1_buffer_record function under certain conditions. In particular this
7540 could occur if an attacker sent repeated DTLS records with the same
7541 sequence number but for the next epoch. The memory leak could be exploited
7542 by an attacker in a Denial of Service attack through memory exhaustion.
7543 Thanks to Chris Mueller for reporting this issue.
7548 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7549 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7550 method would be set to NULL which could later result in a NULL pointer
7551 dereference. Thanks to Frank Schmirler for reporting this issue.
7556 * Abort handshake if server key exchange message is omitted for ephemeral
7559 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7560 reporting this issue.
7565 * Remove non-export ephemeral RSA code on client and server. This code
7566 violated the TLS standard by allowing the use of temporary RSA keys in
7567 non-export ciphersuites and could be used by a server to effectively
7568 downgrade the RSA key length used to a value smaller than the server
7569 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7570 INRIA or reporting this issue.
7575 * Fixed issue where DH client certificates are accepted without verification.
7576 An OpenSSL server will accept a DH certificate for client authentication
7577 without the certificate verify message. This effectively allows a client to
7578 authenticate without the use of a private key. This only affects servers
7579 which trust a client certificate authority which issues certificates
7580 containing DH keys: these are extremely rare and hardly ever encountered.
7581 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7587 * Ensure that the session ID context of an SSL is updated when its
7588 SSL_CTX is updated via SSL_set_SSL_CTX.
7590 The session ID context is typically set from the parent SSL_CTX,
7591 and can vary with the CTX.
7595 * Fix various certificate fingerprint issues.
7597 By using non-DER or invalid encodings outside the signed portion of a
7598 certificate the fingerprint can be changed without breaking the signature.
7599 Although no details of the signed portion of the certificate can be changed
7600 this can cause problems with some applications: e.g. those using the
7601 certificate fingerprint for blacklists.
7603 1. Reject signatures with non zero unused bits.
7605 If the BIT STRING containing the signature has non zero unused bits reject
7606 the signature. All current signature algorithms require zero unused bits.
7608 2. Check certificate algorithm consistency.
7610 Check the AlgorithmIdentifier inside TBS matches the one in the
7611 certificate signature. NB: this will result in signature failure
7612 errors for some broken certificates.
7614 Thanks to Konrad Kraszewski from Google for reporting this issue.
7616 3. Check DSA/ECDSA signatures use DER.
7618 Re-encode DSA/ECDSA signatures and compare with the original received
7619 signature. Return an error if there is a mismatch.
7621 This will reject various cases including garbage after signature
7622 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7623 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7624 (negative or with leading zeroes).
7626 Further analysis was conducted and fixes were developed by Stephen Henson
7627 of the OpenSSL core team.
7633 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7634 results on some platforms, including x86_64. This bug occurs at random
7635 with a very low probability, and is not known to be exploitable in any
7636 way, though its exact impact is difficult to determine. Thanks to Pieter
7637 Wuille (Blockstream) who reported this issue and also suggested an initial
7638 fix. Further analysis was conducted by the OpenSSL development team and
7639 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7640 the OpenSSL core team.
7645 * Do not resume sessions on the server if the negotiated protocol
7646 version does not match the session's version. Resuming with a different
7647 version, while not strictly forbidden by the RFC, is of questionable
7648 sanity and breaks all known clients.
7650 *David Benjamin, Emilia Käsper*
7652 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7653 early CCS messages during renegotiation. (Note that because
7654 renegotiation is encrypted, this early CCS was not exploitable.)
7658 * Tighten client-side session ticket handling during renegotiation:
7659 ensure that the client only accepts a session ticket if the server sends
7660 the extension anew in the ServerHello. Previously, a TLS client would
7661 reuse the old extension state and thus accept a session ticket if one was
7662 announced in the initial ServerHello.
7664 Similarly, ensure that the client requires a session ticket if one
7665 was advertised in the ServerHello. Previously, a TLS client would
7666 ignore a missing NewSessionTicket message.
7670 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7674 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7675 sends a carefully crafted handshake message, to cause OpenSSL to fail
7676 to free up to 64k of memory causing a memory leak. This could be
7677 exploited in a Denial Of Service attack. This issue affects OpenSSL
7678 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7679 whether SRTP is used or configured. Implementations of OpenSSL that
7680 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7682 The fix was developed by the OpenSSL team.
7687 * Session Ticket Memory Leak.
7689 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7690 integrity of that ticket is first verified. In the event of a session
7691 ticket integrity check failing, OpenSSL will fail to free memory
7692 causing a memory leak. By sending a large number of invalid session
7693 tickets an attacker could exploit this issue in a Denial Of Service
7699 * Build option no-ssl3 is incomplete.
7701 When OpenSSL is configured with "no-ssl3" as a build option, servers
7702 could accept and complete an SSL 3.0 handshake, and clients could be
7703 configured to send them.
7706 *Akamai and the OpenSSL team*
7708 * Add support for TLS_FALLBACK_SCSV.
7709 Client applications doing fallback retries should call
7710 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7713 *Adam Langley, Bodo Moeller*
7715 * Add additional DigestInfo checks.
7717 Re-encode DigestInto in DER and check against the original when
7718 verifying RSA signature: this will reject any improperly encoded
7719 DigestInfo structures.
7721 Note: this is a precautionary measure and no attacks are currently known.
7725 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7727 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7728 SRP code can be overrun an internal buffer. Add sanity check that
7729 g, A, B < N to SRP code.
7731 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7732 Group for discovering this issue.
7737 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7738 TLS 1.0 instead of higher protocol versions when the ClientHello message
7739 is badly fragmented. This allows a man-in-the-middle attacker to force a
7740 downgrade to TLS 1.0 even if both the server and the client support a
7741 higher protocol version, by modifying the client's TLS records.
7743 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7744 researching this issue.
7749 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7750 to a denial of service attack. A malicious server can crash the client
7751 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7752 ciphersuite and sending carefully crafted handshake messages.
7754 Thanks to Felix Gröbert (Google) for discovering and researching this
7760 * By sending carefully crafted DTLS packets an attacker could cause openssl
7761 to leak memory. This can be exploited through a Denial of Service attack.
7762 Thanks to Adam Langley for discovering and researching this issue.
7767 * An attacker can force openssl to consume large amounts of memory whilst
7768 processing DTLS handshake messages. This can be exploited through a
7769 Denial of Service attack.
7770 Thanks to Adam Langley for discovering and researching this issue.
7775 * An attacker can force an error condition which causes openssl to crash
7776 whilst processing DTLS packets due to memory being freed twice. This
7777 can be exploited through a Denial of Service attack.
7778 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7784 * If a multithreaded client connects to a malicious server using a resumed
7785 session and the server sends an ec point format extension it could write
7786 up to 255 bytes to freed memory.
7788 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7794 * A malicious server can crash an OpenSSL client with a null pointer
7795 dereference (read) by specifying an SRP ciphersuite even though it was not
7796 properly negotiated with the client. This can be exploited through a
7797 Denial of Service attack.
7799 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7800 discovering and researching this issue.
7805 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7806 X509_name_oneline, X509_name_print_ex et al. to leak some information
7807 from the stack. Applications may be affected if they echo pretty printing
7808 output to the attacker.
7810 Thanks to Ivan Fratric (Google) for discovering this issue.
7813 *Emilia Käsper, and Steve Henson*
7815 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7816 for corner cases. (Certain input points at infinity could lead to
7817 bogus results, with non-infinity inputs mapped to infinity too.)
7821 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7823 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7824 handshake can force the use of weak keying material in OpenSSL
7825 SSL/TLS clients and servers.
7827 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7828 researching this issue. ([CVE-2014-0224])
7830 *KIKUCHI Masashi, Steve Henson*
7832 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7833 OpenSSL DTLS client the code can be made to recurse eventually crashing
7836 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7839 *Imre Rad, Steve Henson*
7841 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7842 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7843 client or server. This is potentially exploitable to run arbitrary
7844 code on a vulnerable client or server.
7846 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7848 *Jüri Aedla, Steve Henson*
7850 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7851 are subject to a denial of service attack.
7853 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7854 this issue. ([CVE-2014-3470])
7856 *Felix Gröbert, Ivan Fratric, Steve Henson*
7858 * Harmonize version and its documentation. -f flag is used to display
7861 *mancha <mancha1@zoho.com>*
7863 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7864 in i2d_ECPrivateKey.
7866 *mancha <mancha1@zoho.com>*
7868 * Fix some double frees. These are not thought to be exploitable.
7870 *mancha <mancha1@zoho.com>*
7872 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7874 * A missing bounds check in the handling of the TLS heartbeat extension
7875 can be used to reveal up to 64k of memory to a connected client or
7878 Thanks for Neel Mehta of Google Security for discovering this bug and to
7879 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7880 preparing the fix ([CVE-2014-0160])
7882 *Adam Langley, Bodo Moeller*
7884 * Fix for the attack described in the paper "Recovering OpenSSL
7885 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7886 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7887 <http://eprint.iacr.org/2014/140>
7889 Thanks to Yuval Yarom and Naomi Benger for discovering this
7890 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7892 *Yuval Yarom and Naomi Benger*
7894 * TLS pad extension: draft-agl-tls-padding-03
7896 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7897 TLS client Hello record length value would otherwise be > 255 and
7898 less that 512 pad with a dummy extension containing zeroes so it
7899 is at least 512 bytes long.
7901 *Adam Langley, Steve Henson*
7903 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7905 * Fix for TLS record tampering bug. A carefully crafted invalid
7906 handshake could crash OpenSSL with a NULL pointer exception.
7907 Thanks to Anton Johansson for reporting this issues.
7910 * Keep original DTLS digest and encryption contexts in retransmission
7911 structures so we can use the previous session parameters if they need
7912 to be resent. ([CVE-2013-6450])
7916 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7917 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7918 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7919 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7920 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7921 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7923 *Rob Stradling, Adam Langley*
7925 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7927 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7928 supporting platforms or when small records were transferred.
7930 *Andy Polyakov, Steve Henson*
7932 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7934 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7936 This addresses the flaw in CBC record processing discovered by
7937 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7938 at: <http://www.isg.rhul.ac.uk/tls/>
7940 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7941 Security Group at Royal Holloway, University of London
7942 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7943 Emilia Käsper for the initial patch.
7946 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7948 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7949 ciphersuites which can be exploited in a denial of service attack.
7950 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7951 and detecting this bug and to Wolfgang Ettlinger
7952 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7957 * Return an error when checking OCSP signatures when key is NULL.
7958 This fixes a DoS attack. ([CVE-2013-0166])
7962 * Make openssl verify return errors.
7964 *Chris Palmer <palmer@google.com> and Ben Laurie*
7966 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7967 the right response is stapled. Also change SSL_get_certificate()
7968 so it returns the certificate actually sent.
7969 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7971 *Rob Stradling <rob.stradling@comodo.com>*
7973 * Fix possible deadlock when decoding public keys.
7977 * Don't use TLS 1.0 record version number in initial client hello
7982 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7984 * Sanity check record length before skipping explicit IV in TLS
7985 1.2, 1.1 and DTLS to fix DoS attack.
7987 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7988 fuzzing as a service testing platform.
7993 * Initialise tkeylen properly when encrypting CMS messages.
7994 Thanks to Solar Designer of Openwall for reporting this issue.
7998 * In FIPS mode don't try to use composite ciphers as they are not
8003 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8005 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8006 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8007 mean any application compiled against OpenSSL 1.0.0 headers setting
8008 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8009 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8010 0x10000000L Any application which was previously compiled against
8011 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8012 will need to be recompiled as a result. Letting be results in
8013 inability to disable specifically TLS 1.1 and in client context,
8014 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8018 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8019 disable just protocol X, but all protocols above X *if* there are
8020 protocols *below* X still enabled. In more practical terms it means
8021 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8022 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8023 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8028 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8030 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8031 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8032 in CRYPTO_realloc_clean.
8034 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8035 issue and to Adam Langley <agl@chromium.org> for fixing it.
8038 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8040 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8044 * Workarounds for some broken servers that "hang" if a client hello
8045 record length exceeds 255 bytes.
8047 1. Do not use record version number > TLS 1.0 in initial client
8048 hello: some (but not all) hanging servers will now work.
8049 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8050 the number of ciphers sent in the client hello. This should be
8051 set to an even number, such as 50, for example by passing:
8052 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8053 Most broken servers should now work.
8054 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8055 TLS 1.2 client support entirely.
8059 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8063 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8065 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8066 STRING form instead of a DigestInfo.
8070 * The format used for MDC2 RSA signatures is inconsistent between EVP
8071 and the RSA_sign/RSA_verify functions. This was made more apparent when
8072 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8073 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8074 the correct format in RSA_verify so both forms transparently work.
8078 * Some servers which support TLS 1.0 can choke if we initially indicate
8079 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8080 encrypted premaster secret. As a workaround use the maximum permitted
8081 client version in client hello, this should keep such servers happy
8082 and still work with previous versions of OpenSSL.
8086 * Add support for TLS/DTLS heartbeats.
8088 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8090 * Add support for SCTP.
8092 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8094 * Improved PRNG seeding for VOS.
8096 *Paul Green <Paul.Green@stratus.com>*
8098 * Extensive assembler packs updates, most notably:
8100 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8101 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8102 - x86_64: bit-sliced AES implementation;
8103 - ARM: NEON support, contemporary platforms optimizations;
8104 - s390x: z196 support;
8105 - `*`: GHASH and GF(2^m) multiplication implementations;
8109 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8110 (removal of unnecessary code)
8112 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8114 * Add TLS key material exporter from RFC 5705.
8118 * Add DTLS-SRTP negotiation from RFC 5764.
8122 * Add Next Protocol Negotiation,
8123 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8124 disabled with a no-npn flag to config or Configure. Code donated
8127 *Adam Langley <agl@google.com> and Ben Laurie*
8129 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8130 NIST-P256, NIST-P521, with constant-time single point multiplication on
8131 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8132 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8133 Code made available under Apache License version 2.0.
8135 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8136 line to include this in your build of OpenSSL, and run "make depend" (or
8137 "make update"). This enables the following EC_METHODs:
8139 EC_GFp_nistp224_method()
8140 EC_GFp_nistp256_method()
8141 EC_GFp_nistp521_method()
8143 EC_GROUP_new_by_curve_name() will automatically use these (while
8144 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8147 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8149 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8150 all platforms. Move ssize_t definition from e_os.h to the public
8151 header file e_os2.h as it now appears in public header file cms.h
8155 * New -sigopt option to the ca, req and x509 utilities. Additional
8156 signature parameters can be passed using this option and in
8161 * Add RSA PSS signing function. This will generate and set the
8162 appropriate AlgorithmIdentifiers for PSS based on those in the
8163 corresponding EVP_MD_CTX structure. No application support yet.
8167 * Support for companion algorithm specific ASN1 signing routines.
8168 New function ASN1_item_sign_ctx() signs a pre-initialised
8169 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8170 the appropriate parameters.
8174 * Add new algorithm specific ASN1 verification initialisation function
8175 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8176 handling will be the same no matter what EVP_PKEY_METHOD is used.
8177 Add a PSS handler to support verification of PSS signatures: checked
8178 against a number of sample certificates.
8182 * Add signature printing for PSS. Add PSS OIDs.
8184 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8186 * Add algorithm specific signature printing. An individual ASN1 method
8187 can now print out signatures instead of the standard hex dump.
8189 More complex signatures (e.g. PSS) can print out more meaningful
8190 information. Include DSA version that prints out the signature
8195 * Password based recipient info support for CMS library: implementing
8200 * Split password based encryption into PBES2 and PBKDF2 functions. This
8201 neatly separates the code into cipher and PBE sections and is required
8202 for some algorithms that split PBES2 into separate pieces (such as
8203 password based CMS).
8207 * Session-handling fixes:
8208 - Fix handling of connections that are resuming with a session ID,
8209 but also support Session Tickets.
8210 - Fix a bug that suppressed issuing of a new ticket if the client
8211 presented a ticket with an expired session.
8212 - Try to set the ticket lifetime hint to something reasonable.
8213 - Make tickets shorter by excluding irrelevant information.
8214 - On the client side, don't ignore renewed tickets.
8216 *Adam Langley, Bodo Moeller (Google)*
8218 * Fix PSK session representation.
8222 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8224 This work was sponsored by Intel.
8228 * Add GCM support to TLS library. Some custom code is needed to split
8229 the IV between the fixed (from PRF) and explicit (from TLS record)
8230 portions. This adds all GCM ciphersuites supported by RFC5288 and
8231 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8232 add a special AESGCM string for GCM only.
8236 * Expand range of ctrls for AES GCM. Permit setting invocation
8237 field on decrypt and retrieval of invocation field only on encrypt.
8241 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8242 As required by RFC5289 these ciphersuites cannot be used if for
8243 versions of TLS earlier than 1.2.
8247 * For FIPS capable OpenSSL interpret a NULL default public key method
8248 as unset and return the appropriate default but do *not* set the default.
8249 This means we can return the appropriate method in applications that
8250 switch between FIPS and non-FIPS modes.
8254 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8255 ENGINE is used then we cannot handle that in the FIPS module so we
8256 keep original code iff non-FIPS operations are allowed.
8260 * Add -attime option to openssl utilities.
8262 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8264 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8268 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8269 FIPS EC methods unconditionally for now.
8273 * New build option no-ec2m to disable characteristic 2 code.
8277 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8278 all cases can be covered as some introduce binary incompatibilities.
8282 * Redirect RSA operations to FIPS module including keygen,
8283 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8287 * Add similar low-level API blocking to ciphers.
8291 * low-level digest APIs are not approved in FIPS mode: any attempt
8292 to use these will cause a fatal error. Applications that *really* want
8293 to use them can use the `private_*` version instead.
8297 * Redirect cipher operations to FIPS module for FIPS builds.
8301 * Redirect digest operations to FIPS module for FIPS builds.
8305 * Update build system to add "fips" flag which will link in fipscanister.o
8306 for static and shared library builds embedding a signature if needed.
8310 * Output TLS supported curves in preference order instead of numerical
8311 order. This is currently hardcoded for the highest order curves first.
8312 This should be configurable so applications can judge speed vs strength.
8316 * Add TLS v1.2 server support for client authentication.
8320 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8325 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8326 FIPS modules versions.
8330 * Add TLS v1.2 client side support for client authentication. Keep cache
8331 of handshake records longer as we don't know the hash algorithm to use
8332 until after the certificate request message is received.
8336 * Initial TLS v1.2 client support. Add a default signature algorithms
8337 extension including all the algorithms we support. Parse new signature
8338 format in client key exchange. Relax some ECC signing restrictions for
8339 TLS v1.2 as indicated in RFC5246.
8343 * Add server support for TLS v1.2 signature algorithms extension. Switch
8344 to new signature format when needed using client digest preference.
8345 All server ciphersuites should now work correctly in TLS v1.2. No client
8346 support yet and no support for client certificates.
8350 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8351 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8352 ciphersuites. At present only RSA key exchange ciphersuites work with
8353 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8354 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8355 and version checking.
8359 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8360 with this defined it will not be affected by any changes to ssl internal
8361 structures. Add several utility functions to allow openssl application
8362 to work with OPENSSL_NO_SSL_INTERN defined.
8366 * A long standing patch to add support for SRP from EdelWeb (Peter
8367 Sylvester and Christophe Renou) was integrated.
8368 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8369 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8372 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8376 * Permit abbreviated handshakes when renegotiating using the function
8377 SSL_renegotiate_abbreviated().
8379 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8381 * Add call to ENGINE_register_all_complete() to
8382 ENGINE_load_builtin_engines(), so some implementations get used
8383 automatically instead of needing explicit application support.
8387 * Add support for TLS key exporter as described in RFC5705.
8389 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8391 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8392 a few changes are required:
8394 Add SSL_OP_NO_TLSv1_1 flag.
8395 Add TLSv1_1 methods.
8396 Update version checking logic to handle version 1.1.
8397 Add explicit IV handling (ported from DTLS code).
8398 Add command line options to s_client/s_server.
8405 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8407 * X509_ATTRIBUTE memory leak
8409 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8410 memory. This structure is used by the PKCS#7 and CMS routines so any
8411 application which reads PKCS#7 or CMS data from untrusted sources is
8412 affected. SSL/TLS is not affected.
8414 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8420 * Race condition handling PSK identify hint
8422 If PSK identity hints are received by a multi-threaded client then
8423 the values are wrongly updated in the parent SSL_CTX structure. This can
8424 result in a race condition potentially leading to a double free of the
8430 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8432 * Malformed ECParameters causes infinite loop
8434 When processing an ECParameters structure OpenSSL enters an infinite loop
8435 if the curve specified is over a specially malformed binary polynomial
8438 This can be used to perform denial of service against any
8439 system which processes public keys, certificate requests or
8440 certificates. This includes TLS clients and TLS servers with
8441 client authentication enabled.
8443 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8448 * Exploitable out-of-bounds read in X509_cmp_time
8450 X509_cmp_time does not properly check the length of the ASN1_TIME
8451 string and can read a few bytes out of bounds. In addition,
8452 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8455 An attacker can use this to craft malformed certificates and CRLs of
8456 various sizes and potentially cause a segmentation fault, resulting in
8457 a DoS on applications that verify certificates or CRLs. TLS clients
8458 that verify CRLs are affected. TLS clients and servers with client
8459 authentication enabled may be affected if they use custom verification
8462 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8463 independently by Hanno Böck.
8468 * PKCS7 crash with missing EnvelopedContent
8470 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8471 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8472 with missing content and trigger a NULL pointer dereference on parsing.
8474 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8475 structures from untrusted sources are affected. OpenSSL clients and
8476 servers are not affected.
8478 This issue was reported to OpenSSL by Michal Zalewski (Google).
8483 * CMS verify infinite loop with unknown hash function
8485 When verifying a signedData message the CMS code can enter an infinite loop
8486 if presented with an unknown hash function OID. This can be used to perform
8487 denial of service against any system which verifies signedData messages using
8489 This issue was reported to OpenSSL by Johannes Bauer.
8494 * Race condition handling NewSessionTicket
8496 If a NewSessionTicket is received by a multi-threaded client when attempting to
8497 reuse a previous ticket then a race condition can occur potentially leading to
8498 a double free of the ticket data.
8503 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8505 * Segmentation fault in ASN1_TYPE_cmp fix
8507 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8508 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8509 certificate signature algorithm consistency this can be used to crash any
8510 certificate verification operation and exploited in a DoS attack. Any
8511 application which performs certificate verification is vulnerable including
8512 OpenSSL clients and servers which enable client authentication.
8517 * ASN.1 structure reuse memory corruption fix
8519 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8520 memory corruption via an invalid write. Such reuse is and has been
8521 strongly discouraged and is believed to be rare.
8523 Applications that parse structures containing CHOICE or ANY DEFINED BY
8524 components may be affected. Certificate parsing (d2i_X509 and related
8525 functions) are however not affected. OpenSSL clients and servers are
8531 * PKCS7 NULL pointer dereferences fix
8533 The PKCS#7 parsing code does not handle missing outer ContentInfo
8534 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8535 missing content and trigger a NULL pointer dereference on parsing.
8537 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8538 otherwise parse PKCS#7 structures from untrusted sources are
8539 affected. OpenSSL clients and servers are not affected.
8541 This issue was reported to OpenSSL by Michal Zalewski (Google).
8546 * DoS via reachable assert in SSLv2 servers fix
8548 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8549 servers that both support SSLv2 and enable export cipher suites by sending
8550 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8552 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8553 (OpenSSL development team).
8558 * Use After Free following d2i_ECPrivatekey error fix
8560 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8561 could cause a use after free condition. This, in turn, could cause a double
8562 free in several private key parsing functions (such as d2i_PrivateKey
8563 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8564 for applications that receive EC private keys from untrusted
8565 sources. This scenario is considered rare.
8567 This issue was discovered by the BoringSSL project and fixed in their
8573 * X509_to_X509_REQ NULL pointer deref fix
8575 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8576 the certificate key is invalid. This function is rarely used in practice.
8578 This issue was discovered by Brian Carpenter.
8583 * Removed the export ciphers from the DEFAULT ciphers
8587 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8589 * Build fixes for the Windows and OpenVMS platforms
8591 *Matt Caswell and Richard Levitte*
8593 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8595 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8596 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8597 dereference. This could lead to a Denial Of Service attack. Thanks to
8598 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8603 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8604 dtls1_buffer_record function under certain conditions. In particular this
8605 could occur if an attacker sent repeated DTLS records with the same
8606 sequence number but for the next epoch. The memory leak could be exploited
8607 by an attacker in a Denial of Service attack through memory exhaustion.
8608 Thanks to Chris Mueller for reporting this issue.
8613 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8614 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8615 method would be set to NULL which could later result in a NULL pointer
8616 dereference. Thanks to Frank Schmirler for reporting this issue.
8621 * Abort handshake if server key exchange message is omitted for ephemeral
8624 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8625 reporting this issue.
8630 * Remove non-export ephemeral RSA code on client and server. This code
8631 violated the TLS standard by allowing the use of temporary RSA keys in
8632 non-export ciphersuites and could be used by a server to effectively
8633 downgrade the RSA key length used to a value smaller than the server
8634 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8635 INRIA or reporting this issue.
8640 * Fixed issue where DH client certificates are accepted without verification.
8641 An OpenSSL server will accept a DH certificate for client authentication
8642 without the certificate verify message. This effectively allows a client to
8643 authenticate without the use of a private key. This only affects servers
8644 which trust a client certificate authority which issues certificates
8645 containing DH keys: these are extremely rare and hardly ever encountered.
8646 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8652 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8653 results on some platforms, including x86_64. This bug occurs at random
8654 with a very low probability, and is not known to be exploitable in any
8655 way, though its exact impact is difficult to determine. Thanks to Pieter
8656 Wuille (Blockstream) who reported this issue and also suggested an initial
8657 fix. Further analysis was conducted by the OpenSSL development team and
8658 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8659 the OpenSSL core team.
8664 * Fix various certificate fingerprint issues.
8666 By using non-DER or invalid encodings outside the signed portion of a
8667 certificate the fingerprint can be changed without breaking the signature.
8668 Although no details of the signed portion of the certificate can be changed
8669 this can cause problems with some applications: e.g. those using the
8670 certificate fingerprint for blacklists.
8672 1. Reject signatures with non zero unused bits.
8674 If the BIT STRING containing the signature has non zero unused bits reject
8675 the signature. All current signature algorithms require zero unused bits.
8677 2. Check certificate algorithm consistency.
8679 Check the AlgorithmIdentifier inside TBS matches the one in the
8680 certificate signature. NB: this will result in signature failure
8681 errors for some broken certificates.
8683 Thanks to Konrad Kraszewski from Google for reporting this issue.
8685 3. Check DSA/ECDSA signatures use DER.
8687 Reencode DSA/ECDSA signatures and compare with the original received
8688 signature. Return an error if there is a mismatch.
8690 This will reject various cases including garbage after signature
8691 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8692 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8693 (negative or with leading zeroes).
8695 Further analysis was conducted and fixes were developed by Stephen Henson
8696 of the OpenSSL core team.
8702 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8704 * Session Ticket Memory Leak.
8706 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8707 integrity of that ticket is first verified. In the event of a session
8708 ticket integrity check failing, OpenSSL will fail to free memory
8709 causing a memory leak. By sending a large number of invalid session
8710 tickets an attacker could exploit this issue in a Denial Of Service
8716 * Build option no-ssl3 is incomplete.
8718 When OpenSSL is configured with "no-ssl3" as a build option, servers
8719 could accept and complete an SSL 3.0 handshake, and clients could be
8720 configured to send them.
8723 *Akamai and the OpenSSL team*
8725 * Add support for TLS_FALLBACK_SCSV.
8726 Client applications doing fallback retries should call
8727 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8730 *Adam Langley, Bodo Moeller*
8732 * Add additional DigestInfo checks.
8734 Reencode DigestInto in DER and check against the original when
8735 verifying RSA signature: this will reject any improperly encoded
8736 DigestInfo structures.
8738 Note: this is a precautionary measure and no attacks are currently known.
8742 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8744 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8745 to a denial of service attack. A malicious server can crash the client
8746 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8747 ciphersuite and sending carefully crafted handshake messages.
8749 Thanks to Felix Gröbert (Google) for discovering and researching this
8755 * By sending carefully crafted DTLS packets an attacker could cause openssl
8756 to leak memory. This can be exploited through a Denial of Service attack.
8757 Thanks to Adam Langley for discovering and researching this issue.
8762 * An attacker can force openssl to consume large amounts of memory whilst
8763 processing DTLS handshake messages. This can be exploited through a
8764 Denial of Service attack.
8765 Thanks to Adam Langley for discovering and researching this issue.
8770 * An attacker can force an error condition which causes openssl to crash
8771 whilst processing DTLS packets due to memory being freed twice. This
8772 can be exploited through a Denial of Service attack.
8773 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8779 * If a multithreaded client connects to a malicious server using a resumed
8780 session and the server sends an ec point format extension it could write
8781 up to 255 bytes to freed memory.
8783 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8789 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8790 X509_name_oneline, X509_name_print_ex et al. to leak some information
8791 from the stack. Applications may be affected if they echo pretty printing
8792 output to the attacker.
8794 Thanks to Ivan Fratric (Google) for discovering this issue.
8797 *Emilia Käsper, and Steve Henson*
8799 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8800 for corner cases. (Certain input points at infinity could lead to
8801 bogus results, with non-infinity inputs mapped to infinity too.)
8805 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8807 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8808 handshake can force the use of weak keying material in OpenSSL
8809 SSL/TLS clients and servers.
8811 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8812 researching this issue. ([CVE-2014-0224])
8814 *KIKUCHI Masashi, Steve Henson*
8816 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8817 OpenSSL DTLS client the code can be made to recurse eventually crashing
8820 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8823 *Imre Rad, Steve Henson*
8825 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8826 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8827 client or server. This is potentially exploitable to run arbitrary
8828 code on a vulnerable client or server.
8830 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8832 *Jüri Aedla, Steve Henson*
8834 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8835 are subject to a denial of service attack.
8837 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8838 this issue. ([CVE-2014-3470])
8840 *Felix Gröbert, Ivan Fratric, Steve Henson*
8842 * Harmonize version and its documentation. -f flag is used to display
8845 *mancha <mancha1@zoho.com>*
8847 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8848 in i2d_ECPrivateKey.
8850 *mancha <mancha1@zoho.com>*
8852 * Fix some double frees. These are not thought to be exploitable.
8854 *mancha <mancha1@zoho.com>*
8856 * Fix for the attack described in the paper "Recovering OpenSSL
8857 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8858 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8859 <http://eprint.iacr.org/2014/140>
8861 Thanks to Yuval Yarom and Naomi Benger for discovering this
8862 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8864 *Yuval Yarom and Naomi Benger*
8866 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8868 * Keep original DTLS digest and encryption contexts in retransmission
8869 structures so we can use the previous session parameters if they need
8870 to be resent. ([CVE-2013-6450])
8874 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8875 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8876 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8877 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8878 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8879 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8881 *Rob Stradling, Adam Langley*
8883 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8885 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8887 This addresses the flaw in CBC record processing discovered by
8888 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8889 at: <http://www.isg.rhul.ac.uk/tls/>
8891 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8892 Security Group at Royal Holloway, University of London
8893 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8894 Emilia Käsper for the initial patch.
8897 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8899 * Return an error when checking OCSP signatures when key is NULL.
8900 This fixes a DoS attack. ([CVE-2013-0166])
8904 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8905 the right response is stapled. Also change SSL_get_certificate()
8906 so it returns the certificate actually sent.
8907 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8908 (This is a backport)
8910 *Rob Stradling <rob.stradling@comodo.com>*
8912 * Fix possible deadlock when decoding public keys.
8916 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8918 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8921 * Sanity check record length before skipping explicit IV in DTLS
8924 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8925 fuzzing as a service testing platform.
8930 * Initialise tkeylen properly when encrypting CMS messages.
8931 Thanks to Solar Designer of Openwall for reporting this issue.
8935 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8937 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8938 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8939 in CRYPTO_realloc_clean.
8941 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8942 issue and to Adam Langley <agl@chromium.org> for fixing it.
8945 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8947 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8949 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8950 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8951 content decryption and always return the same error. Note: this attack
8952 needs on average 2^20 messages so it only affects automated senders. The
8953 old behaviour can be re-enabled in the CMS code by setting the
8954 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8955 an MMA defence is not necessary.
8956 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8957 this issue. ([CVE-2012-0884])
8961 * Fix CVE-2011-4619: make sure we really are receiving a
8962 client hello before rejecting multiple SGC restarts. Thanks to
8963 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8967 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8969 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8970 Thanks to Antonio Martin, Enterprise Secure Access Research and
8971 Development, Cisco Systems, Inc. for discovering this bug and
8972 preparing a fix. ([CVE-2012-0050])
8976 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8978 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8979 of the Vaudenay padding oracle attack on CBC mode encryption
8980 which enables an efficient plaintext recovery attack against
8981 the OpenSSL implementation of DTLS. Their attack exploits timing
8982 differences arising during decryption processing. A research
8983 paper describing this attack can be found at:
8984 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8985 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8986 Security Group at Royal Holloway, University of London
8987 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8988 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8989 for preparing the fix. ([CVE-2011-4108])
8991 *Robin Seggelmann, Michael Tuexen*
8993 * Clear bytes used for block padding of SSL 3.0 records.
8996 *Adam Langley (Google)*
8998 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8999 Kadianakis <desnacked@gmail.com> for discovering this issue and
9000 Adam Langley for preparing the fix. ([CVE-2011-4619])
9002 *Adam Langley (Google)*
9004 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9006 *Andrey Kulikov <amdeich@gmail.com>*
9008 * Prevent malformed RFC3779 data triggering an assertion failure.
9009 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9010 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9012 *Rob Austein <sra@hactrn.net>*
9014 * Improved PRNG seeding for VOS.
9016 *Paul Green <Paul.Green@stratus.com>*
9018 * Fix ssl_ciph.c set-up race.
9020 *Adam Langley (Google)*
9022 * Fix spurious failures in ecdsatest.c.
9024 *Emilia Käsper (Google)*
9026 * Fix the BIO_f_buffer() implementation (which was mixing different
9027 interpretations of the `..._len` fields).
9029 *Adam Langley (Google)*
9031 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9032 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9033 threads won't reuse the same blinding coefficients.
9035 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9036 lock to call BN_BLINDING_invert_ex, and avoids one use of
9037 BN_BLINDING_update for each BN_BLINDING structure (previously,
9038 the last update always remained unused).
9040 *Emilia Käsper (Google)*
9042 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9044 *Bob Buckholz (Google)*
9046 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9048 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9049 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9051 *Kaspar Brand <ossl@velox.ch>*
9053 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9054 for multi-threaded use of ECDH. ([CVE-2011-3210])
9056 *Adam Langley (Google)*
9058 * Fix x509_name_ex_d2i memory leak on bad inputs.
9062 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9063 signature public key algorithm by using OID xref utilities instead.
9064 Before this you could only use some ECC ciphersuites with SHA1 only.
9068 * Add protection against ECDSA timing attacks as mentioned in the paper
9069 by Billy Bob Brumley and Nicola Tuveri, see:
9070 <http://eprint.iacr.org/2011/232.pdf>
9072 *Billy Bob Brumley and Nicola Tuveri*
9074 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9076 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9078 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9080 * Fix bug in string printing code: if *any* escaping is enabled we must
9081 escape the escape character (backslash) or the resulting string is
9086 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9088 * Disable code workaround for ancient and obsolete Netscape browsers
9089 and servers: an attacker can use it in a ciphersuite downgrade attack.
9090 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9094 * Fixed J-PAKE implementation error, originally discovered by
9095 Sebastien Martini, further info and confirmation from Stefan
9096 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9100 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9102 * Fix extension code to avoid race conditions which can result in a buffer
9103 overrun vulnerability: resumed sessions must not be modified as they can
9104 be shared by multiple threads. CVE-2010-3864
9108 * Fix WIN32 build system to correctly link an ENGINE directory into
9113 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9115 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9118 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9120 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9122 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9123 context. The operation can be customised via the ctrl mechanism in
9124 case ENGINEs want to include additional functionality.
9128 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9132 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9133 output hashes compatible with older versions of OpenSSL.
9135 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9137 * Fix compression algorithm handling: if resuming a session use the
9138 compression algorithm of the resumed session instead of determining
9139 it from client hello again. Don't allow server to change algorithm.
9143 * Add load_crls() function to commands tidying load_certs() too. Add option
9144 to verify utility to allow additional CRLs to be included.
9148 * Update OCSP request code to permit adding custom headers to the request:
9149 some responders need this.
9153 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9156 *Julia Lawall <julia@diku.dk>*
9158 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9159 needlessly dereferenced structures, used obsolete functions and
9160 didn't handle all updated verify codes correctly.
9164 * Disable MD2 in the default configuration.
9168 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9169 indicate the initial BIO being pushed or popped. This makes it possible
9170 to determine whether the BIO is the one explicitly called or as a result
9171 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9172 it handles reference counts correctly and doesn't zero out the I/O bio
9173 when it is not being explicitly popped. WARNING: applications which
9174 included workarounds for the old buggy behaviour will need to be modified
9175 or they could free up already freed BIOs.
9179 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9180 renaming to all platforms (within the 0.9.8 branch, this was
9181 done conditionally on Netware platforms to avoid a name clash).
9183 *Guenter <lists@gknw.net>*
9185 * Add ECDHE and PSK support to DTLS.
9187 *Michael Tuexen <tuexen@fh-muenster.de>*
9189 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9194 * Add "missing" function EVP_MD_flags() (without this the only way to
9195 retrieve a digest flags is by accessing the structure directly. Update
9196 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9197 or cipher is registered as in the "from" argument. Print out all
9198 registered digests in the dgst usage message instead of manually
9199 attempting to work them out.
9203 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9204 this allows the use of compression and extensions. Change default cipher
9205 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9206 by default unless an application cipher string requests it.
9210 * Alter match criteria in PKCS12_parse(). It used to try to use local
9211 key ids to find matching certificates and keys but some PKCS#12 files
9212 don't follow the (somewhat unwritten) rules and this strategy fails.
9213 Now just gather all certificates together and the first private key
9214 then look for the first certificate that matches the key.
9218 * Support use of registered digest and cipher names for dgst and cipher
9219 commands instead of having to add each one as a special case. So now
9226 openssl dgst -sha256 foo
9228 and this works for ENGINE based algorithms too.
9232 * Update Gost ENGINE to support parameter files.
9234 *Victor B. Wagner <vitus@cryptocom.ru>*
9236 * Support GeneralizedTime in ca utility.
9238 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9240 * Enhance the hash format used for certificate directory links. The new
9241 form uses the canonical encoding (meaning equivalent names will work
9242 even if they aren't identical) and uses SHA1 instead of MD5. This form
9243 is incompatible with the older format and as a result c_rehash should
9244 be used to rebuild symbolic links.
9248 * Make PKCS#8 the default write format for private keys, replacing the
9249 traditional format. This form is standardised, more secure and doesn't
9250 include an implicit MD5 dependency.
9254 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9255 committed to OpenSSL should pass this lot as a minimum.
9259 * Add session ticket override functionality for use by EAP-FAST.
9261 *Jouni Malinen <j@w1.fi>*
9263 * Modify HMAC functions to return a value. Since these can be implemented
9264 in an ENGINE errors can occur.
9268 * Type-checked OBJ_bsearch_ex.
9272 * Type-checked OBJ_bsearch. Also some constification necessitated
9273 by type-checking. Still to come: TXT_DB, bsearch(?),
9274 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9279 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9280 seconds to a tm structure directly, instead of going through OS
9281 specific date routines. This avoids any issues with OS routines such
9282 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9283 and X509_time_adj_ex() to cover the extended range. The existing
9284 X509_time_adj() is still usable and will no longer have any date issues.
9288 * Delta CRL support. New use deltas option which will attempt to locate
9289 and search any appropriate delta CRLs available.
9291 This work was sponsored by Google.
9295 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9296 code and add additional score elements. Validate alternate CRL paths
9297 as part of the CRL checking and indicate a new error "CRL path validation
9298 error" in this case. Applications wanting additional details can use
9299 the verify callback and check the new "parent" field. If this is not
9300 NULL CRL path validation is taking place. Existing applications won't
9301 see this because it requires extended CRL support which is off by
9304 This work was sponsored by Google.
9308 * Support for freshest CRL extension.
9310 This work was sponsored by Google.
9314 * Initial indirect CRL support. Currently only supported in the CRLs
9315 passed directly and not via lookup. Process certificate issuer
9316 CRL entry extension and lookup CRL entries by bother issuer name
9317 and serial number. Check and process CRL issuer entry in IDP extension.
9319 This work was sponsored by Google.
9323 * Add support for distinct certificate and CRL paths. The CRL issuer
9324 certificate is validated separately in this case. Only enabled if
9325 an extended CRL support flag is set: this flag will enable additional
9326 CRL functionality in future.
9328 This work was sponsored by Google.
9332 * Add support for policy mappings extension.
9334 This work was sponsored by Google.
9338 * Fixes to pathlength constraint, self issued certificate handling,
9339 policy processing to align with RFC3280 and PKITS tests.
9341 This work was sponsored by Google.
9345 * Support for name constraints certificate extension. DN, email, DNS
9346 and URI types are currently supported.
9348 This work was sponsored by Google.
9352 * To cater for systems that provide a pointer-based thread ID rather
9353 than numeric, deprecate the current numeric thread ID mechanism and
9354 replace it with a structure and associated callback type. This
9355 mechanism allows a numeric "hash" to be extracted from a thread ID in
9356 either case, and on platforms where pointers are larger than 'long',
9357 mixing is done to help ensure the numeric 'hash' is usable even if it
9358 can't be guaranteed unique. The default mechanism is to use "&errno"
9359 as a pointer-based thread ID to distinguish between threads.
9361 Applications that want to provide their own thread IDs should now use
9362 CRYPTO_THREADID_set_callback() to register a callback that will call
9363 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9365 Note that ERR_remove_state() is now deprecated, because it is tied
9366 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9367 to free the current thread's error state should be replaced by
9368 ERR_remove_thread_state(NULL).
9370 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9371 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9372 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9373 application was previously providing a numeric thread callback that
9374 was inappropriate for distinguishing threads, then uniqueness might
9375 have been obtained with &errno that happened immediately in the
9376 intermediate development versions of OpenSSL; this is no longer the
9377 case, the numeric thread callback will now override the automatic use
9380 *Geoff Thorpe, with help from Bodo Moeller*
9382 * Initial support for different CRL issuing certificates. This covers a
9383 simple case where the self issued certificates in the chain exist and
9384 the real CRL issuer is higher in the existing chain.
9386 This work was sponsored by Google.
9390 * Removed effectively defunct crypto/store from the build.
9394 * Revamp of STACK to provide stronger type-checking. Still to come:
9395 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9396 ASN1_STRING, CONF_VALUE.
9400 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9401 RAM on SSL connections. This option can save about 34k per idle SSL.
9405 * Revamp of LHASH to provide stronger type-checking. Still to come:
9406 STACK, TXT_DB, bsearch, qsort.
9410 * Initial support for Cryptographic Message Syntax (aka CMS) based
9411 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9412 support for data, signedData, compressedData, digestedData and
9413 encryptedData, envelopedData types included. Scripts to check against
9414 RFC4134 examples draft and interop and consistency checks of many
9415 content types and variants.
9419 * Add options to enc utility to support use of zlib compression BIO.
9423 * Extend mk1mf to support importing of options and assembly language
9424 files from Configure script, currently only included in VC-WIN32.
9425 The assembly language rules can now optionally generate the source
9426 files from the associated perl scripts.
9430 * Implement remaining functionality needed to support GOST ciphersuites.
9431 Interop testing has been performed using CryptoPro implementations.
9433 *Victor B. Wagner <vitus@cryptocom.ru>*
9435 * s390x assembler pack.
9439 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9444 * Implement Opaque PRF Input TLS extension as specified in
9445 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9446 official specification yet and no extension type assignment by
9447 IANA exists, this extension (for now) will have to be explicitly
9448 enabled when building OpenSSL by providing the extension number
9449 to use. For example, specify an option
9451 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9453 to the "config" or "Configure" script to enable the extension,
9454 assuming extension number 0x9527 (which is a completely arbitrary
9455 and unofficial assignment based on the MD5 hash of the Internet
9456 Draft). Note that by doing so, you potentially lose
9457 interoperability with other TLS implementations since these might
9458 be using the same extension number for other purposes.
9460 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9461 opaque PRF input value to use in the handshake. This will create
9462 an internal copy of the length-'len' string at 'src', and will
9463 return non-zero for success.
9465 To get more control and flexibility, provide a callback function
9468 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9469 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9473 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9476 Callback function 'cb' will be called in handshakes, and is
9477 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9478 Argument 'arg' is for application purposes (the value as given to
9479 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9480 be provided to the callback function). The callback function
9481 has to return non-zero to report success: usually 1 to use opaque
9482 PRF input just if possible, or 2 to enforce use of the opaque PRF
9483 input. In the latter case, the library will abort the handshake
9484 if opaque PRF input is not successfully negotiated.
9486 Arguments 'peerinput' and 'len' given to the callback function
9487 will always be NULL and 0 in the case of a client. A server will
9488 see the client's opaque PRF input through these variables if
9489 available (NULL and 0 otherwise). Note that if the server
9490 provides an opaque PRF input, the length must be the same as the
9491 length of the client's opaque PRF input.
9493 Note that the callback function will only be called when creating
9494 a new session (session resumption can resume whatever was
9495 previously negotiated), and will not be called in SSL 2.0
9496 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9497 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9498 for applications that need to enforce opaque PRF input.
9502 * Update ssl code to support digests other than SHA1+MD5 for handshake
9505 *Victor B. Wagner <vitus@cryptocom.ru>*
9507 * Add RFC4507 support to OpenSSL. This includes the corrections in
9508 RFC4507bis. The encrypted ticket format is an encrypted encoded
9509 SSL_SESSION structure, that way new session features are automatically
9512 If a client application caches session in an SSL_SESSION structure
9513 support is transparent because tickets are now stored in the encoded
9516 The SSL_CTX structure automatically generates keys for ticket
9517 protection in servers so again support should be possible
9518 with no application modification.
9520 If a client or server wishes to disable RFC4507 support then the option
9521 SSL_OP_NO_TICKET can be set.
9523 Add a TLS extension debugging callback to allow the contents of any client
9524 or server extensions to be examined.
9526 This work was sponsored by Google.
9530 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9531 OpenSSL should now compile cleanly on gcc 4.2
9533 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9535 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9536 support including streaming MAC support: this is required for GOST
9537 ciphersuite support.
9539 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9541 * Add option -stream to use PKCS#7 streaming in smime utility. New
9542 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9543 to output in BER and PEM format.
9547 * Experimental support for use of HMAC via EVP_PKEY interface. This
9548 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9549 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9550 ENGINE support for HMAC keys which are unextractable. New -mac and
9551 -macopt options to dgst utility.
9555 * New option -sigopt to dgst utility. Update dgst to use
9556 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9557 alternative signing parameters such as X9.31 or PSS in the dgst
9562 * Change ssl_cipher_apply_rule(), the internal function that does
9563 the work each time a ciphersuite string requests enabling
9564 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9565 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9566 the order of disabled ciphersuites such that those ciphersuites
9567 that most recently went from enabled to disabled not only stay
9568 in order with respect to each other, but also have higher priority
9569 than other disabled ciphersuites the next time ciphersuites are
9572 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9573 the same ciphersuites as with "HIGH" alone, but in a specific
9574 order where the PSK ciphersuites come first (since they are the
9575 most recently disabled ciphersuites when "HIGH" is parsed).
9577 Also, change ssl_create_cipher_list() (using this new
9578 functionality) such that between otherwise identical
9579 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9584 * Change ssl_create_cipher_list() so that it automatically
9585 arranges the ciphersuites in reasonable order before starting
9586 to process the rule string. Thus, the definition for "DEFAULT"
9587 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9588 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9589 This makes it much easier to arrive at a reasonable default order
9590 in applications for which anonymous ciphers are OK (meaning
9591 that you can't actually use DEFAULT).
9593 *Bodo Moeller; suggested by Victor Duchovni*
9595 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9596 processing) into multiple integers instead of setting
9597 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9598 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9599 (These masks as well as the individual bit definitions are hidden
9600 away into the non-exported interface ssl/ssl_locl.h, so this
9601 change to the definition of the SSL_CIPHER structure shouldn't
9602 affect applications.) This give us more bits for each of these
9603 categories, so there is no longer a need to coagulate AES128 and
9604 AES256 into a single algorithm bit, and to coagulate Camellia128
9605 and Camellia256 into a single algorithm bit, which has led to all
9608 Thus, among other things, the kludge introduced in 0.9.7m and
9609 0.9.8e for masking out AES256 independently of AES128 or masking
9610 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9612 With the change, we also introduce new ciphersuite aliases that
9613 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9618 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9619 Use the leftmost N bytes of the signature input if the input is
9620 larger than the prime q (with N being the size in bytes of q).
9624 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9625 it yet and it is largely untested.
9629 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9633 * Initial incomplete changes to avoid need for function casts in OpenSSL
9634 some compilers (gcc 4.2 and later) reject their use. Safestack is
9635 reimplemented. Update ASN1 to avoid use of legacy functions.
9639 * Win32/64 targets are linked with Winsock2.
9643 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9644 to external functions. This can be used to increase CRL handling
9645 efficiency especially when CRLs are very large by (for example) storing
9646 the CRL revoked certificates in a database.
9650 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9651 new CRLs added to a directory can be used. New command line option
9652 -verify_return_error to s_client and s_server. This causes real errors
9653 to be returned by the verify callback instead of carrying on no matter
9654 what. This reflects the way a "real world" verify callback would behave.
9658 * GOST engine, supporting several GOST algorithms and public key formats.
9659 Kindly donated by Cryptocom.
9663 * Partial support for Issuing Distribution Point CRL extension. CRLs
9664 partitioned by DP are handled but no indirect CRL or reason partitioning
9665 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9666 selected via a scoring technique which handles IDP and AKID in CRLs.
9670 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9671 will ultimately be used for all verify operations: this will remove the
9672 X509_STORE dependency on certificate verification and allow alternative
9673 lookup methods. X509_STORE based implementations of these two callbacks.
9677 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9678 Modify get_crl() to find a valid (unexpired) CRL if possible.
9682 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9683 this would be called X509_CRL_cmp() but that name is already used by
9684 a function that just compares CRL issuer names. Cache several CRL
9685 extensions in X509_CRL structure and cache CRLDP in X509.
9689 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9690 this maps equivalent X509_NAME structures into a consistent structure.
9691 Name comparison can then be performed rapidly using memcmp().
9695 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9700 * Allow digests to supply their own micalg string for S/MIME type using
9701 the ctrl EVP_MD_CTRL_MICALG.
9705 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9706 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9707 ctrl. It can then customise the structure before and/or after signing
9712 * New function OBJ_add_sigid() to allow application defined signature OIDs
9713 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9714 to free up any added signature OIDs.
9718 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9719 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9720 digest and cipher tables. New options added to openssl utility:
9721 list-message-digest-algorithms and list-cipher-algorithms.
9725 * Change the array representation of binary polynomials: the list
9726 of degrees of non-zero coefficients is now terminated with -1.
9727 Previously it was terminated with 0, which was also part of the
9728 value; thus, the array representation was not applicable to
9729 polynomials where t^0 has coefficient zero. This change makes
9730 the array representation useful in a more general context.
9734 * Various modifications and fixes to SSL/TLS cipher string
9735 handling. For ECC, the code now distinguishes between fixed ECDH
9736 with RSA certificates on the one hand and with ECDSA certificates
9737 on the other hand, since these are separate ciphersuites. The
9738 unused code for Fortezza ciphersuites has been removed.
9740 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9741 (not "ECDHE"). For consistency with the code for DH
9742 certificates, use of ECDH certificates is now considered ECDH
9743 authentication, not RSA or ECDSA authentication (the latter is
9744 merely the CA's signing algorithm and not actively used in the
9747 The temporary ciphersuite alias "ECCdraft" is no longer
9748 available, and ECC ciphersuites are no longer excluded from "ALL"
9749 and "DEFAULT". The following aliases now exist for RFC 4492
9750 ciphersuites, most of these by analogy with the DH case:
9752 kECDHr - ECDH cert, signed with RSA
9753 kECDHe - ECDH cert, signed with ECDSA
9754 kECDH - ECDH cert (signed with either RSA or ECDSA)
9755 kEECDH - ephemeral ECDH
9756 ECDH - ECDH cert or ephemeral ECDH
9762 AECDH - anonymous ECDH
9763 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9767 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9768 Use correct micalg parameters depending on digest(s) in signed message.
9772 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9773 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9777 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9778 an engine to register a method. Add ENGINE lookups for methods and
9779 functional reference processing.
9783 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9784 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9789 * New -resign option to smime utility. This adds one or more signers
9790 to an existing PKCS#7 signedData structure. Also -md option to use an
9791 alternative message digest algorithm for signing.
9795 * Tidy up PKCS#7 routines and add new functions to make it easier to
9796 create PKCS7 structures containing multiple signers. Update smime
9797 application to support multiple signers.
9801 * New -macalg option to pkcs12 utility to allow setting of an alternative
9806 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9807 Reorganize PBE internals to lookup from a static table using NIDs,
9808 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9809 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9810 PRF which will be automatically used with PBES2.
9814 * Replace the algorithm specific calls to generate keys in "req" with the
9819 * Update PKCS#7 enveloped data routines to use new API. This is now
9820 supported by any public key method supporting the encrypt operation. A
9821 ctrl is added to allow the public key algorithm to examine or modify
9822 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9827 * Add a ctrl to asn1 method to allow a public key algorithm to express
9828 a default digest type to use. In most cases this will be SHA1 but some
9829 algorithms (such as GOST) need to specify an alternative digest. The
9830 return value indicates how strong the preference is 1 means optional and
9831 2 is mandatory (that is it is the only supported type). Modify
9832 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9833 use the default md. Update openssl utilities to use the default digest
9834 type for signing if it is not explicitly indicated.
9838 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9839 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9840 signing method from the key type. This effectively removes the link
9841 between digests and public key types.
9845 * Add an OID cross reference table and utility functions. Its purpose is to
9846 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9847 rsaEncryption. This will allow some of the algorithm specific hackery
9848 needed to use the correct OID to be removed.
9852 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9853 structures for PKCS7_sign(). They are now set up by the relevant public
9858 * Add provisional EC pkey method with support for ECDSA and ECDH.
9862 * Add support for key derivation (agreement) in the API, DH method and
9867 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9868 public and private key formats. As a side effect these add additional
9869 command line functionality not previously available: DSA signatures can be
9870 generated and verified using pkeyutl and DH key support and generation in
9877 *Oliver Tappe <zooey@hirschkaefer.de>*
9879 * New make target "install_html_docs" installs HTML renditions of the
9882 *Oliver Tappe <zooey@hirschkaefer.de>*
9884 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9885 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9886 support key and parameter generation and add initial key generation
9887 functionality for RSA.
9891 * Add functions for main EVP_PKEY_method operations. The undocumented
9892 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9893 `EVP_PKEY_{encrypt,decrypt}_old`.
9897 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9898 key API, doesn't do much yet.
9902 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9903 public key algorithms. New option to openssl utility:
9904 "list-public-key-algorithms" to print out info.
9908 * Implement the Supported Elliptic Curves Extension for
9909 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9913 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9914 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9918 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9919 utilities such as rsa, dsa, dsaparam etc except they process any key
9924 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9925 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9926 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9931 * Initial support for pluggable public key ASN1.
9932 De-spaghettify the public key ASN1 handling. Move public and private
9933 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9934 algorithm specific handling to a single module within the relevant
9935 algorithm directory. Add functions to allow (near) opaque processing
9936 of public and private key structures.
9940 * Implement the Supported Point Formats Extension for
9941 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9945 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9946 for the psk identity [hint] and the psk callback functions to the
9947 SSL_SESSION, SSL and SSL_CTX structure.
9950 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9954 SSL_CTX_use_psk_identity_hint
9955 SSL_get_psk_identity_hint
9956 SSL_get_psk_identity
9957 SSL_use_psk_identity_hint
9959 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9961 * Add RFC 3161 compliant time stamp request creation, response generation
9962 and response verification functionality.
9964 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9966 * Add initial support for TLS extensions, specifically for the server_name
9967 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9968 have new members for a hostname. The SSL data structure has an
9969 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9970 stored in that context to allow for session resumption, even after the
9971 SSL has been switched to a new SSL_CTX in reaction to a client's
9972 server_name extension.
9974 New functions (subject to change):
9976 SSL_get_servername()
9977 SSL_get_servername_type()
9980 New CTRL codes and macros (subject to change):
9982 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9983 - SSL_CTX_set_tlsext_servername_callback()
9984 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9985 - SSL_CTX_set_tlsext_servername_arg()
9986 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9988 openssl s_client has a new '-servername ...' option.
9990 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9991 '-key2 ...', '-servername_fatal' (subject to change). This allows
9992 testing the HostName extension for a specific single hostname ('-cert'
9993 and '-key' remain fallbacks for handshakes without HostName
9994 negotiation). If the unrecognized_name alert has to be sent, this by
9995 default is a warning; it becomes fatal with the '-servername_fatal'
9998 *Peter Sylvester, Remy Allais, Christophe Renou*
10000 * Whirlpool hash implementation is added.
10004 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10005 bn(64,32). Because of instruction set limitations it doesn't have
10006 any negative impact on performance. This was done mostly in order
10007 to make it possible to share assembler modules, such as bn_mul_mont
10008 implementations, between 32- and 64-bit builds without hassle.
10012 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10013 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10018 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10019 dedicated Montgomery multiplication procedure, is introduced.
10020 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10021 "64-bit" performance on certain 32-bit targets.
10025 * New option SSL_OP_NO_COMP to disable use of compression selectively
10026 in SSL structures. New SSL ctrl to set maximum send fragment size.
10027 Save memory by setting the I/O buffer sizes dynamically instead of
10028 using the maximum available value.
10032 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10033 in addition to the text details.
10037 * Very, very preliminary EXPERIMENTAL support for printing of general
10038 ASN1 structures. This currently produces rather ugly output and doesn't
10039 handle several customised structures at all.
10043 * Integrated support for PVK file format and some related formats such
10044 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10045 these in the 'rsa' and 'dsa' utilities.
10049 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10053 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10054 place for the (very old) "NETSCAPE" format certificates which are now
10055 handled using new ASN1 code equivalents.
10059 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10060 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10061 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10065 * Modify CRL distribution points extension code to print out previously
10066 unsupported fields. Enhance extension setting code to allow setting of
10071 * Add print and set support for Issuing Distribution Point CRL extension.
10075 * Change 'Configure' script to enable Camellia by default.
10082 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10084 * When rejecting SSL/TLS records due to an incorrect version number, never
10085 update s->server with a new major version number. As of
10086 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10087 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10088 the previous behavior could result in a read attempt at NULL when
10089 receiving specific incorrect SSL/TLS records once record payload
10090 protection is active. ([CVE-2010-0740])
10092 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10094 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10095 could be crashed if the relevant tables were not present (e.g. chrooted).
10097 *Tomas Hoger <thoger@redhat.com>*
10099 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10101 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10103 *Martin Olsson, Neel Mehta*
10105 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10106 accommodate for stack sorting, always a write lock!).
10110 * On some versions of WIN32 Heap32Next is very slow. This can cause
10111 excessive delays in the RAND_poll(): over a minute. As a workaround
10112 include a time check in the inner Heap32Next loop too.
10116 * The code that handled flushing of data in SSL/TLS originally used the
10117 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10118 the problem outlined in PR#1949. The fix suggested there however can
10119 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10120 of Apache). So instead simplify the code to flush unconditionally.
10121 This should be fine since flushing with no data to flush is a no op.
10125 * Handle TLS versions 2.0 and later properly and correctly use the
10126 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10127 off ancient servers have a habit of sticking around for a while...
10131 * Modify compression code so it frees up structures without using the
10132 ex_data callbacks. This works around a problem where some applications
10133 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10134 restarting) then use compression (e.g. SSL with compression) later.
10135 This results in significant per-connection memory leaks and
10136 has caused some security issues including CVE-2008-1678 and
10141 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10142 change when encrypting or decrypting.
10146 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10147 connect and renegotiate with servers which do not support RI.
10148 Until RI is more widely deployed this option is enabled by default.
10152 * Add "missing" ssl ctrls to clear options and mode.
10156 * If client attempts to renegotiate and doesn't support RI respond with
10157 a no_renegotiation alert as required by RFC5746. Some renegotiating
10158 TLS clients will continue a connection gracefully when they receive
10159 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10160 waiting for a server hello which it will never receive. Now we treat a
10161 received no_renegotiation alert as a fatal error. This is because
10162 applications requesting a renegotiation might well expect it to succeed
10163 and would have no code in place to handle the server denying it so the
10164 only safe thing to do is to terminate the connection.
10168 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10169 peer supports secure renegotiation and 0 otherwise. Print out peer
10170 renegotiation support in s_client/s_server.
10174 * Replace the highly broken and deprecated SPKAC certification method with
10175 the updated NID creation version. This should correctly handle UTF8.
10179 * Implement RFC5746. Re-enable renegotiation but require the extension
10180 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10181 turns out to be a bad idea. It has been replaced by
10182 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10183 SSL_CTX_set_options(). This is really not recommended unless you
10184 know what you are doing.
10186 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10188 * Fixes to stateless session resumption handling. Use initial_ctx when
10189 issuing and attempting to decrypt tickets in case it has changed during
10190 servername handling. Use a non-zero length session ID when attempting
10191 stateless session resumption: this makes it possible to determine if
10192 a resumption has occurred immediately after receiving server hello
10193 (several places in OpenSSL subtly assume this) instead of later in
10198 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10199 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10200 fixes for a few places where the return code is not checked
10203 *Julia Lawall <julia@diku.dk>*
10205 * Add --strict-warnings option to Configure script to include devteam
10206 warnings in other configurations.
10210 * Add support for --libdir option and LIBDIR variable in makefiles. This
10211 makes it possible to install openssl libraries in locations which
10212 have names other than "lib", for example "/usr/lib64" which some
10215 *Steve Henson, based on patch from Jeremy Utley*
10217 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10218 X690 8.9.12 and can produce some misleading textual output of OIDs.
10220 *Steve Henson, reported by Dan Kaminsky*
10222 * Delete MD2 from algorithm tables. This follows the recommendation in
10223 several standards that it is not used in new applications due to
10224 several cryptographic weaknesses. For binary compatibility reasons
10225 the MD2 API is still compiled in by default.
10229 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10234 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10235 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10238 *Guenter <lists@gknw.net>*
10240 * Fix the server certificate chain building code to use X509_verify_cert(),
10241 it used to have an ad-hoc builder which was unable to cope with anything
10242 other than a simple chain.
10244 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10246 * Don't check self signed certificate signatures in X509_verify_cert()
10247 by default (a flag can override this): it just wastes time without
10248 adding any security. As a useful side effect self signed root CAs
10249 with non-FIPS digests are now usable in FIPS mode.
10253 * In dtls1_process_out_of_seq_message() the check if the current message
10254 is already buffered was missing. For every new message was memory
10255 allocated, allowing an attacker to perform an denial of service attack
10256 with sending out of seq handshake messages until there is no memory
10257 left. Additionally every future message was buffered, even if the
10258 sequence number made no sense and would be part of another handshake.
10259 So only messages with sequence numbers less than 10 in advance will be
10260 buffered. ([CVE-2009-1378])
10262 *Robin Seggelmann, discovered by Daniel Mentz*
10264 * Records are buffered if they arrive with a future epoch to be
10265 processed after finishing the corresponding handshake. There is
10266 currently no limitation to this buffer allowing an attacker to perform
10267 a DOS attack with sending records with future epochs until there is no
10268 memory left. This patch adds the pqueue_size() function to determine
10269 the size of a buffer and limits the record buffer to 100 entries.
10272 *Robin Seggelmann, discovered by Daniel Mentz*
10274 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10275 parent structure is freed. ([CVE-2009-1379])
10279 * Handle non-blocking I/O properly in SSL_shutdown() call.
10281 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10283 * Add `2.5.4.*` OIDs
10285 *Ilya O. <vrghost@gmail.com>*
10287 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10289 * Disable renegotiation completely - this fixes a severe security
10290 problem ([CVE-2009-3555]) at the cost of breaking all
10291 renegotiation. Renegotiation can be re-enabled by setting
10292 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10293 run-time. This is really not recommended unless you know what
10298 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10300 * Don't set val to NULL when freeing up structures, it is freed up by
10301 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10302 zeroing past the valid field. ([CVE-2009-0789])
10304 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10306 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10307 checked correctly. This would allow some invalid signed attributes to
10308 appear to verify correctly. ([CVE-2009-0591])
10310 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10312 * Reject UniversalString and BMPString types with invalid lengths. This
10313 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10314 a legal length. ([CVE-2009-0590])
10318 * Set S/MIME signing as the default purpose rather than setting it
10319 unconditionally. This allows applications to override it at the store
10324 * Permit restricted recursion of ASN1 strings. This is needed in practice
10325 to handle some structures.
10329 * Improve efficiency of mem_gets: don't search whole buffer each time
10332 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10334 * New -hex option for openssl rand.
10338 * Print out UTF8String and NumericString when parsing ASN1.
10342 * Support NumericString type for name components.
10346 * Allow CC in the environment to override the automatically chosen
10347 compiler. Note that nothing is done to ensure flags work with the
10352 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10354 * Properly check EVP_VerifyFinal() and similar return values
10357 *Ben Laurie, Bodo Moeller, Google Security Team*
10359 * Enable TLS extensions by default.
10363 * Allow the CHIL engine to be loaded, whether the application is
10364 multithreaded or not. (This does not release the developer from the
10365 obligation to set up the dynamic locking callbacks.)
10367 *Sander Temme <sander@temme.net>*
10369 * Use correct exit code if there is an error in dgst command.
10371 *Steve Henson; problem pointed out by Roland Dirlewanger*
10373 * Tweak Configure so that you need to say "experimental-jpake" to enable
10374 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10378 * Add experimental JPAKE support, including demo authentication in
10379 s_client and s_server.
10383 * Set the comparison function in v3_addr_canonize().
10385 *Rob Austein <sra@hactrn.net>*
10387 * Add support for XMPP STARTTLS in s_client.
10389 *Philip Paeps <philip@freebsd.org>*
10391 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10392 to ensure that even with this option, only ciphersuites in the
10393 server's preference list will be accepted. (Note that the option
10394 applies only when resuming a session, so the earlier behavior was
10395 just about the algorithm choice for symmetric cryptography.)
10399 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10401 * Fix NULL pointer dereference if a DTLS server received
10402 ChangeCipherSpec as first record ([CVE-2009-1386]).
10406 * Fix a state transition in s3_srvr.c and d1_srvr.c
10407 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10409 *Nagendra Modadugu*
10411 * The fix in 0.9.8c that supposedly got rid of unsafe
10412 double-checked locking was incomplete for RSA blinding,
10413 addressing just one layer of what turns out to have been
10414 doubly unsafe triple-checked locking.
10416 So now fix this for real by retiring the MONT_HELPER macro
10417 in crypto/rsa/rsa_eay.c.
10419 *Bodo Moeller; problem pointed out by Marius Schilder*
10421 * Various precautionary measures:
10423 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10425 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10426 (NB: This would require knowledge of the secret session ticket key
10427 to exploit, in which case you'd be SOL either way.)
10429 - Change bn_nist.c so that it will properly handle input BIGNUMs
10430 outside the expected range.
10432 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10435 *Neel Mehta, Bodo Moeller*
10437 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10438 the load fails. Useful for distros.
10440 *Ben Laurie and the FreeBSD team*
10442 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10446 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10450 * Expand ENGINE to support engine supplied SSL client certificate functions.
10452 This work was sponsored by Logica.
10456 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10457 keystores. Support for SSL/TLS client authentication too.
10458 Not compiled unless enable-capieng specified to Configure.
10460 This work was sponsored by Logica.
10464 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10465 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10466 attribute creation routines such as certificate requests and PKCS#12
10471 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10473 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10474 handshake which could lead to a client crash as found using the
10475 Codenomicon TLS test suite ([CVE-2008-1672])
10477 *Steve Henson, Mark Cox*
10479 * Fix double free in TLS server name extensions which could lead to
10480 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10484 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10486 Clear the error queue to ensure that error entries left from
10487 older function calls do not interfere with the correct operation.
10489 *Lutz Jaenicke, Erik de Castro Lopo*
10491 * Remove root CA certificates of commercial CAs:
10493 The OpenSSL project does not recommend any specific CA and does not
10494 have any policy with respect to including or excluding any CA.
10495 Therefore, it does not make any sense to ship an arbitrary selection
10496 of root CA certificates with the OpenSSL software.
10500 * RSA OAEP patches to fix two separate invalid memory reads.
10501 The first one involves inputs when 'lzero' is greater than
10502 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10503 before the beginning of from). The second one involves inputs where
10504 the 'db' section contains nothing but zeroes (there is a one-byte
10505 invalid read after the end of 'db').
10507 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10509 * Partial backport from 0.9.9-dev:
10511 Introduce bn_mul_mont (dedicated Montgomery multiplication
10512 procedure) as a candidate for BIGNUM assembler implementation.
10513 While 0.9.9-dev uses assembler for various architectures, only
10514 x86_64 is available by default here in the 0.9.8 branch, and
10515 32-bit x86 is available through a compile-time setting.
10517 To try the 32-bit x86 assembler implementation, use Configure
10518 option "enable-montasm" (which exists only for this backport).
10520 As "enable-montasm" for 32-bit x86 disclaims code stability
10521 anyway, in this constellation we activate additional code
10522 backported from 0.9.9-dev for further performance improvements,
10523 namely BN_from_montgomery_word. (To enable this otherwise,
10524 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10526 *Andy Polyakov (backport partially by Bodo Moeller)*
10528 * Add TLS session ticket callback. This allows an application to set
10529 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10530 values. This is useful for key rollover for example where several key
10531 sets may exist with different names.
10535 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10536 This was broken until now in 0.9.8 releases, such that the only way
10537 a registered ENGINE could be used (assuming it initialises
10538 successfully on the host) was to explicitly set it as the default
10539 for the relevant algorithms. This is in contradiction with 0.9.7
10540 behaviour and the documentation. With this fix, when an ENGINE is
10541 registered into a given algorithm's table of implementations, the
10542 'uptodate' flag is reset so that auto-discovery will be used next
10543 time a new context for that algorithm attempts to select an
10546 *Ian Lister (tweaked by Geoff Thorpe)*
10548 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10549 implementation in the following ways:
10551 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10554 Lack of BER streaming support means one pass streaming processing is
10555 only supported if data is detached: setting the streaming flag is
10556 ignored for embedded content.
10558 CMS support is disabled by default and must be explicitly enabled
10559 with the enable-cms configuration option.
10563 * Update the GMP engine glue to do direct copies between BIGNUM and
10564 mpz_t when openssl and GMP use the same limb size. Otherwise the
10565 existing "conversion via a text string export" trick is still used.
10567 *Paul Sheer <paulsheer@gmail.com>*
10569 * Zlib compression BIO. This is a filter BIO which compressed and
10570 uncompresses any data passed through it.
10574 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10575 RFC3394 compatible AES key wrapping.
10579 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10580 sets string data without copying. X509_ALGOR_set0() and
10581 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10582 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10583 from an X509_ATTRIBUTE structure optionally checking it occurs only
10584 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10589 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10590 to get the expected BN_FLG_CONSTTIME behavior.
10592 *Bodo Moeller (Google)*
10596 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10597 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10598 - added some more tests to do_tests.pl
10599 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10600 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10601 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10602 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10603 - various changes to netware.pl to enable gcc-cross builds on Win32
10605 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10606 - various changes to fix missing prototype warnings
10607 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10608 - added AES, WHIRLPOOL and CPUID assembler code to build files
10609 - added missing AES assembler make rules to mk1mf.pl
10610 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10612 *Guenter Knauf <eflash@gmx.net>*
10614 * Implement certificate status request TLS extension defined in RFC3546.
10615 A client can set the appropriate parameters and receive the encoded
10616 OCSP response via a callback. A server can query the supplied parameters
10617 and set the encoded OCSP response in the callback. Add simplified examples
10618 to s_client and s_server.
10622 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10624 * Fix various bugs:
10625 + Binary incompatibility of ssl_ctx_st structure
10626 + DTLS interoperation with non-compliant servers
10627 + Don't call get_session_cb() without proposed session
10628 + Fix ia64 assembler code
10630 *Andy Polyakov, Steve Henson*
10632 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10634 * DTLS Handshake overhaul. There were longstanding issues with
10635 OpenSSL DTLS implementation, which were making it impossible for
10636 RFC 4347 compliant client to communicate with OpenSSL server.
10637 Unfortunately just fixing these incompatibilities would "cut off"
10638 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10639 server keeps tolerating non RFC compliant syntax. The opposite is
10640 not true, 0.9.8f client can not communicate with earlier server.
10641 This update even addresses CVE-2007-4995.
10645 * Changes to avoid need for function casts in OpenSSL: some compilers
10646 (gcc 4.2 and later) reject their use.
10647 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10650 * Add RFC4507 support to OpenSSL. This includes the corrections in
10651 RFC4507bis. The encrypted ticket format is an encrypted encoded
10652 SSL_SESSION structure, that way new session features are automatically
10655 If a client application caches session in an SSL_SESSION structure
10656 support is transparent because tickets are now stored in the encoded
10659 The SSL_CTX structure automatically generates keys for ticket
10660 protection in servers so again support should be possible
10661 with no application modification.
10663 If a client or server wishes to disable RFC4507 support then the option
10664 SSL_OP_NO_TICKET can be set.
10666 Add a TLS extension debugging callback to allow the contents of any client
10667 or server extensions to be examined.
10669 This work was sponsored by Google.
10673 * Add initial support for TLS extensions, specifically for the server_name
10674 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10675 have new members for a hostname. The SSL data structure has an
10676 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10677 stored in that context to allow for session resumption, even after the
10678 SSL has been switched to a new SSL_CTX in reaction to a client's
10679 server_name extension.
10681 New functions (subject to change):
10683 SSL_get_servername()
10684 SSL_get_servername_type()
10687 New CTRL codes and macros (subject to change):
10689 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10690 - SSL_CTX_set_tlsext_servername_callback()
10691 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10692 - SSL_CTX_set_tlsext_servername_arg()
10693 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10695 openssl s_client has a new '-servername ...' option.
10697 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10698 '-key2 ...', '-servername_fatal' (subject to change). This allows
10699 testing the HostName extension for a specific single hostname ('-cert'
10700 and '-key' remain fallbacks for handshakes without HostName
10701 negotiation). If the unrecognized_name alert has to be sent, this by
10702 default is a warning; it becomes fatal with the '-servername_fatal'
10705 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10707 * Add AES and SSE2 assembly language support to VC++ build.
10711 * Mitigate attack on final subtraction in Montgomery reduction.
10715 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10716 (which previously caused an internal error).
10720 * Squeeze another 10% out of IGE mode when in != out.
10724 * AES IGE mode speedup.
10726 *Dean Gaudet (Google)*
10728 * Add the Korean symmetric 128-bit cipher SEED (see
10729 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10730 add SEED ciphersuites from RFC 4162:
10732 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10733 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10734 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10735 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10737 To minimize changes between patchlevels in the OpenSSL 0.9.8
10738 series, SEED remains excluded from compilation unless OpenSSL
10739 is configured with 'enable-seed'.
10741 *KISA, Bodo Moeller*
10743 * Mitigate branch prediction attacks, which can be practical if a
10744 single processor is shared, allowing a spy process to extract
10745 information. For detailed background information, see
10746 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10747 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10748 and Necessary Software Countermeasures"). The core of the change
10749 are new versions BN_div_no_branch() and
10750 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10751 respectively, which are slower, but avoid the security-relevant
10752 conditional branches. These are automatically called by BN_div()
10753 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10754 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10755 remove a conditional branch.
10757 BN_FLG_CONSTTIME is the new name for the previous
10758 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10759 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10760 in the exponent causes BN_mod_exp_mont() to use the alternative
10761 implementation in BN_mod_exp_mont_consttime().) The old name
10762 remains as a deprecated alias.
10764 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10765 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10766 constant-time implementations for more than just exponentiation.
10767 Here too the old name is kept as a deprecated alias.
10769 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10770 the BN_BLINDING structure gets an independent copy of the
10771 modulus. This means that the previous `BIGNUM *m` argument to
10772 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10773 essentially becomes `const BIGNUM *m`, although we can't actually
10774 change this in the header file before 0.9.9. It allows
10775 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10776 enable BN_FLG_CONSTTIME.
10778 *Matthew D Wood (Intel Corp)*
10780 * In the SSL/TLS server implementation, be strict about session ID
10781 context matching (which matters if an application uses a single
10782 external cache for different purposes). Previously,
10783 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10784 set. This did ensure strict client verification, but meant that,
10785 with applications using a single external cache for quite
10786 different requirements, clients could circumvent ciphersuite
10787 restrictions for a given session ID context by starting a session
10788 in a different context.
10792 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10793 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10794 authentication-only ciphersuites.
10798 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10799 not complete and could lead to a possible single byte overflow
10800 ([CVE-2007-5135]) [Ben Laurie]
10802 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10804 * Since AES128 and AES256 (and similarly Camellia128 and
10805 Camellia256) share a single mask bit in the logic of
10806 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10807 kludge to work properly if AES128 is available and AES256 isn't
10808 (or if Camellia128 is available and Camellia256 isn't).
10812 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10813 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10814 When a point or a seed is encoded in a BIT STRING, we need to
10815 prevent the removal of trailing zero bits to get the proper DER
10816 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10817 of a NamedBitList, for which trailing 0 bits need to be removed.)
10821 * Have SSL/TLS server implementation tolerate "mismatched" record
10822 protocol version while receiving ClientHello even if the
10823 ClientHello is fragmented. (The server can't insist on the
10824 particular protocol version it has chosen before the ServerHello
10825 message has informed the client about his choice.)
10829 * Add RFC 3779 support.
10831 *Rob Austein for ARIN, Ben Laurie*
10833 * Load error codes if they are not already present instead of using a
10834 static variable. This allows them to be cleanly unloaded and reloaded.
10835 Improve header file function name parsing.
10839 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10840 or CAPABILITY handshake as required by RFCs.
10842 *Goetz Babin-Ebell*
10844 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10846 * Introduce limits to prevent malicious keys being able to
10847 cause a denial of service. ([CVE-2006-2940])
10849 *Steve Henson, Bodo Moeller*
10851 * Fix ASN.1 parsing of certain invalid structures that can result
10852 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10854 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10855 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10857 * Fix SSL client code which could crash if connecting to a
10858 malicious SSLv2 server. ([CVE-2006-4343])
10860 *Tavis Ormandy and Will Drewry, Google Security Team*
10862 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10863 match only those. Before that, "AES256-SHA" would be interpreted
10864 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10865 the same strength classification in 0.9.7h) as we currently only
10866 have a single AES bit in the ciphersuite description bitmap.
10867 That change, however, also applied to ciphersuite strings such as
10868 "RC4-MD5" that intentionally matched multiple ciphersuites --
10869 namely, SSL 2.0 ciphersuites in addition to the more common ones
10870 from SSL 3.0/TLS 1.0.
10872 So we change the selection algorithm again: Naming an explicit
10873 ciphersuite selects this one ciphersuite, and any other similar
10874 ciphersuite (same bitmap) from *other* protocol versions.
10875 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10876 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10878 Since SSL 2.0 does not have any ciphersuites for which the
10879 128/256 bit distinction would be relevant, this works for now.
10880 The proper fix will be to use different bits for AES128 and
10881 AES256, which would have avoided the problems from the beginning;
10882 however, bits are scarce, so we can only do this in a new release
10883 (not just a patchlevel) when we can change the SSL_CIPHER
10884 definition to split the single 'unsigned long mask' bitmap into
10885 multiple values to extend the available space.
10889 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10891 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10892 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10894 * Add AES IGE and biIGE modes.
10898 * Change the Unix randomness entropy gathering to use poll() when
10899 possible instead of select(), since the latter has some
10900 undesirable limitations.
10902 *Darryl Miles via Richard Levitte and Bodo Moeller*
10904 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10905 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10906 cannot be implicitly activated as part of, e.g., the "AES" alias.
10907 However, please upgrade to OpenSSL 0.9.9[-dev] for
10908 non-experimental use of the ECC ciphersuites to get TLS extension
10909 support, which is required for curve and point format negotiation
10910 to avoid potential handshake problems.
10914 * Disable rogue ciphersuites:
10916 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10917 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10918 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10920 The latter two were purportedly from
10921 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10924 Also deactivate the remaining ciphersuites from
10925 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10926 unofficial, and the ID has long expired.
10930 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10931 dual-core machines) and other potential thread-safety issues.
10935 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10936 versions), which is now available for royalty-free use
10937 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10938 Also, add Camellia TLS ciphersuites from RFC 4132.
10940 To minimize changes between patchlevels in the OpenSSL 0.9.8
10941 series, Camellia remains excluded from compilation unless OpenSSL
10942 is configured with 'enable-camellia'.
10946 * Disable the padding bug check when compression is in use. The padding
10947 bug check assumes the first packet is of even length, this is not
10948 necessarily true if compression is enabled and can result in false
10949 positives causing handshake failure. The actual bug test is ancient
10950 code so it is hoped that implementations will either have fixed it by
10951 now or any which still have the bug do not support compression.
10955 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10957 * When applying a cipher rule check to see if string match is an explicit
10958 cipher suite and only match that one cipher suite if it is.
10962 * Link in manifests for VC++ if needed.
10964 *Austin Ziegler <halostatue@gmail.com>*
10966 * Update support for ECC-based TLS ciphersuites according to
10967 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10968 TLS extensions, which are supported starting with the 0.9.9
10969 branch, not in the OpenSSL 0.9.8 branch).
10973 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10974 opaque EVP_CIPHER_CTX handling.
10978 * Fixes and enhancements to zlib compression code. We now only use
10979 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10980 to conform with the standards mentioned here:
10981 <http://www.zlib.net/DLL_FAQ.txt>
10982 Static zlib linking now works on Windows and the new --with-zlib-include
10983 --with-zlib-lib options to Configure can be used to supply the location
10984 of the headers and library. Gracefully handle case where zlib library
10989 * Several fixes and enhancements to the OID generation code. The old code
10990 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10991 handle numbers larger than ULONG_MAX, truncated printing and had a
10992 non standard OBJ_obj2txt() behaviour.
10996 * Add support for building of engines under engine/ as shared libraries
10997 under VC++ build system.
11001 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11002 Hopefully, we will not see any false combination of paths any more.
11006 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11008 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11009 (part of SSL_OP_ALL). This option used to disable the
11010 countermeasure against man-in-the-middle protocol-version
11011 rollback in the SSL 2.0 server implementation, which is a bad
11012 idea. ([CVE-2005-2969])
11014 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11015 for Information Security, National Institute of Advanced Industrial
11016 Science and Technology [AIST], Japan)*
11018 * Add two function to clear and return the verify parameter flags.
11022 * Keep cipherlists sorted in the source instead of sorting them at
11023 runtime, thus removing the need for a lock.
11027 * Avoid some small subgroup attacks in Diffie-Hellman.
11029 *Nick Mathewson and Ben Laurie*
11031 * Add functions for well-known primes.
11035 * Extended Windows CE support.
11037 *Satoshi Nakamura and Andy Polyakov*
11039 * Initialize SSL_METHOD structures at compile time instead of during
11040 runtime, thus removing the need for a lock.
11044 * Make PKCS7_decrypt() work even if no certificate is supplied by
11045 attempting to decrypt each encrypted key in turn. Add support to
11050 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11052 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11055 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11059 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11060 key into the same file any more.
11064 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11068 * Add -utf8 command line and config file option to 'ca'.
11070 *Stefan <stf@udoma.org*
11072 * Removed the macro des_crypt(), as it seems to conflict with some
11073 libraries. Use DES_crypt().
11077 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11078 involves renaming the source and generated shared-libs for
11079 both. The engines will accept the corrected or legacy ids
11080 ('ncipher' and '4758_cca' respectively) when binding. NB,
11081 this only applies when building 'shared'.
11083 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11085 * Add attribute functions to EVP_PKEY structure. Modify
11086 PKCS12_create() to recognize a CSP name attribute and
11087 use it. Make -CSP option work again in pkcs12 utility.
11091 * Add new functionality to the bn blinding code:
11092 - automatic re-creation of the BN_BLINDING parameters after
11093 a fixed number of uses (currently 32)
11094 - add new function for parameter creation
11095 - introduce flags to control the update behaviour of the
11096 BN_BLINDING parameters
11097 - hide BN_BLINDING structure
11098 Add a second BN_BLINDING slot to the RSA structure to improve
11099 performance when a single RSA object is shared among several
11104 * Add support for DTLS.
11106 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11108 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11109 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11113 * Remove buggy and incomplete DH cert support from
11114 ssl/ssl_rsa.c and ssl/s3_both.c
11118 * Use SHA-1 instead of MD5 as the default digest algorithm for
11119 the `apps/openssl` commands.
11123 * Compile clean with "-Wall -Wmissing-prototypes
11124 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11125 DEBUG_SAFESTACK must also be set.
11129 * Change ./Configure so that certain algorithms can be disabled by default.
11130 The new counterpiece to "no-xxx" is "enable-xxx".
11132 The patented RC5 and MDC2 algorithms will now be disabled unless
11133 "enable-rc5" and "enable-mdc2", respectively, are specified.
11135 (IDEA remains enabled despite being patented. This is because IDEA
11136 is frequently required for interoperability, and there is no license
11137 fee for non-commercial use. As before, "no-idea" can be used to
11138 avoid this algorithm.)
11142 * Add processing of proxy certificates (see RFC 3820). This work was
11143 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11144 EGEE (Enabling Grids for E-science in Europe).
11148 * RC4 performance overhaul on modern architectures/implementations, such
11149 as Intel P4, IA-64 and AMD64.
11153 * New utility extract-section.pl. This can be used specify an alternative
11154 section number in a pod file instead of having to treat each file as
11155 a separate case in Makefile. This can be done by adding two lines to the
11158 =for comment openssl_section:XXX
11160 The blank line is mandatory.
11164 * New arguments -certform, -keyform and -pass for s_client and s_server
11165 to allow alternative format key and certificate files and passphrase
11170 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11171 update associated structures and add various utility functions.
11173 Add new policy related verify parameters, include policy checking in
11174 standard verify code. Enhance 'smime' application with extra parameters
11175 to support policy checking and print out.
11179 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11180 Nehemiah processors. These extensions support AES encryption in hardware
11181 as well as RNG (though RNG support is currently disabled).
11183 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11185 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11189 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11191 *Andy Polyakov and a number of other people*
11193 * Improved PowerPC platform support. Most notably BIGNUM assembler
11194 implementation contributed by IBM.
11196 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11198 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11199 exponent rather than 'unsigned long'. There is a corresponding change to
11200 the new 'rsa_keygen' element of the RSA_METHOD structure.
11202 *Jelte Jansen, Geoff Thorpe*
11204 * Functionality for creating the initial serial number file is now
11205 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11207 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11208 number file to 1, which is bound to cause problems. To avoid
11209 the problems while respecting compatibility between different 0.9.7
11210 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11211 CA.pl for serial number initialization. With the new release 0.9.8,
11212 we can fix the problem directly in the 'ca' utility.)
11216 * Reduced header interdependencies by declaring more opaque objects in
11217 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11218 give fewer recursive includes, which could break lazy source code - so
11219 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11220 developers should define this symbol when building and using openssl to
11221 ensure they track the recommended behaviour, interfaces, [etc], but
11222 backwards-compatible behaviour prevails when this isn't defined.
11226 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11230 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11231 This will generate a random key of the appropriate length based on the
11232 cipher context. The EVP_CIPHER can provide its own random key generation
11233 routine to support keys of a specific form. This is used in the des and
11234 3des routines to generate a key of the correct parity. Update S/MIME
11235 code to use new functions and hence generate correct parity DES keys.
11236 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11237 valid (weak or incorrect parity).
11241 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11242 as looking them up. This is useful when the verified structure may contain
11243 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11244 present unless the new PKCS7_NO_CRL flag is asserted.
11248 * Extend ASN1 oid configuration module. It now additionally accepts the
11251 shortName = some long name, 1.2.3.4
11255 * Reimplemented the BN_CTX implementation. There is now no more static
11256 limitation on the number of variables it can handle nor the depth of the
11257 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11258 information can now expand as required, and rather than having a single
11259 static array of bignums, BN_CTX now uses a linked-list of such arrays
11260 allowing it to expand on demand whilst maintaining the usefulness of
11261 BN_CTX's "bundling".
11265 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11266 to allow all RSA operations to function using a single BN_CTX.
11270 * Preliminary support for certificate policy evaluation and checking. This
11271 is initially intended to pass the tests outlined in "Conformance Testing
11272 of Relying Party Client Certificate Path Processing Logic" v1.07.
11276 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11277 remained unused and not that useful. A variety of other little bignum
11278 tweaks and fixes have also been made continuing on from the audit (see
11283 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11284 associated ASN1, EVP and SSL functions and old ASN1 macros.
11288 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11289 and this should never fail. So the return value from the use of
11290 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11291 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11295 * BN_CTX_get() should return zero-valued bignums, providing the same
11296 initialised value as BN_new().
11298 *Geoff Thorpe, suggested by Ulf Möller*
11300 * Support for inhibitAnyPolicy certificate extension.
11304 * An audit of the BIGNUM code is underway, for which debugging code is
11305 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11306 is considered valid when processing BIGNUMs, and causes execution to
11307 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11308 further steps are taken to deliberately pollute unused data in BIGNUM
11309 structures to try and expose faulty code further on. For now, openssl will
11310 (in its default mode of operation) continue to tolerate the inconsistent
11311 forms that it has tolerated in the past, but authors and packagers should
11312 consider trying openssl and their own applications when compiled with
11313 these debugging symbols defined. It will help highlight potential bugs in
11314 their own code, and will improve the test coverage for OpenSSL itself. At
11315 some point, these tighter rules will become openssl's default to improve
11316 maintainability, though the assert()s and other overheads will remain only
11317 in debugging configurations. See bn.h for more details.
11319 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11321 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11322 that can only be obtained through BN_CTX_new() (which implicitly
11323 initialises it). The presence of this function only made it possible
11324 to overwrite an existing structure (and cause memory leaks).
11328 * Because of the callback-based approach for implementing LHASH as a
11329 template type, lh_insert() adds opaque objects to hash-tables and
11330 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11331 to clean up those corresponding objects before destroying the hash table
11332 (and losing the object pointers). So some over-zealous constifications in
11333 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11334 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11335 prototyped to have "const" restrictions on the object pointers they are
11336 given (and so aren't required to cast them away any more).
11340 * The tmdiff.h API was so ugly and minimal that our own timing utility
11341 (speed) prefers to use its own implementation. The two implementations
11342 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11343 its object type properly exposed (MS_TM) instead of casting to/from
11344 `char *`. This may still change yet if someone realises MS_TM and
11346 aren't necessarily the greatest nomenclatures - but this is what was used
11347 internally to the implementation so I've used that for now.
11351 * Ensure that deprecated functions do not get compiled when
11352 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11353 the self-tests were still using deprecated key-generation functions so
11354 these have been updated also.
11358 * Reorganise PKCS#7 code to separate the digest location functionality
11359 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11360 New function PKCS7_set_digest() to set the digest type for PKCS#7
11361 digestedData type. Add additional code to correctly generate the
11362 digestedData type and add support for this type in PKCS7 initialization
11367 * New function PKCS7_set0_type_other() this initializes a PKCS7
11368 structure of type "other".
11372 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11373 sure the loop does correctly stop and breaking ("division by zero")
11374 modulus operations are not performed. The (pre-generated) prime
11375 table crypto/bn/bn_prime.h was already correct, but it could not be
11376 re-generated on some platforms because of the "division by zero"
11377 situation in the script.
11379 *Ralf S. Engelschall*
11381 * Update support for ECC-based TLS ciphersuites according to
11382 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11383 SHA-1 now is only used for "small" curves (where the
11384 representation of a field element takes up to 24 bytes); for
11385 larger curves, the field element resulting from ECDH is directly
11386 used as premaster secret.
11388 *Douglas Stebila (Sun Microsystems Laboratories)*
11390 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11391 curve secp160r1 to the tests.
11393 *Douglas Stebila (Sun Microsystems Laboratories)*
11395 * Add the possibility to load symbols globally with DSO.
11397 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11399 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11400 control of the error stack.
11404 * Add support for STORE in ENGINE.
11408 * Add the STORE type. The intention is to provide a common interface
11409 to certificate and key stores, be they simple file-based stores, or
11410 HSM-type store, or LDAP stores, or...
11411 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11415 * Add a generic structure called OPENSSL_ITEM. This can be used to
11416 pass a list of arguments to any function as well as provide a way
11417 for a function to pass data back to the caller.
11421 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11422 works like BUF_strdup() but can be used to duplicate a portion of
11423 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11428 * Add the function sk_find_ex() which works like sk_find(), but will
11429 return an index to an element even if an exact match couldn't be
11430 found. The index is guaranteed to point at the element where the
11431 searched-for key would be inserted to preserve sorting order.
11435 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11436 takes an extra flags argument for optional functionality. Currently,
11437 the following flags are defined:
11439 OBJ_BSEARCH_VALUE_ON_NOMATCH
11440 This one gets OBJ_bsearch_ex() to return a pointer to the first
11441 element where the comparing function returns a negative or zero
11444 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11445 This one gets OBJ_bsearch_ex() to return a pointer to the first
11446 element where the comparing function returns zero. This is useful
11447 if there are more than one element where the comparing function
11452 * Make it possible to create self-signed certificates with 'openssl ca'
11453 in such a way that the self-signed certificate becomes part of the
11454 CA database and uses the same mechanisms for serial number generation
11455 as all other certificate signing. The new flag '-selfsign' enables
11456 this functionality. Adapt CA.sh and CA.pl.in.
11460 * Add functionality to check the public key of a certificate request
11461 against a given private. This is useful to check that a certificate
11462 request can be signed by that key (self-signing).
11466 * Make it possible to have multiple active certificates with the same
11467 subject in the CA index file. This is done only if the keyword
11468 'unique_subject' is set to 'no' in the main CA section (default
11469 if 'CA_default') of the configuration file. The value is saved
11470 with the database itself in a separate index attribute file,
11471 named like the index file with '.attr' appended to the name.
11475 * Generate multi-valued AVAs using '+' notation in config files for
11480 * Support for nameConstraints certificate extension.
11484 * Support for policyConstraints certificate extension.
11488 * Support for policyMappings certificate extension.
11492 * Make sure the default DSA_METHOD implementation only uses its
11493 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11494 and change its own handlers to be NULL so as to remove unnecessary
11495 indirection. This lets alternative implementations fallback to the
11496 default implementation more easily.
11500 * Support for directoryName in GeneralName related extensions
11505 * Make it possible to link applications using Makefile.shared.
11506 Make that possible even when linking against static libraries!
11510 * Support for single pass processing for S/MIME signing. This now
11511 means that S/MIME signing can be done from a pipe, in addition
11512 cleartext signing (multipart/signed type) is effectively streaming
11513 and the signed data does not need to be all held in memory.
11515 This is done with a new flag PKCS7_STREAM. When this flag is set
11516 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11517 is done after the data is output (and digests calculated) in
11518 SMIME_write_PKCS7().
11522 * Add full support for -rpath/-R, both in shared libraries and
11523 applications, at least on the platforms where it's known how
11528 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11529 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11530 will now compute a table of multiples of the generator that
11531 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11532 faster (notably in the case of a single point multiplication,
11533 scalar * generator).
11535 *Nils Larsch, Bodo Moeller*
11537 * IPv6 support for certificate extensions. The various extensions
11538 which use the IP:a.b.c.d can now take IPv6 addresses using the
11539 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11544 * Added an ENGINE that implements RSA by performing private key
11545 exponentiations with the GMP library. The conversions to and from
11546 GMP's mpz_t format aren't optimised nor are any montgomery forms
11547 cached, and on x86 it appears OpenSSL's own performance has caught up.
11548 However there are likely to be other architectures where GMP could
11549 provide a boost. This ENGINE is not built in by default, but it can be
11550 specified at Configure time and should be accompanied by the necessary
11551 linker additions, eg;
11552 ./config -DOPENSSL_USE_GMP -lgmp
11556 * "openssl engine" will not display ENGINE/DSO load failure errors when
11557 testing availability of engines with "-t" - the old behaviour is
11558 produced by increasing the feature's verbosity with "-tt".
11562 * ECDSA routines: under certain error conditions uninitialized BN objects
11563 could be freed. Solution: make sure initialization is performed early
11564 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11569 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11570 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11571 software implementations. For DSA and DH, parameter generation can
11572 also be overridden by providing the appropriate method callbacks.
11576 * Change the "progress" mechanism used in key-generation and
11577 primality testing to functions that take a new BN_GENCB pointer in
11578 place of callback/argument pairs. The new API functions have `_ex`
11579 postfixes and the older functions are reimplemented as wrappers for
11580 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11581 declarations of the old functions to help (graceful) attempts to
11582 migrate to the new functions. Also, the new key-generation API
11583 functions operate on a caller-supplied key-structure and return
11584 success/failure rather than returning a key or NULL - this is to
11585 help make "keygen" another member function of RSA_METHOD etc.
11587 Example for using the new callback interface:
11589 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11590 void *my_arg = ...;
11593 BN_GENCB_set(&my_cb, my_callback, my_arg);
11595 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11596 /* For the meaning of a, b in calls to my_callback(), see the
11597 * documentation of the function that calls the callback.
11598 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11599 * my_callback should return 1 if it wants BN_is_prime_ex()
11600 * to continue, or 0 to stop.
11605 * Change the ZLIB compression method to be stateful, and make it
11606 available to TLS with the number defined in
11607 draft-ietf-tls-compression-04.txt.
11611 * Add the ASN.1 structures and functions for CertificatePair, which
11612 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11614 CertificatePair ::= SEQUENCE {
11615 forward [0] Certificate OPTIONAL,
11616 reverse [1] Certificate OPTIONAL,
11617 -- at least one of the pair shall be present -- }
11619 Also implement the PEM functions to read and write certificate
11620 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11622 This needed to be defined, mostly for the sake of the LDAP
11623 attribute crossCertificatePair, but may prove useful elsewhere as
11628 * Make it possible to inhibit symlinking of shared libraries in
11629 Makefile.shared, for Cygwin's sake.
11633 * Extend the BIGNUM API by creating a function
11634 void BN_set_negative(BIGNUM *a, int neg);
11635 and a macro that behave like
11636 int BN_is_negative(const BIGNUM *a);
11638 to avoid the need to access 'a->neg' directly in applications.
11642 * Implement fast modular reduction for pseudo-Mersenne primes
11643 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11644 EC_GROUP_new_curve_GFp() will now automatically use this
11647 *Nils Larsch <nla@trustcenter.de>*
11649 * Add new lock type (CRYPTO_LOCK_BN).
11653 * Change the ENGINE framework to automatically load engines
11654 dynamically from specific directories unless they could be
11655 found to already be built in or loaded. Move all the
11656 current engines except for the cryptodev one to a new
11657 directory engines/.
11658 The engines in engines/ are built as shared libraries if
11659 the "shared" options was given to ./Configure or ./config.
11660 Otherwise, they are inserted in libcrypto.a.
11661 /usr/local/ssl/engines is the default directory for dynamic
11662 engines, but that can be overridden at configure time through
11663 the usual use of --prefix and/or --openssldir, and at run
11664 time with the environment variable OPENSSL_ENGINES.
11666 *Geoff Thorpe and Richard Levitte*
11668 * Add Makefile.shared, a helper makefile to build shared
11669 libraries. Adapt Makefile.org.
11673 * Add version info to Win32 DLLs.
11675 *Peter 'Luna' Runestig" <peter@runestig.com>*
11677 * Add new 'medium level' PKCS#12 API. Certificates and keys
11678 can be added using this API to created arbitrary PKCS#12
11679 files while avoiding the low-level API.
11681 New options to PKCS12_create(), key or cert can be NULL and
11682 will then be omitted from the output file. The encryption
11683 algorithm NIDs can be set to -1 for no encryption, the mac
11684 iteration count can be set to 0 to omit the mac.
11686 Enhance pkcs12 utility by making the -nokeys and -nocerts
11687 options work when creating a PKCS#12 file. New option -nomac
11688 to omit the mac, NONE can be set for an encryption algorithm.
11689 New code is modified to use the enhanced PKCS12_create()
11690 instead of the low-level API.
11694 * Extend ASN1 encoder to support indefinite length constructed
11695 encoding. This can output sequences tags and octet strings in
11696 this form. Modify pk7_asn1.c to support indefinite length
11697 encoding. This is experimental and needs additional code to
11698 be useful, such as an ASN1 bio and some enhanced streaming
11701 Extend template encode functionality so that tagging is passed
11702 down to the template encoder.
11706 * Let 'openssl req' fail if an argument to '-newkey' is not
11707 recognized instead of using RSA as a default.
11711 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11712 As these are not official, they are not included in "ALL";
11713 the "ECCdraft" ciphersuite group alias can be used to select them.
11715 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11717 * Add ECDH engine support.
11719 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11721 * Add ECDH in new directory crypto/ecdh/.
11723 *Douglas Stebila (Sun Microsystems Laboratories)*
11725 * Let BN_rand_range() abort with an error after 100 iterations
11726 without success (which indicates a broken PRNG).
11730 * Change BN_mod_sqrt() so that it verifies that the input value
11731 is really the square of the return value. (Previously,
11732 BN_mod_sqrt would show GIGO behaviour.)
11736 * Add named elliptic curves over binary fields from X9.62, SECG,
11737 and WAP/WTLS; add OIDs that were still missing.
11739 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11741 * Extend the EC library for elliptic curves over binary fields
11742 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11745 EC_GF2m_simple_method
11749 EC_GROUP_new_curve_GF2m
11750 EC_GROUP_set_curve_GF2m
11751 EC_GROUP_get_curve_GF2m
11752 EC_POINT_set_affine_coordinates_GF2m
11753 EC_POINT_get_affine_coordinates_GF2m
11754 EC_POINT_set_compressed_coordinates_GF2m
11756 Point compression for binary fields is disabled by default for
11757 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11760 As binary polynomials are represented as BIGNUMs, various members
11761 of the EC_GROUP and EC_POINT data structures can be shared
11762 between the implementations for prime fields and binary fields;
11763 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11764 are essentially identical to their `..._GFp` counterparts.
11765 (For simplicity, the `..._GFp` prefix has been dropped from
11766 various internal method names.)
11768 An internal 'field_div' method (similar to 'field_mul' and
11769 'field_sqr') has been added; this is used only for binary fields.
11771 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11773 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11774 through methods ('mul', 'precompute_mult').
11776 The generic implementations (now internally called 'ec_wNAF_mul'
11777 and 'ec_wNAF_precomputed_mult') remain the default if these
11778 methods are undefined.
11780 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11782 * New function EC_GROUP_get_degree, which is defined through
11783 EC_METHOD. For curves over prime fields, this returns the bit
11784 length of the modulus.
11786 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11788 * New functions EC_GROUP_dup, EC_POINT_dup.
11789 (These simply call ..._new and ..._copy).
11791 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11793 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11794 Polynomials are represented as BIGNUMs (where the sign bit is not
11795 used) in the following functions [macros]:
11798 BN_GF2m_sub [= BN_GF2m_add]
11799 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11800 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11801 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11803 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11804 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11805 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11806 BN_GF2m_cmp [= BN_ucmp]
11808 (Note that only the 'mod' functions are actually for fields GF(2^m).
11809 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11811 For some functions, an the irreducible polynomial defining a
11812 field can be given as an 'unsigned int[]' with strictly
11813 decreasing elements giving the indices of those bits that are set;
11814 i.e., p[] represents the polynomial
11815 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11817 p[0] > p[1] > ... > p[k] = 0.
11818 This applies to the following functions:
11821 BN_GF2m_mod_mul_arr
11822 BN_GF2m_mod_sqr_arr
11823 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11824 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11825 BN_GF2m_mod_exp_arr
11826 BN_GF2m_mod_sqrt_arr
11827 BN_GF2m_mod_solve_quad_arr
11831 Conversion can be performed by the following functions:
11836 bntest.c has additional tests for binary polynomial arithmetic.
11838 Two implementations for BN_GF2m_mod_div() are available.
11839 The default algorithm simply uses BN_GF2m_mod_inv() and
11840 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11841 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11842 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11844 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11846 * Add new error code 'ERR_R_DISABLED' that can be used when some
11847 functionality is disabled at compile-time.
11849 *Douglas Stebila <douglas.stebila@sun.com>*
11851 * Change default behaviour of 'openssl asn1parse' so that more
11852 information is visible when viewing, e.g., a certificate:
11854 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11855 mode the content of non-printable OCTET STRINGs is output in a
11856 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11857 avoid the appearance of a printable string.
11859 *Nils Larsch <nla@trustcenter.de>*
11861 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11863 EC_GROUP_set_asn1_flag()
11864 EC_GROUP_get_asn1_flag()
11865 EC_GROUP_set_point_conversion_form()
11866 EC_GROUP_get_point_conversion_form()
11867 These control ASN1 encoding details:
11868 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11869 has been set to OPENSSL_EC_NAMED_CURVE.
11870 - Points are encoded in uncompressed form by default; options for
11871 asn1_for are as for point2oct, namely
11872 POINT_CONVERSION_COMPRESSED
11873 POINT_CONVERSION_UNCOMPRESSED
11874 POINT_CONVERSION_HYBRID
11876 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11878 EC_GROUP_set_seed()
11879 EC_GROUP_get0_seed()
11880 EC_GROUP_get_seed_len()
11881 This is used only for ASN1 purposes (so far).
11883 *Nils Larsch <nla@trustcenter.de>*
11885 * Add 'field_type' member to EC_METHOD, which holds the NID
11886 of the appropriate field type OID. The new function
11887 EC_METHOD_get_field_type() returns this value.
11889 *Nils Larsch <nla@trustcenter.de>*
11892 EC_POINT_point2bn()
11893 EC_POINT_bn2point()
11894 EC_POINT_point2hex()
11895 EC_POINT_hex2point()
11896 providing useful interfaces to EC_POINT_point2oct() and
11897 EC_POINT_oct2point().
11899 *Nils Larsch <nla@trustcenter.de>*
11901 * Change internals of the EC library so that the functions
11902 EC_GROUP_set_generator()
11903 EC_GROUP_get_generator()
11904 EC_GROUP_get_order()
11905 EC_GROUP_get_cofactor()
11906 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11907 to methods, which would lead to unnecessary code duplication when
11908 adding different types of curves.
11910 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11912 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11913 arithmetic, and such that modified wNAFs are generated
11914 (which avoid length expansion in many cases).
11918 * Add a function EC_GROUP_check_discriminant() (defined via
11919 EC_METHOD) that verifies that the curve discriminant is non-zero.
11921 Add a function EC_GROUP_check() that makes some sanity tests
11922 on a EC_GROUP, its generator and order. This includes
11923 EC_GROUP_check_discriminant().
11925 *Nils Larsch <nla@trustcenter.de>*
11927 * Add ECDSA in new directory crypto/ecdsa/.
11929 Add applications 'openssl ecparam' and 'openssl ecdsa'
11930 (these are based on 'openssl dsaparam' and 'openssl dsa').
11932 ECDSA support is also included in various other files across the
11933 library. Most notably,
11934 - 'openssl req' now has a '-newkey ecdsa:file' option;
11935 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11936 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11937 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11938 them suitable for ECDSA where domain parameters must be
11939 extracted before the specific public key;
11940 - ECDSA engine support has been added.
11942 *Nils Larsch <nla@trustcenter.de>*
11944 * Include some named elliptic curves, and add OIDs from X9.62,
11945 SECG, and WAP/WTLS. Each curve can be obtained from the new
11947 EC_GROUP_new_by_curve_name(),
11948 and the list of available named curves can be obtained with
11949 EC_get_builtin_curves().
11950 Also add a 'curve_name' member to EC_GROUP objects, which can be
11952 EC_GROUP_set_curve_name()
11953 EC_GROUP_get_curve_name()
11955 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11957 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11958 was actually never needed) and in BN_mul(). The removal in BN_mul()
11959 required a small change in bn_mul_part_recursive() and the addition
11960 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11961 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11962 bn_sub_words() and bn_add_words() except they take arrays with
11967 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11969 * Cleanse PEM buffers before freeing them since they may contain
11972 *Benjamin Bennett <ben@psc.edu>*
11974 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11975 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11976 authentication-only ciphersuites.
11980 * Since AES128 and AES256 share a single mask bit in the logic of
11981 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11982 kludge to work properly if AES128 is available and AES256 isn't.
11986 * Expand security boundary to match 1.1.1 module.
11990 * Remove redundant features: hash file source, editing of test vectors
11991 modify fipsld to use external fips_premain.c signature.
11995 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11996 run algorithm test programs.
12000 * Make algorithm test programs more tolerant of whitespace.
12004 * Have SSL/TLS server implementation tolerate "mismatched" record
12005 protocol version while receiving ClientHello even if the
12006 ClientHello is fragmented. (The server can't insist on the
12007 particular protocol version it has chosen before the ServerHello
12008 message has informed the client about his choice.)
12012 * Load error codes if they are not already present instead of using a
12013 static variable. This allows them to be cleanly unloaded and reloaded.
12017 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12019 * Introduce limits to prevent malicious keys being able to
12020 cause a denial of service. ([CVE-2006-2940])
12022 *Steve Henson, Bodo Moeller*
12024 * Fix ASN.1 parsing of certain invalid structures that can result
12025 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12027 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12028 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12030 * Fix SSL client code which could crash if connecting to a
12031 malicious SSLv2 server. ([CVE-2006-4343])
12033 *Tavis Ormandy and Will Drewry, Google Security Team*
12035 * Change ciphersuite string processing so that an explicit
12036 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12037 will no longer include "AES128-SHA"), and any other similar
12038 ciphersuite (same bitmap) from *other* protocol versions (so that
12039 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12040 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12041 changes from 0.9.8b and 0.9.8d.
12045 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12047 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12048 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12050 * Change the Unix randomness entropy gathering to use poll() when
12051 possible instead of select(), since the latter has some
12052 undesirable limitations.
12054 *Darryl Miles via Richard Levitte and Bodo Moeller*
12056 * Disable rogue ciphersuites:
12058 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12059 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12060 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12062 The latter two were purportedly from
12063 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12066 Also deactivate the remaining ciphersuites from
12067 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12068 unofficial, and the ID has long expired.
12072 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12073 dual-core machines) and other potential thread-safety issues.
12077 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12079 * Adapt fipsld and the build system to link against the validated FIPS
12080 module in FIPS mode.
12084 * Fixes for VC++ 2005 build under Windows.
12088 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12089 from a Windows bash shell such as MSYS. It is autodetected from the
12090 "config" script when run from a VC++ environment. Modify standard VC++
12091 build to use fipscanister.o from the GNU make build.
12095 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12097 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12098 The value now differs depending on if you build for FIPS or not.
12099 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12100 safely run with a non-FIPSed libcrypto, as it may crash because of
12101 the difference induced by this change.
12105 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12107 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12108 (part of SSL_OP_ALL). This option used to disable the
12109 countermeasure against man-in-the-middle protocol-version
12110 rollback in the SSL 2.0 server implementation, which is a bad
12111 idea. ([CVE-2005-2969])
12113 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12114 for Information Security, National Institute of Advanced Industrial
12115 Science and Technology [AIST, Japan)]*
12117 * Minimal support for X9.31 signatures and PSS padding modes. This is
12118 mainly for FIPS compliance and not fully integrated at this stage.
12122 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12123 the exponentiation using a fixed-length exponent. (Otherwise,
12124 the information leaked through timing could expose the secret key
12125 after many signatures; cf. Bleichenbacher's attack on DSA with
12130 * Make a new fixed-window mod_exp implementation the default for
12131 RSA, DSA, and DH private-key operations so that the sequence of
12132 squares and multiplies and the memory access pattern are
12133 independent of the particular secret key. This will mitigate
12134 cache-timing and potential related attacks.
12136 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12137 and this is automatically used by BN_mod_exp_mont() if the new flag
12138 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12139 will use this BN flag for private exponents unless the flag
12140 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12141 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12143 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12145 * Change the client implementation for SSLv23_method() and
12146 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12147 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12148 (Previously, the SSL 2.0 backwards compatible Client Hello
12149 message format would be used even with SSL_OP_NO_SSLv2.)
12153 * Add support for smime-type MIME parameter in S/MIME messages which some
12158 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12159 a threadsafe manner. Modify rsa code to use new function and add calls
12160 to dsa and dh code (which had race conditions before).
12164 * Include the fixed error library code in the C error file definitions
12165 instead of fixing them up at runtime. This keeps the error code
12166 structures constant.
12170 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12172 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12175 * Fixes for newer kerberos headers. NB: the casts are needed because
12176 the 'length' field is signed on one version and unsigned on another
12177 with no (?) obvious way to tell the difference, without these VC++
12178 complains. Also the "definition" of FAR (blank) is no longer included
12179 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12180 some needed definitions.
12184 * Undo Cygwin change.
12188 * Added support for proxy certificates according to RFC 3820.
12189 Because they may be a security thread to unaware applications,
12190 they must be explicitly allowed in run-time. See
12191 docs/HOWTO/proxy_certificates.txt for further information.
12195 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12197 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12198 server and client random values. Previously
12199 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12200 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12202 This change has negligible security impact because:
12204 1. Server and client random values still have 24 bytes of pseudo random
12207 2. Server and client random values are sent in the clear in the initial
12210 3. The master secret is derived using the premaster secret (48 bytes in
12211 size for static RSA ciphersuites) as well as client server and random
12214 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12217 *Stephen Henson, reported by UK NISCC*
12219 * Use Windows randomness collection on Cygwin.
12223 * Fix hang in EGD/PRNGD query when communication socket is closed
12224 prematurely by EGD/PRNGD.
12226 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12228 * Prompt for pass phrases when appropriate for PKCS12 input format.
12232 * Back-port of selected performance improvements from development
12233 branch, as well as improved support for PowerPC platforms.
12237 * Add lots of checks for memory allocation failure, error codes to indicate
12238 failure and freeing up memory if a failure occurs.
12240 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12242 * Add new -passin argument to dgst.
12246 * Perform some character comparisons of different types in X509_NAME_cmp:
12247 this is needed for some certificates that re-encode DNs into UTF8Strings
12248 (in violation of RFC3280) and can't or won't issue name rollover
12253 * Make an explicit check during certificate validation to see that
12254 the CA setting in each certificate on the chain is correct. As a
12255 side effect always do the following basic checks on extensions,
12256 not just when there's an associated purpose to the check:
12258 - if there is an unhandled critical extension (unless the user
12259 has chosen to ignore this fault)
12260 - if the path length has been exceeded (if one is set at all)
12261 - that certain extensions fit the associated purpose (if one has
12266 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12268 * Avoid a race condition when CRLs are checked in a multi threaded
12269 environment. This would happen due to the reordering of the revoked
12270 entries during signature checking and serial number lookup. Now the
12271 encoding is cached and the serial number sort performed under a lock.
12272 Add new STACK function sk_is_sorted().
12276 * Add Delta CRL to the extension code.
12280 * Various fixes to s3_pkt.c so alerts are sent properly.
12282 *David Holmes <d.holmes@f5.com>*
12284 * Reduce the chances of duplicate issuer name and serial numbers (in
12285 violation of RFC3280) using the OpenSSL certificate creation utilities.
12286 This is done by creating a random 64 bit value for the initial serial
12287 number when a serial number file is created or when a self signed
12288 certificate is created using 'openssl req -x509'. The initial serial
12289 number file is created using 'openssl x509 -next_serial' in CA.pl
12290 rather than being initialized to 1.
12294 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12296 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12297 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12299 *Joe Orton, Steve Henson*
12301 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12304 *Joe Orton, Steve Henson*
12306 * Make it possible to have multiple active certificates with the same
12307 subject in the CA index file. This is done only if the keyword
12308 'unique_subject' is set to 'no' in the main CA section (default
12309 if 'CA_default') of the configuration file. The value is saved
12310 with the database itself in a separate index attribute file,
12311 named like the index file with '.attr' appended to the name.
12315 * X509 verify fixes. Disable broken certificate workarounds when
12316 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12317 keyUsage extension present. Don't accept CRLs with unhandled critical
12318 extensions: since verify currently doesn't process CRL extensions this
12319 rejects a CRL with *any* critical extensions. Add new verify error codes
12324 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12325 A clarification of RFC2560 will require the use of OCTET STRINGs and
12326 some implementations cannot handle the current raw format. Since OpenSSL
12327 copies and compares OCSP nonces as opaque blobs without any attempt at
12328 parsing them this should not create any compatibility issues.
12332 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12333 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12334 this HMAC (and other) operations are several times slower than OpenSSL
12339 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12341 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12343 * Use the correct content when signing type "other".
12347 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12349 * Fix various bugs revealed by running the NISCC test suite:
12351 Stop out of bounds reads in the ASN1 code when presented with
12352 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12354 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12356 If verify callback ignores invalid public key errors don't try to check
12357 certificate signature with the NULL public key.
12361 * New -ignore_err option in ocsp application to stop the server
12362 exiting on the first error in a request.
12366 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12367 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12372 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12373 extra data after the compression methods not only for TLS 1.0
12374 but also for SSL 3.0 (as required by the specification).
12376 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12378 * Change X509_certificate_type() to mark the key as exported/exportable
12379 when it's 512 *bits* long, not 512 bytes.
12383 * Change AES_cbc_encrypt() so it outputs exact multiple of
12384 blocks during encryption.
12388 * Various fixes to base64 BIO and non blocking I/O. On write
12389 flushes were not handled properly if the BIO retried. On read
12390 data was not being buffered properly and had various logic bugs.
12391 This also affects blocking I/O when the data being decoded is a
12396 * Various S/MIME bugfixes and compatibility changes:
12397 output correct application/pkcs7 MIME type if
12398 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12399 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12400 of files as .eml work). Correctly handle very long lines in MIME
12405 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12407 * Countermeasure against the Klima-Pokorny-Rosa extension of
12408 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12409 a protocol version number mismatch like a decryption error
12410 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12414 * Turn on RSA blinding by default in the default implementation
12415 to avoid a timing attack. Applications that don't want it can call
12416 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12417 They would be ill-advised to do so in most cases.
12419 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12421 * Change RSA blinding code so that it works when the PRNG is not
12422 seeded (in this case, the secret RSA exponent is abused as
12423 an unpredictable seed -- if it is not unpredictable, there
12424 is no point in blinding anyway). Make RSA blinding thread-safe
12425 by remembering the creator's thread ID in rsa->blinding and
12426 having all other threads use local one-time blinding factors
12427 (this requires more computation than sharing rsa->blinding, but
12428 avoids excessive locking; and if an RSA object is not shared
12429 between threads, blinding will still be very fast).
12433 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12434 ENGINE as defaults for all supported algorithms irrespective of
12435 the 'flags' parameter. 'flags' is now honoured, so applications
12436 should make sure they are passing it correctly.
12440 * Target "mingw" now allows native Windows code to be generated in
12441 the Cygwin environment as well as with the MinGW compiler.
12445 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12447 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12448 via timing by performing a MAC computation even if incorrect
12449 block cipher padding has been found. This is a countermeasure
12450 against active attacks where the attacker has to distinguish
12451 between bad padding and a MAC verification error. ([CVE-2003-0078])
12453 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12454 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12455 Martin Vuagnoux (EPFL, Ilion)*
12457 * Make the no-err option work as intended. The intention with no-err
12458 is not to have the whole error stack handling routines removed from
12459 libcrypto, it's only intended to remove all the function name and
12460 reason texts, thereby removing some of the footprint that may not
12461 be interesting if those errors aren't displayed anyway.
12463 NOTE: it's still possible for any application or module to have its
12464 own set of error texts inserted. The routines are there, just not
12465 used by default when no-err is given.
12469 * Add support for FreeBSD on IA64.
12471 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12473 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12474 Kerberos function mit_des_cbc_cksum(). Before this change,
12475 the value returned by DES_cbc_cksum() was like the one from
12476 mit_des_cbc_cksum(), except the bytes were swapped.
12478 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12480 * Allow an application to disable the automatic SSL chain building.
12481 Before this a rather primitive chain build was always performed in
12482 ssl3_output_cert_chain(): an application had no way to send the
12483 correct chain if the automatic operation produced an incorrect result.
12485 Now the chain builder is disabled if either:
12487 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12489 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12491 The reasoning behind this is that an application would not want the
12492 auto chain building to take place if extra chain certificates are
12493 present and it might also want a means of sending no additional
12494 certificates (for example the chain has two certificates and the
12499 * Add the possibility to build without the ENGINE framework.
12501 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12503 * Under Win32 gmtime() can return NULL: check return value in
12504 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12508 * DSA routines: under certain error conditions uninitialized BN objects
12509 could be freed. Solution: make sure initialization is performed early
12510 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12511 Nils Larsch <nla@trustcenter.de> via PR#459)
12515 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12516 checked on reconnect on the client side, therefore session resumption
12517 could still fail with a "ssl session id is different" error. This
12518 behaviour is masked when SSL_OP_ALL is used due to
12519 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12520 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12521 followup to PR #377.
12525 * IA-32 assembler support enhancements: unified ELF targets, support
12526 for SCO/Caldera platforms, fix for Cygwin shared build.
12530 * Add support for FreeBSD on sparc64. As a consequence, support for
12531 FreeBSD on non-x86 processors is separate from x86 processors on
12532 the config script, much like the NetBSD support.
12534 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12536 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12538 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12541 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12542 code (06) was taken as the first octet of the session ID and the last
12543 octet was ignored consequently. As a result SSLv2 client side session
12544 caching could not have worked due to the session ID mismatch between
12546 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12551 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12552 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12557 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12558 seems that in spite of existing for more than a year, many application
12559 author have done nothing to provide the necessary callbacks, which
12560 means that this particular engine will not work properly anywhere.
12561 This is a very unfortunate situation which forces us, in the name
12562 of usability, to give the hw_ncipher.c a static lock, which is part
12564 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12565 appear in 0.9.8 or later. We EXPECT application authors to have
12566 dealt properly with this when 0.9.8 is released (unless we actually
12567 make such changes in the libcrypto locking code that changes will
12568 have to be made anyway).
12572 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12573 octets have been read, EOF or an error occurs. Without this change
12574 some truncated ASN1 structures will not produce an error.
12578 * Disable Heimdal support, since it hasn't been fully implemented.
12579 Still give the possibility to force the use of Heimdal, but with
12580 warnings and a request that patches get sent to openssl-dev.
12584 * Add the VC-CE target, introduce the WINCE sysname, and add
12585 INSTALL.WCE and appropriate conditionals to make it build.
12587 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12589 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12590 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12591 edit numbers of the version.
12593 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12595 * Introduce safe string copy and catenation functions
12596 (BUF_strlcpy() and BUF_strlcat()).
12598 *Ben Laurie (CHATS) and Richard Levitte*
12600 * Avoid using fixed-size buffers for one-line DNs.
12602 *Ben Laurie (CHATS)*
12604 * Add BUF_MEM_grow_clean() to avoid information leakage when
12605 resizing buffers containing secrets, and use where appropriate.
12607 *Ben Laurie (CHATS)*
12609 * Avoid using fixed size buffers for configuration file location.
12611 *Ben Laurie (CHATS)*
12613 * Avoid filename truncation for various CA files.
12615 *Ben Laurie (CHATS)*
12617 * Use sizeof in preference to magic numbers.
12619 *Ben Laurie (CHATS)*
12621 * Avoid filename truncation in cert requests.
12623 *Ben Laurie (CHATS)*
12625 * Add assertions to check for (supposedly impossible) buffer
12628 *Ben Laurie (CHATS)*
12630 * Don't cache truncated DNS entries in the local cache (this could
12631 potentially lead to a spoofing attack).
12633 *Ben Laurie (CHATS)*
12635 * Fix various buffers to be large enough for hex/decimal
12636 representations in a platform independent manner.
12638 *Ben Laurie (CHATS)*
12640 * Add CRYPTO_realloc_clean() to avoid information leakage when
12641 resizing buffers containing secrets, and use where appropriate.
12643 *Ben Laurie (CHATS)*
12645 * Add BIO_indent() to avoid much slightly worrying code to do
12648 *Ben Laurie (CHATS)*
12650 * Convert sprintf()/BIO_puts() to BIO_printf().
12652 *Ben Laurie (CHATS)*
12654 * buffer_gets() could terminate with the buffer only half
12657 *Ben Laurie (CHATS)*
12659 * Add assertions to prevent user-supplied crypto functions from
12660 overflowing internal buffers by having large block sizes, etc.
12662 *Ben Laurie (CHATS)*
12664 * New OPENSSL_assert() macro (similar to assert(), but enabled
12667 *Ben Laurie (CHATS)*
12669 * Eliminate unused copy of key in RC4.
12671 *Ben Laurie (CHATS)*
12673 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12675 *Ben Laurie (CHATS)*
12677 * Fix off-by-one error in EGD path.
12679 *Ben Laurie (CHATS)*
12681 * If RANDFILE path is too long, ignore instead of truncating.
12683 *Ben Laurie (CHATS)*
12685 * Eliminate unused and incorrectly sized X.509 structure
12688 *Ben Laurie (CHATS)*
12690 * Eliminate unused and dangerous function knumber().
12692 *Ben Laurie (CHATS)*
12694 * Eliminate unused and dangerous structure, KSSL_ERR.
12696 *Ben Laurie (CHATS)*
12698 * Protect against overlong session ID context length in an encoded
12699 session object. Since these are local, this does not appear to be
12702 *Ben Laurie (CHATS)*
12704 * Change from security patch (see 0.9.6e below) that did not affect
12705 the 0.9.6 release series:
12707 Remote buffer overflow in SSL3 protocol - an attacker could
12708 supply an oversized master key in Kerberos-enabled versions.
12711 *Ben Laurie (CHATS)*
12713 * Change the SSL kerb5 codes to match RFC 2712.
12717 * Make -nameopt work fully for req and add -reqopt switch.
12719 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12721 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12723 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12725 * Make sure tests can be performed even if the corresponding algorithms
12726 have been removed entirely. This was also the last step to make
12727 OpenSSL compilable with DJGPP under all reasonable conditions.
12729 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12731 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12732 to allow version independent disabling of normally unselected ciphers,
12733 which may be activated as a side-effect of selecting a single cipher.
12735 (E.g., cipher list string "RSA" enables ciphersuites that are left
12736 out of "ALL" because they do not provide symmetric encryption.
12737 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12739 *Lutz Jaenicke, Bodo Moeller*
12741 * Add appropriate support for separate platform-dependent build
12742 directories. The recommended way to make a platform-dependent
12743 build directory is the following (tested on Linux), maybe with
12746 # Place yourself outside of the OpenSSL source tree. In
12747 # this example, the environment variable OPENSSL_SOURCE
12748 # is assumed to contain the absolute OpenSSL source directory.
12749 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12750 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12751 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12752 mkdir -p `dirname $F`
12753 ln -s $OPENSSL_SOURCE/$F $F
12756 To be absolutely sure not to disturb the source tree, a "make clean"
12757 is a good thing. If it isn't successful, don't worry about it,
12758 it probably means the source directory is very clean.
12762 * Make sure any ENGINE control commands make local copies of string
12763 pointers passed to them whenever necessary. Otherwise it is possible
12764 the caller may have overwritten (or deallocated) the original string
12765 data when a later ENGINE operation tries to use the stored values.
12767 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12769 * Improve diagnostics in file reading and command-line digests.
12771 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12773 * Add AES modes CFB and OFB to the object database. Correct an
12774 error in AES-CFB decryption.
12778 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12779 allows existing EVP_CIPHER_CTX structures to be reused after
12780 calling `EVP_*Final()`. This behaviour is used by encryption
12781 BIOs and some applications. This has the side effect that
12782 applications must explicitly clean up cipher contexts with
12783 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12787 * Check the values of dna and dnb in bn_mul_recursive before calling
12788 bn_mul_comba (a non zero value means the a or b arrays do not contain
12789 n2 elements) and fallback to bn_mul_normal if either is not zero.
12793 * Fix escaping of non-ASCII characters when using the -subj option
12794 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12798 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12799 form for "surname", serialNumber has no short form.
12800 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12801 therefore remove "mail" short name for "internet 7".
12802 The OID for unique identifiers in X509 certificates is
12803 x500UniqueIdentifier, not uniqueIdentifier.
12804 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12808 * Add an "init" command to the ENGINE config module and auto initialize
12809 ENGINEs. Without any "init" command the ENGINE will be initialized
12810 after all ctrl commands have been executed on it. If init=1 the
12811 ENGINE is initialized at that point (ctrls before that point are run
12812 on the uninitialized ENGINE and after on the initialized one). If
12813 init=0 then the ENGINE will not be initialized at all.
12817 * Fix the 'app_verify_callback' interface so that the user-defined
12818 argument is actually passed to the callback: In the
12819 SSL_CTX_set_cert_verify_callback() prototype, the callback
12820 declaration has been changed from
12823 int (*cb)(X509_STORE_CTX *,void *);
12824 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12825 i=s->ctx->app_verify_callback(&ctx)
12826 has been changed into
12827 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12829 To update applications using SSL_CTX_set_cert_verify_callback(),
12830 a dummy argument can be added to their callback functions.
12832 *D. K. Smetters <smetters@parc.xerox.com>*
12834 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12836 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12838 * Add and OPENSSL_LOAD_CONF define which will cause
12839 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12840 This allows older applications to transparently support certain
12841 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12842 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12843 load the config file and OPENSSL_add_all_algorithms_conf() which will
12844 always load it have also been added.
12848 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12849 Adjust NIDs and EVP layer.
12851 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12853 * Config modules support in openssl utility.
12855 Most commands now load modules from the config file,
12856 though in a few (such as version) this isn't done
12857 because it couldn't be used for anything.
12859 In the case of ca and req the config file used is
12860 the same as the utility itself: that is the -config
12861 command line option can be used to specify an
12866 * Move default behaviour from OPENSSL_config(). If appname is NULL
12867 use "openssl_conf" if filename is NULL use default openssl config file.
12871 * Add an argument to OPENSSL_config() to allow the use of an alternative
12872 config section name. Add a new flag to tolerate a missing config file
12873 and move code to CONF_modules_load_file().
12877 * Support for crypto accelerator cards from Accelerated Encryption
12878 Processing, www.aep.ie. (Use engine 'aep')
12879 The support was copied from 0.9.6c [engine] and adapted/corrected
12880 to work with the new engine framework.
12882 *AEP Inc. and Richard Levitte*
12884 * Support for SureWare crypto accelerator cards from Baltimore
12885 Technologies. (Use engine 'sureware')
12886 The support was copied from 0.9.6c [engine] and adapted
12887 to work with the new engine framework.
12891 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12892 make the newer ENGINE framework commands for the CHIL engine work.
12894 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12896 * Make it possible to produce shared libraries on ReliantUNIX.
12898 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12900 * Add the configuration target debug-linux-ppro.
12901 Make 'openssl rsa' use the general key loading routines
12902 implemented in `apps.c`, and make those routines able to
12903 handle the key format FORMAT_NETSCAPE and the variant
12906 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12908 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12910 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12912 * Add -keyform to rsautl, and document -engine.
12914 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12916 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12917 BIO_R_NO_SUCH_FILE error code rather than the generic
12918 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12922 * Add new functions
12923 ERR_peek_last_error
12924 ERR_peek_last_error_line
12925 ERR_peek_last_error_line_data.
12926 These are similar to
12928 ERR_peek_error_line
12929 ERR_peek_error_line_data,
12930 but report on the latest error recorded rather than the first one
12931 still in the error queue.
12933 *Ben Laurie, Bodo Moeller*
12935 * default_algorithms option in ENGINE config module. This allows things
12937 default_algorithms = ALL
12938 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12942 * Preliminary ENGINE config module.
12946 * New experimental application configuration code.
12950 * Change the AES code to follow the same name structure as all other
12951 symmetric ciphers, and behave the same way. Move everything to
12952 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12954 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12956 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12958 *Ben Laurie and Theo de Raadt*
12960 * Add option to output public keys in req command.
12962 *Massimiliano Pala madwolf@openca.org*
12964 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12965 (up to about 10% better than before for P-192 and P-224).
12969 * New functions/macros
12971 SSL_CTX_set_msg_callback(ctx, cb)
12972 SSL_CTX_set_msg_callback_arg(ctx, arg)
12973 SSL_set_msg_callback(ssl, cb)
12974 SSL_set_msg_callback_arg(ssl, arg)
12976 to request calling a callback function
12978 void cb(int write_p, int version, int content_type,
12979 const void *buf, size_t len, SSL *ssl, void *arg)
12981 whenever a protocol message has been completely received
12982 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12983 protocol version according to which the SSL library interprets
12984 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12985 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12986 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12987 specification (change_cipher_spec(20), alert(21), handshake(22)).
12988 'buf' and 'len' point to the actual message, 'ssl' to the
12989 SSL object, and 'arg' is the application-defined value set by
12990 SSL[_CTX]_set_msg_callback_arg().
12992 'openssl s_client' and 'openssl s_server' have new '-msg' options
12993 to enable a callback that displays all protocol messages.
12997 * Change the shared library support so shared libraries are built as
12998 soon as the corresponding static library is finished, and thereby get
12999 openssl and the test programs linked against the shared library.
13000 This still only happens when the keyword "shard" has been given to
13001 the configuration scripts.
13003 NOTE: shared library support is still an experimental thing, and
13004 backward binary compatibility is still not guaranteed.
13006 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13008 * Add support for Subject Information Access extension.
13010 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13012 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13013 additional bytes when new memory had to be allocated, not just
13014 when reusing an existing buffer.
13018 * New command line and configuration option 'utf8' for the req command.
13019 This allows field values to be specified as UTF8 strings.
13023 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13024 runs for the former and machine-readable output for the latter.
13028 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13029 of the e-mail address in the DN (i.e., it will go into a certificate
13030 extension only). The new configuration file option 'email_in_dn = no'
13031 has the same effect.
13033 *Massimiliano Pala madwolf@openca.org*
13035 * Change all functions with names starting with `des_` to be starting
13036 with `DES_` instead. Add wrappers that are compatible with libdes,
13037 but are named `_ossl_old_des_*`. Finally, add macros that map the
13038 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13039 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13040 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13043 Since we provide two compatibility mappings, the user needs to
13044 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13045 compatibility is desired. The default (i.e., when that macro
13046 isn't defined) is OpenSSL 0.9.6c compatibility.
13048 There are also macros that enable and disable the support of old
13049 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13050 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13051 are defined, the default will apply: to support the old des routines.
13053 In either case, one must include openssl/des.h to get the correct
13054 definitions. Do not try to just include openssl/des_old.h, that
13057 NOTE: This is a major break of an old API into a new one. Software
13058 authors are encouraged to switch to the `DES_` style functions. Some
13059 time in the future, des_old.h and the libdes compatibility functions
13060 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13061 default), and then completely removed.
13065 * Test for certificates which contain unsupported critical extensions.
13066 If such a certificate is found during a verify operation it is
13067 rejected by default: this behaviour can be overridden by either
13068 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13069 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13070 X509_supported_extension() has also been added which returns 1 if a
13071 particular extension is supported.
13075 * Modify the behaviour of EVP cipher functions in similar way to digests
13076 to retain compatibility with existing code.
13080 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13081 compatibility with existing code. In particular the 'ctx' parameter does
13082 not have to be to be initialized before the call to EVP_DigestInit() and
13083 it is tidied up after a call to EVP_DigestFinal(). New function
13084 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13085 EVP_MD_CTX_copy() changed to not require the destination to be
13086 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13087 requires the destination to be valid.
13089 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13090 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13094 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13095 so that complete 'Handshake' protocol structures are kept in memory
13096 instead of overwriting 'msg_type' and 'length' with 'body' data.
13100 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13102 *Massimo Santin via Richard Levitte*
13104 * Major restructuring to the underlying ENGINE code. This includes
13105 reduction of linker bloat, separation of pure "ENGINE" manipulation
13106 (initialisation, etc) from functionality dealing with implementations
13107 of specific crypto interfaces. This change also introduces integrated
13108 support for symmetric ciphers and digest implementations - so ENGINEs
13109 can now accelerate these by providing EVP_CIPHER and EVP_MD
13110 implementations of their own. This is detailed in
13111 [crypto/engine/README.md](crypto/engine/README.md)
13112 as it couldn't be adequately described here. However, there are a few
13113 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13114 were changed in the original introduction of ENGINE code have now
13115 reverted back - the hooking from this code to ENGINE is now a good
13116 deal more passive and at run-time, operations deal directly with
13117 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13118 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13119 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13120 they were not being used by the framework as there is no concept of a
13121 BIGNUM_METHOD and they could not be generalised to the new
13122 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13123 ENGINE_cpy() has been removed as it cannot be consistently defined in
13128 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13132 * Change mkdef.pl to sort symbols that get the same entry number,
13133 and make sure the automatically generated functions `ERR_load_*`
13134 become part of libeay.num as well.
13138 * New function SSL_renegotiate_pending(). This returns true once
13139 renegotiation has been requested (either SSL_renegotiate() call
13140 or HelloRequest/ClientHello received from the peer) and becomes
13141 false once a handshake has been completed.
13142 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13143 sends a HelloRequest, but does not ensure that a handshake takes
13144 place. SSL_renegotiate_pending() is useful for checking if the
13145 client has followed the request.)
13149 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13150 By default, clients may request session resumption even during
13151 renegotiation (if session ID contexts permit); with this option,
13152 session resumption is possible only in the first handshake.
13154 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13155 more bits available for options that should not be part of
13156 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13160 * Add some demos for certificate and certificate request creation.
13164 * Make maximum certificate chain size accepted from the peer application
13165 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13166 "Douglas E. Engert" <deengert@anl.gov>.
13170 * Add support for shared libraries for Unixware-7
13171 (Boyd Lynn Gerber <gerberb@zenez.com>).
13175 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13176 be done prior to destruction. Use this to unload error strings from
13177 ENGINEs that load their own error strings. NB: This adds two new API
13178 functions to "get" and "set" this destroy handler in an ENGINE.
13182 * Alter all existing ENGINE implementations (except "openssl" and
13183 "openbsd") to dynamically instantiate their own error strings. This
13184 makes them more flexible to be built both as statically-linked ENGINEs
13185 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13186 Also, add stub code to each that makes building them as self-contained
13187 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13191 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13192 implementations into applications that are completely implemented in
13193 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13194 commands that can be used to configure what shared-library to load and
13195 to control aspects of the way it is handled. Also, made an update to
13196 the [README-Engine.md](README-Engine.md) file
13197 that brings its information up-to-date and
13198 provides some information and instructions on the "dynamic" ENGINE
13199 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13203 * Make it possible to unload ranges of ERR strings with a new
13204 "ERR_unload_strings" function.
13208 * Add a copy() function to EVP_MD.
13212 * Make EVP_MD routines take a context pointer instead of just the
13213 md_data void pointer.
13217 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13218 that the digest can only process a single chunk of data
13219 (typically because it is provided by a piece of
13220 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13221 is only going to provide a single chunk of data, and hence the
13222 framework needn't accumulate the data for oneshot drivers.
13226 * As with "ERR", make it possible to replace the underlying "ex_data"
13227 functions. This change also alters the storage and management of global
13228 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13229 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13230 index counters. The API functions that use this state have been changed
13231 to take a "class_index" rather than pointers to the class's local STACK
13232 and counter, and there is now an API function to dynamically create new
13233 classes. This centralisation allows us to (a) plug a lot of the
13234 thread-safety problems that existed, and (b) makes it possible to clean
13235 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13236 such data would previously have always leaked in application code and
13237 workarounds were in place to make the memory debugging turn a blind eye
13238 to it. Application code that doesn't use this new function will still
13239 leak as before, but their memory debugging output will announce it now
13240 rather than letting it slide.
13242 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13243 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13244 has a return value to indicate success or failure.
13248 * Make it possible to replace the underlying "ERR" functions such that the
13249 global state (2 LHASH tables and 2 locks) is only used by the "default"
13250 implementation. This change also adds two functions to "get" and "set"
13251 the implementation prior to it being automatically set the first time
13252 any other ERR function takes place. Ie. an application can call "get",
13253 pass the return value to a module it has just loaded, and that module
13254 can call its own "set" function using that value. This means the
13255 module's "ERR" operations will use (and modify) the error state in the
13256 application and not in its own statically linked copy of OpenSSL code.
13260 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13261 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13262 the operation, and provides a more encapsulated way for external code
13263 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13264 to use these functions rather than manually incrementing the counts.
13266 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13270 * Add EVP test program.
13274 * Add symmetric cipher support to ENGINE. Expect the API to change!
13278 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13279 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13280 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13281 These allow a CRL to be built without having to access X509_CRL fields
13282 directly. Modify 'ca' application to use new functions.
13286 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13287 bug workarounds. Rollback attack detection is a security feature.
13288 The problem will only arise on OpenSSL servers when TLSv1 is not
13289 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13290 Software authors not wanting to support TLSv1 will have special reasons
13291 for their choice and can explicitly enable this option.
13293 *Bodo Moeller, Lutz Jaenicke*
13295 * Rationalise EVP so it can be extended: don't include a union of
13296 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13297 (similar to those existing for EVP_CIPHER_CTX).
13302 EVP_MD_CTX_init(&md); /* new function call */
13303 EVP_DigestInit(&md, EVP_sha1());
13304 EVP_DigestUpdate(&md, in, len);
13305 EVP_DigestFinal(&md, out, NULL);
13306 EVP_MD_CTX_cleanup(&md); /* new function call */
13310 * Make DES key schedule conform to the usual scheme, as well as
13311 correcting its structure. This means that calls to DES functions
13312 now have to pass a pointer to a des_key_schedule instead of a
13313 plain des_key_schedule (which was actually always a pointer
13316 des_key_schedule ks;
13318 des_set_key_checked(..., &ks);
13319 des_ncbc_encrypt(..., &ks, ...);
13321 (Note that a later change renames 'des_...' into 'DES_...'.)
13325 * Initial reduction of linker bloat: the use of some functions, such as
13326 PEM causes large amounts of unused functions to be linked in due to
13327 poor organisation. For example pem_all.c contains every PEM function
13328 which has a knock on effect of linking in large amounts of (unused)
13329 ASN1 code. Grouping together similar functions and splitting unrelated
13330 functions prevents this.
13334 * Cleanup of EVP macros.
13338 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13339 correct `_ecb suffix`.
13343 * Add initial OCSP responder support to ocsp application. The
13344 revocation information is handled using the text based index
13345 use by the ca application. The responder can either handle
13346 requests generated internally, supplied in files (for example
13347 via a CGI script) or using an internal minimal server.
13351 * Add configuration choices to get zlib compression for TLS.
13355 * Changes to Kerberos SSL for RFC 2712 compliance:
13356 1. Implemented real KerberosWrapper, instead of just using
13357 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13358 2. Implemented optional authenticator field of KerberosWrapper.
13360 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13361 and authenticator structs; see crypto/krb5/.
13363 Generalized Kerberos calls to support multiple Kerberos libraries.
13364 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13365 via Richard Levitte*
13367 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13368 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13369 values for each of the key sizes rather than having just
13370 parameters (and 'speed' generating keys each time).
13374 * Speed up EVP routines.
13377 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13378 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13379 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13380 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13382 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13383 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13384 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13387 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13389 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13393 * Added the OS2-EMX target.
13395 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13397 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13398 New functions to support `NCONF` routines in extension code.
13399 New function `CONF_set_nconf()`
13400 to allow functions which take an `NCONF` to also handle the old `LHASH`
13401 structure: this means that the old `CONF` compatible routines can be
13402 retained (in particular w.rt. extensions) without having to duplicate the
13403 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13407 * Enhance the general user interface with mechanisms for inner control
13408 and with possibilities to have yes/no kind of prompts.
13412 * Change all calls to low-level digest routines in the library and
13413 applications to use EVP. Add missing calls to HMAC_cleanup() and
13414 don't assume HMAC_CTX can be copied using memcpy().
13416 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13418 * Add the possibility to control engines through control names but with
13419 arbitrary arguments instead of just a string.
13420 Change the key loaders to take a UI_METHOD instead of a callback
13421 function pointer. NOTE: this breaks binary compatibility with earlier
13422 versions of OpenSSL [engine].
13423 Adapt the nCipher code for these new conditions and add a card insertion
13428 * Enhance the general user interface with mechanisms to better support
13429 dialog box interfaces, application-defined prompts, the possibility
13430 to use defaults (for example default passwords from somewhere else)
13431 and interrupts/cancellations.
13435 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13436 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13440 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13441 tidy up some unnecessarily weird code in 'sk_new()').
13443 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13445 * Change the key loading routines for ENGINEs to use the same kind
13446 callback (pem_password_cb) as all other routines that need this
13451 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13452 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13453 than this minimum value is recommended.
13457 * New random seeder for OpenVMS, using the system process statistics
13458 that are easily reachable.
13462 * Windows apparently can't transparently handle global
13463 variables defined in DLLs. Initialisations such as:
13465 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13467 won't compile. This is used by the any applications that need to
13468 declare their own ASN1 modules. This was fixed by adding the option
13469 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13470 needed for static libraries under Win32.
13474 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13475 setting of purpose and trust fields. New X509_STORE trust and
13476 purpose functions and tidy up setting in other SSL functions.
13480 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13481 structure. These are inherited by X509_STORE_CTX when it is
13482 initialised. This allows various defaults to be set in the
13483 X509_STORE structure (such as flags for CRL checking and custom
13484 purpose or trust settings) for functions which only use X509_STORE_CTX
13485 internally such as S/MIME.
13487 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13488 trust settings if they are not set in X509_STORE. This allows X509_STORE
13489 purposes and trust (in S/MIME for example) to override any set by default.
13491 Add command line options for CRL checking to smime, s_client and s_server
13496 * Initial CRL based revocation checking. If the CRL checking flag(s)
13497 are set then the CRL is looked up in the X509_STORE structure and
13498 its validity and signature checked, then if the certificate is found
13499 in the CRL the verify fails with a revoked error.
13501 Various new CRL related callbacks added to X509_STORE_CTX structure.
13503 Command line options added to 'verify' application to support this.
13505 This needs some additional work, such as being able to handle multiple
13506 CRLs with different times, extension based lookup (rather than just
13507 by subject name) and ultimately more complete V2 CRL extension
13512 * Add a general user interface API (crypto/ui/). This is designed
13513 to replace things like des_read_password and friends (backward
13514 compatibility functions using this new API are provided).
13515 The purpose is to remove prompting functions from the DES code
13516 section as well as provide for prompting through dialog boxes in
13517 a window system and the like.
13521 * Add "ex_data" support to ENGINE so implementations can add state at a
13522 per-structure level rather than having to store it globally.
13526 * Make it possible for ENGINE structures to be copied when retrieved by
13527 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13528 This causes the "original" ENGINE structure to act like a template,
13529 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13530 operational state can be localised to each ENGINE structure, despite the
13531 fact they all share the same "methods". New ENGINE structures returned in
13532 this case have no functional references and the return value is the single
13533 structural reference. This matches the single structural reference returned
13534 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13539 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13540 needs to match any other type at all we need to manually clear the
13545 * Changes to the "openssl engine" utility to include;
13546 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13547 about an ENGINE's available control commands.
13548 - executing control commands from command line arguments using the
13549 '-pre' and '-post' switches. '-post' is only used if '-t' is
13550 specified and the ENGINE is successfully initialised. The syntax for
13551 the individual commands are colon-separated, for example;
13552 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13556 * New dynamic control command support for ENGINEs. ENGINEs can now
13557 declare their own commands (numbers), names (strings), descriptions,
13558 and input types for run-time discovery by calling applications. A
13559 subset of these commands are implicitly classed as "executable"
13560 depending on their input type, and only these can be invoked through
13561 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13562 can be based on user input, config files, etc). The distinction is
13563 that "executable" commands cannot return anything other than a boolean
13564 result and can only support numeric or string input, whereas some
13565 discoverable commands may only be for direct use through
13566 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13567 pointers, or other custom uses. The "executable" commands are to
13568 support parameterisations of ENGINE behaviour that can be
13569 unambiguously defined by ENGINEs and used consistently across any
13570 OpenSSL-based application. Commands have been added to all the
13571 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13572 control over shared-library paths without source code alterations.
13576 * Changed all ENGINE implementations to dynamically allocate their
13577 ENGINEs rather than declaring them statically. Apart from this being
13578 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13579 this also allows the implementations to compile without using the
13580 internal engine_int.h header.
13584 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13585 'const' value. Any code that should be able to modify a RAND_METHOD
13586 should already have non-const pointers to it (ie. they should only
13587 modify their own ones).
13591 * Made a variety of little tweaks to the ENGINE code.
13592 - "atalla" and "ubsec" string definitions were moved from header files
13593 to C code. "nuron" string definitions were placed in variables
13594 rather than hard-coded - allowing parameterisation of these values
13595 later on via ctrl() commands.
13596 - Removed unused "#if 0"'d code.
13597 - Fixed engine list iteration code so it uses ENGINE_free() to release
13598 structural references.
13599 - Constified the RAND_METHOD element of ENGINE structures.
13600 - Constified various get/set functions as appropriate and added
13601 missing functions (including a catch-all ENGINE_cpy that duplicates
13602 all ENGINE values onto a new ENGINE except reference counts/state).
13603 - Removed NULL parameter checks in get/set functions. Setting a method
13604 or function to NULL is a way of cancelling out a previously set
13605 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13606 and doesn't justify the extra error symbols and code.
13607 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13608 flags from engine_int.h to engine.h.
13609 - Changed prototypes for ENGINE handler functions (init(), finish(),
13610 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13614 * Implement binary inversion algorithm for BN_mod_inverse in addition
13615 to the algorithm using long division. The binary algorithm can be
13616 used only if the modulus is odd. On 32-bit systems, it is faster
13617 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13618 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13619 up to 450 bits. In 64-bit environments, the binary algorithm
13620 appears to be advantageous for much longer moduli; here we use it
13621 for moduli up to 2048 bits.
13625 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13626 could not support the combine flag in choice fields.
13630 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13631 extensions from a certificate request to the certificate.
13635 * Allow multiple 'certopt' and 'nameopt' options to be separated
13636 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13637 file: this allows the display of the certificate about to be
13638 signed to be customised, to allow certain fields to be included
13639 or excluded and extension details. The old system didn't display
13640 multicharacter strings properly, omitted fields not in the policy
13641 and couldn't display additional details such as extensions.
13645 * Function EC_POINTs_mul for multiple scalar multiplication
13646 of an arbitrary number of elliptic curve points
13647 \sum scalars[i]*points[i],
13648 optionally including the generator defined for the EC_GROUP:
13649 scalar*generator + \sum scalars[i]*points[i].
13651 EC_POINT_mul is a simple wrapper function for the typical case
13652 that the point list has just one item (besides the optional
13657 * First EC_METHODs for curves over GF(p):
13659 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13660 operations and provides various method functions that can also
13661 operate with faster implementations of modular arithmetic.
13663 EC_GFp_mont_method() reuses most functions that are part of
13664 EC_GFp_simple_method, but uses Montgomery arithmetic.
13666 *Bodo Moeller; point addition and point doubling
13667 implementation directly derived from source code provided by
13668 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13670 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13671 crypto/ec/ec_lib.c):
13673 Curves are EC_GROUP objects (with an optional group generator)
13674 based on EC_METHODs that are built into the library.
13676 Points are EC_POINT objects based on EC_GROUP objects.
13678 Most of the framework would be able to handle curves over arbitrary
13679 finite fields, but as there are no obvious types for fields other
13680 than GF(p), some functions are limited to that for now.
13684 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13685 that the file contains a complete HTTP response.
13689 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13690 change the def and num file printf format specifier from "%-40sXXX"
13691 to "%-39s XXX". The latter will always guarantee a space after the
13692 field while the former will cause them to run together if the field
13693 is 40 of more characters long.
13697 * Constify the cipher and digest 'method' functions and structures
13698 and modify related functions to take constant EVP_MD and EVP_CIPHER
13703 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13704 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13708 * Modify `EVP_Digest*()` routines so they now return values. Although the
13709 internal software routines can never fail additional hardware versions
13714 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13716 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13717 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13720 ERR_R_NESTED_ASN1_ERROR
13722 ERR_R_MISSING_ASN1_EOS
13723 were 4 .. 9, conflicting with
13724 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13726 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13727 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13729 Add new error code 'ERR_R_INTERNAL_ERROR'.
13733 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13738 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13739 sets the subject name for a new request or supersedes the
13740 subject name in a given request. Formats that can be parsed are
13741 'CN=Some Name, OU=myOU, C=IT'
13743 'CN=Some Name/OU=myOU/C=IT'.
13745 Add options '-batch' and '-verbose' to 'openssl req'.
13747 *Massimiliano Pala <madwolf@hackmasters.net>*
13749 * Introduce the possibility to access global variables through
13750 functions on platform were that's the best way to handle exporting
13751 global variables in shared libraries. To enable this functionality,
13752 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13753 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13754 is normally done by Configure or something similar).
13756 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13757 in the source file (foo.c) like this:
13759 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13760 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13762 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13763 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13765 OPENSSL_DECLARE_GLOBAL(int,foo);
13766 #define foo OPENSSL_GLOBAL_REF(foo)
13767 OPENSSL_DECLARE_GLOBAL(double,bar);
13768 #define bar OPENSSL_GLOBAL_REF(bar)
13770 The #defines are very important, and therefore so is including the
13771 header file everywhere where the defined globals are used.
13773 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13774 of ASN.1 items, but that structure is a bit different.
13776 The largest change is in util/mkdef.pl which has been enhanced with
13777 better and easier to understand logic to choose which symbols should
13778 go into the Windows .def files as well as a number of fixes and code
13779 cleanup (among others, algorithm keywords are now sorted
13780 lexicographically to avoid constant rewrites).
13784 * In BN_div() keep a copy of the sign of 'num' before writing the
13785 result to 'rm' because if rm==num the value will be overwritten
13786 and produce the wrong result if 'num' is negative: this caused
13787 problems with BN_mod() and BN_nnmod().
13791 * Function OCSP_request_verify(). This checks the signature on an
13792 OCSP request and verifies the signer certificate. The signer
13793 certificate is just checked for a generic purpose and OCSP request
13798 * Add OCSP_check_validity() function to check the validity of OCSP
13799 responses. OCSP responses are prepared in real time and may only
13800 be a few seconds old. Simply checking that the current time lies
13801 between thisUpdate and nextUpdate max reject otherwise valid responses
13802 caused by either OCSP responder or client clock inaccuracy. Instead
13803 we allow thisUpdate and nextUpdate to fall within a certain period of
13804 the current time. The age of the response can also optionally be
13805 checked. Two new options -validity_period and -status_age added to
13810 * If signature or public key algorithm is unrecognized print out its
13811 OID rather that just UNKNOWN.
13815 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13816 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13817 ID to be generated from the issuer certificate alone which can then be
13818 passed to OCSP_id_issuer_cmp().
13822 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13823 ASN1 modules to export functions returning ASN1_ITEM pointers
13824 instead of the ASN1_ITEM structures themselves. This adds several
13825 new macros which allow the underlying ASN1 function/structure to
13826 be accessed transparently. As a result code should not use ASN1_ITEM
13827 references directly (such as &X509_it) but instead use the relevant
13828 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13829 use of the new ASN1 code on platforms where exporting structures
13830 is problematical (for example in shared libraries) but exporting
13831 functions returning pointers to structures is not.
13835 * Add support for overriding the generation of SSL/TLS session IDs.
13836 These callbacks can be registered either in an SSL_CTX or per SSL.
13837 The purpose of this is to allow applications to control, if they wish,
13838 the arbitrary values chosen for use as session IDs, particularly as it
13839 can be useful for session caching in multiple-server environments. A
13840 command-line switch for testing this (and any client code that wishes
13841 to use such a feature) has been added to "s_server".
13843 *Geoff Thorpe, Lutz Jaenicke*
13845 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13846 of the form `#if defined(...) || defined(...) || ...` and
13847 `#if !defined(...) && !defined(...) && ...`. This also avoids
13848 the growing number of special cases it was previously handling.
13852 * Make all configuration macros available for application by making
13853 sure they are available in opensslconf.h, by giving them names starting
13854 with `OPENSSL_` to avoid conflicts with other packages and by making
13855 sure e_os2.h will cover all platform-specific cases together with
13857 Additionally, it is now possible to define configuration/platform-
13858 specific names (called "system identities"). In the C code, these
13859 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13860 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13861 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13866 * New option -set_serial to 'req' and 'x509' this allows the serial
13867 number to use to be specified on the command line. Previously self
13868 signed certificates were hard coded with serial number 0 and the
13869 CA options of 'x509' had to use a serial number in a file which was
13874 * New options to 'ca' utility to support V2 CRL entry extensions.
13875 Currently CRL reason, invalidity date and hold instruction are
13876 supported. Add new CRL extensions to V3 code and some new objects.
13880 * New function EVP_CIPHER_CTX_set_padding() this is used to
13881 disable standard block padding (aka PKCS#5 padding) in the EVP
13882 API, which was previously mandatory. This means that the data is
13883 not padded in any way and so the total length much be a multiple
13884 of the block size, otherwise an error occurs.
13888 * Initial (incomplete) OCSP SSL support.
13892 * New function OCSP_parse_url(). This splits up a URL into its host,
13893 port and path components: primarily to parse OCSP URLs. New -url
13894 option to ocsp utility.
13898 * New nonce behavior. The return value of OCSP_check_nonce() now
13899 reflects the various checks performed. Applications can decide
13900 whether to tolerate certain situations such as an absent nonce
13901 in a response when one was present in a request: the ocsp application
13902 just prints out a warning. New function OCSP_add1_basic_nonce()
13903 this is to allow responders to include a nonce in a response even if
13904 the request is nonce-less.
13908 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13909 skipped when using openssl x509 multiple times on a single input file,
13910 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13914 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13915 set string type: to handle setting ASN1_TIME structures. Fix ca
13916 utility to correctly initialize revocation date of CRLs.
13920 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13921 the clients preferred ciphersuites and rather use its own preferences.
13922 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13923 Internet Explorer by ensuring unchanged hash method during stepup.
13924 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13928 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13929 to aes and add a new 'exist' option to print out symbols that don't
13934 * Additional options to ocsp utility to allow flags to be set and
13935 additional certificates supplied.
13939 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13940 OCSP client a number of certificate to only verify the response
13945 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13946 handle the new API. Currently only ECB, CBC modes supported. Add new
13949 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13950 Encryption Standard (AES) Ciphersuites for Transport Layer
13951 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13952 not enabled by default and were not part of the "ALL" ciphersuite
13953 alias because they were not yet official; they could be
13954 explicitly requested by specifying the "AESdraft" ciphersuite
13955 group alias. In the final release of OpenSSL 0.9.7, the group
13956 alias is called "AES" and is part of "ALL".)
13958 *Ben Laurie, Steve Henson, Bodo Moeller*
13960 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13961 request to response.
13965 * Functions for OCSP responders. OCSP_request_onereq_count(),
13966 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13967 extract information from a certificate request. OCSP_response_create()
13968 creates a response and optionally adds a basic response structure.
13969 OCSP_basic_add1_status() adds a complete single response to a basic
13970 response and returns the OCSP_SINGLERESP structure just added (to allow
13971 extensions to be included for example). OCSP_basic_add1_cert() adds a
13972 certificate to a basic response and OCSP_basic_sign() signs a basic
13973 response with various flags. New helper functions ASN1_TIME_check()
13974 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13975 (converts ASN1_TIME to GeneralizedTime).
13979 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13980 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13981 structure from a certificate. X509_pubkey_digest() digests the public_key
13982 contents: this is used in various key identifiers.
13986 * Make sk_sort() tolerate a NULL argument.
13988 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13990 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13991 passed by the function are trusted implicitly. If any of them signed the
13992 response then it is assumed to be valid and is not verified.
13996 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13997 to data. This was previously part of the PKCS7 ASN1 code. This
13998 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13999 *Steve Henson, reported by Kenneth R. Robinette
14000 <support@securenetterm.com>*
14002 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14003 routines: without these tracing memory leaks is very painful.
14004 Fix leaks in PKCS12 and PKCS7 routines.
14008 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14009 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14010 effectively meant GeneralizedTime would never be used. Now it
14011 is initialised to -1 but X509_time_adj() now has to check the value
14012 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14013 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14014 *Steve Henson, reported by Kenneth R. Robinette
14015 <support@securenetterm.com>*
14017 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14018 result in a zero length in the ASN1_INTEGER structure which was
14019 not consistent with the structure when d2i_ASN1_INTEGER() was used
14020 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14021 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14022 where it did not print out a minus for negative ASN1_INTEGER.
14026 * Add summary printout to ocsp utility. The various functions which
14027 convert status values to strings have been renamed to:
14028 OCSP_response_status_str(), OCSP_cert_status_str() and
14029 OCSP_crl_reason_str() and are no longer static. New options
14030 to verify nonce values and to disable verification. OCSP response
14031 printout format cleaned up.
14035 * Add additional OCSP certificate checks. These are those specified
14036 in RFC2560. This consists of two separate checks: the CA of the
14037 certificate being checked must either be the OCSP signer certificate
14038 or the issuer of the OCSP signer certificate. In the latter case the
14039 OCSP signer certificate must contain the OCSP signing extended key
14040 usage. This check is performed by attempting to match the OCSP
14041 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14042 in the OCSP_CERTID structures of the response.
14046 * Initial OCSP certificate verification added to OCSP_basic_verify()
14047 and related routines. This uses the standard OpenSSL certificate
14048 verify routines to perform initial checks (just CA validity) and
14049 to obtain the certificate chain. Then additional checks will be
14050 performed on the chain. Currently the root CA is checked to see
14051 if it is explicitly trusted for OCSP signing. This is used to set
14052 a root CA as a global signing root: that is any certificate that
14053 chains to that CA is an acceptable OCSP signing certificate.
14057 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14058 extensions from a separate configuration file.
14059 As when reading extensions from the main configuration file,
14060 the '-extensions ...' option may be used for specifying the
14063 *Massimiliano Pala <madwolf@comune.modena.it>*
14065 * New OCSP utility. Allows OCSP requests to be generated or
14066 read. The request can be sent to a responder and the output
14067 parsed, outputted or printed in text form. Not complete yet:
14068 still needs to check the OCSP response validity.
14072 * New subcommands for 'openssl ca':
14073 `openssl ca -status <serial>` prints the status of the cert with
14074 the given serial number (according to the index file).
14075 `openssl ca -updatedb` updates the expiry status of certificates
14078 *Massimiliano Pala <madwolf@comune.modena.it>*
14080 * New '-newreq-nodes' command option to CA.pl. This is like
14081 '-newreq', but calls 'openssl req' with the '-nodes' option
14082 so that the resulting key is not encrypted.
14084 *Damien Miller <djm@mindrot.org>*
14086 * New configuration for the GNU Hurd.
14088 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14090 * Initial code to implement OCSP basic response verify. This
14091 is currently incomplete. Currently just finds the signer's
14092 certificate and verifies the signature on the response.
14096 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14097 value of OPENSSLDIR. This is available via the new '-d' option
14098 to 'openssl version', and is also included in 'openssl version -a'.
14102 * Allowing defining memory allocation callbacks that will be given
14103 file name and line number information in additional arguments
14104 (a `const char*` and an int). The basic functionality remains, as
14105 well as the original possibility to just replace malloc(),
14106 realloc() and free() by functions that do not know about these
14107 additional arguments. To register and find out the current
14108 settings for extended allocation functions, the following
14109 functions are provided:
14111 CRYPTO_set_mem_ex_functions
14112 CRYPTO_set_locked_mem_ex_functions
14113 CRYPTO_get_mem_ex_functions
14114 CRYPTO_get_locked_mem_ex_functions
14116 These work the same way as CRYPTO_set_mem_functions and friends.
14117 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14118 extended allocation function is enabled.
14119 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14120 a conventional allocation function is enabled.
14122 *Richard Levitte, Bodo Moeller*
14124 * Finish off removing the remaining LHASH function pointer casts.
14125 There should no longer be any prototype-casting required when using
14126 the LHASH abstraction, and any casts that remain are "bugs". See
14127 the callback types and macros at the head of lhash.h for details
14128 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14132 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14133 If /dev/[u]random devices are not available or do not return enough
14134 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14136 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14137 /etc/entropy will be queried once each in this sequence, querying stops
14138 when enough entropy was collected without querying more sockets.
14142 * Change the Unix RAND_poll() variant to be able to poll several
14143 random devices, as specified by DEVRANDOM, until a sufficient amount
14144 of data has been collected. We spend at most 10 ms on each file
14145 (select timeout) and read in non-blocking mode. DEVRANDOM now
14146 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14147 (previously it was just the string "/dev/urandom"), so on typical
14148 platforms the 10 ms delay will never occur.
14149 Also separate out the Unix variant to its own file, rand_unix.c.
14150 For VMS, there's a currently-empty rand_vms.c.
14154 * Move OCSP client related routines to ocsp_cl.c. These
14155 provide utility functions which an application needing
14156 to issue a request to an OCSP responder and analyse the
14157 response will typically need: as opposed to those which an
14158 OCSP responder itself would need which will be added later.
14160 OCSP_request_sign() signs an OCSP request with an API similar
14161 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14162 response. OCSP_response_get1_basic() extracts basic response
14163 from response. OCSP_resp_find_status(): finds and extracts status
14164 information from an OCSP_CERTID structure (which will be created
14165 when the request structure is built). These are built from lower
14166 level functions which work on OCSP_SINGLERESP structures but
14167 won't normally be used unless the application wishes to examine
14168 extensions in the OCSP response for example.
14170 Replace nonce routines with a pair of functions.
14171 OCSP_request_add1_nonce() adds a nonce value and optionally
14172 generates a random value. OCSP_check_nonce() checks the
14173 validity of the nonce in an OCSP response.
14177 * Change function OCSP_request_add() to OCSP_request_add0_id().
14178 This doesn't copy the supplied OCSP_CERTID and avoids the
14179 need to free up the newly created id. Change return type
14180 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14181 This can then be used to add extensions to the request.
14182 Deleted OCSP_request_new(), since most of its functionality
14183 is now in OCSP_REQUEST_new() (and the case insensitive name
14184 clash) apart from the ability to set the request name which
14185 will be added elsewhere.
14189 * Update OCSP API. Remove obsolete extensions argument from
14190 various functions. Extensions are now handled using the new
14191 OCSP extension code. New simple OCSP HTTP function which
14192 can be used to send requests and parse the response.
14196 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14197 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14198 uses the special reorder version of SET OF to sort the attributes
14199 and reorder them to match the encoded order. This resolves a long
14200 standing problem: a verify on a PKCS7 structure just after signing
14201 it used to fail because the attribute order did not match the
14202 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14203 it uses the received order. This is necessary to tolerate some broken
14204 software that does not order SET OF. This is handled by encoding
14205 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14206 to produce the required SET OF.
14210 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14211 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14212 files to get correct declarations of the ASN.1 item variables.
14216 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14217 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14218 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14219 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14220 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14221 ASN1_ITEM and no wrapper functions.
14225 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14226 replace the old function pointer based I/O routines. Change most of
14227 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14231 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14232 lines, recognize more "algorithms" that can be deselected, and make
14233 it complain about algorithm deselection that isn't recognised.
14237 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14238 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14239 to use new functions. Add NO_ASN1_OLD which can be set to remove
14240 some old style ASN1 functions: this can be used to determine if old
14241 code will still work when these eventually go away.
14245 * New extension functions for OCSP structures, these follow the
14246 same conventions as certificates and CRLs.
14250 * New function X509V3_add1_i2d(). This automatically encodes and
14251 adds an extension. Its behaviour can be customised with various
14252 flags to append, replace or delete. Various wrappers added for
14253 certificates and CRLs.
14257 * Fix to avoid calling the underlying ASN1 print routine when
14258 an extension cannot be parsed. Correct a typo in the
14259 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14263 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14264 entries for variables.
14268 * Add functionality to `apps/openssl.c` for detecting locking
14269 problems: As the program is single-threaded, all we have
14270 to do is register a locking callback using an array for
14271 storing which locks are currently held by the program.
14275 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14276 SSL_get_ex_data_X509_STORE_idx(), which is used in
14277 ssl_verify_cert_chain() and thus can be called at any time
14278 during TLS/SSL handshakes so that thread-safety is essential.
14279 Unfortunately, the ex_data design is not at all suited
14280 for multi-threaded use, so it probably should be abolished.
14284 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14286 *Broadcom, tweaked and integrated by Geoff Thorpe*
14288 * Move common extension printing code to new function
14289 X509V3_print_extensions(). Reorganise OCSP print routines and
14290 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14294 * New function X509_signature_print() to remove duplication in some
14299 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14300 set (this was treated exactly the same as SET OF previously). This
14301 is used to reorder the STACK representing the structure to match the
14302 encoding. This will be used to get round a problem where a PKCS7
14303 structure which was signed could not be verified because the STACK
14304 order did not reflect the encoded order.
14308 * Reimplement the OCSP ASN1 module using the new code.
14312 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14313 for its ASN1 operations. The old style function pointers still exist
14314 for now but they will eventually go away.
14318 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14319 completely replaces the old ASN1 functionality with a table driven
14320 encoder and decoder which interprets an ASN1_ITEM structure describing
14321 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14322 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14323 has also been converted to the new form.
14327 * Change BN_mod_exp_recp so that negative moduli are tolerated
14328 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14329 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14330 for negative moduli.
14334 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14335 of not touching the result's sign bit.
14339 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14344 * Changed the LHASH code to use prototypes for callbacks, and created
14345 macros to declare and implement thin (optionally static) functions
14346 that provide type-safety and avoid function pointer casting for the
14347 type-specific callbacks.
14351 * Added Kerberos Cipher Suites to be used with TLS, as written in
14353 *Veers Staats <staatsvr@asc.hpc.mil>,
14354 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14356 * Reformat the FAQ so the different questions and answers can be divided
14357 in sections depending on the subject.
14361 * Have the zlib compression code load ZLIB.DLL dynamically under
14366 * New function BN_mod_sqrt for computing square roots modulo a prime
14367 (using the probabilistic Tonelli-Shanks algorithm unless
14368 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14369 be handled deterministically).
14371 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14373 * Make BN_mod_inverse faster by explicitly handling small quotients
14374 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14375 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14379 * New function BN_kronecker.
14383 * Fix BN_gcd so that it works on negative inputs; the result is
14384 positive unless both parameters are zero.
14385 Previously something reasonably close to an infinite loop was
14386 possible because numbers could be growing instead of shrinking
14387 in the implementation of Euclid's algorithm.
14391 * Fix BN_is_word() and BN_is_one() macros to take into account the
14392 sign of the number in question.
14394 Fix BN_is_word(a,w) to work correctly for w == 0.
14396 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14397 because its test if the absolute value of 'a' equals 'w'.
14398 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14399 it exists mostly for use in the implementations of BN_is_zero(),
14400 BN_is_one(), and BN_is_word().
14404 * New function BN_swap.
14408 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14409 the exponentiation functions are more likely to produce reasonable
14410 results on negative inputs.
14414 * Change BN_mod_mul so that the result is always non-negative.
14415 Previously, it could be negative if one of the factors was negative;
14416 I don't think anyone really wanted that behaviour.
14420 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14421 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14422 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14423 and add new functions:
14432 BN_mod_lshift1_quick
14434 BN_mod_lshift_quick
14436 These functions always generate non-negative results.
14438 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14439 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14441 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14442 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14443 be reduced modulo `m`.
14445 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14448 The following entry accidentally appeared in the CHANGES file
14449 distributed with OpenSSL 0.9.7. The modifications described in
14450 it do *not* apply to OpenSSL 0.9.7.
14452 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14453 was actually never needed) and in BN_mul(). The removal in BN_mul()
14454 required a small change in bn_mul_part_recursive() and the addition
14455 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14456 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14457 bn_sub_words() and bn_add_words() except they take arrays with
14463 * In 'openssl passwd', verify passwords read from the terminal
14464 unless the '-salt' option is used (which usually means that
14465 verification would just waste user's time since the resulting
14466 hash is going to be compared with some given password hash)
14467 or the new '-noverify' option is used.
14469 This is an incompatible change, but it does not affect
14470 non-interactive use of 'openssl passwd' (passwords on the command
14471 line, '-stdin' option, '-in ...' option) and thus should not
14472 cause any problems.
14476 * Remove all references to RSAref, since there's no more need for it.
14480 * Make DSO load along a path given through an environment variable
14481 (SHLIB_PATH) with shl_load().
14485 * Constify the ENGINE code as a result of BIGNUM constification.
14486 Also constify the RSA code and most things related to it. In a
14487 few places, most notable in the depth of the ASN.1 code, ugly
14488 casts back to non-const were required (to be solved at a later
14493 * Make it so the openssl application has all engines loaded by default.
14497 * Constify the BIGNUM routines a little more.
14501 * Add the following functions:
14503 ENGINE_load_cswift()
14505 ENGINE_load_atalla()
14506 ENGINE_load_nuron()
14507 ENGINE_load_builtin_engines()
14509 That way, an application can itself choose if external engines that
14510 are built-in in OpenSSL shall ever be used or not. The benefit is
14511 that applications won't have to be linked with libdl or other dso
14512 libraries unless it's really needed.
14514 Changed 'openssl engine' to load all engines on demand.
14515 Changed the engine header files to avoid the duplication of some
14516 declarations (they differed!).
14520 * 'openssl engine' can now list capabilities.
14524 * Better error reporting in 'openssl engine'.
14528 * Never call load_dh_param(NULL) in s_server.
14532 * Add engine application. It can currently list engines by name and
14533 identity, and test if they are actually available.
14537 * Improve RPM specification file by forcing symbolic linking and making
14538 sure the installed documentation is also owned by root.root.
14540 *Damien Miller <djm@mindrot.org>*
14542 * Give the OpenSSL applications more possibilities to make use of
14543 keys (public as well as private) handled by engines.
14547 * Add OCSP code that comes from CertCo.
14551 * Add VMS support for the Rijndael code.
14555 * Added untested support for Nuron crypto accelerator.
14559 * Add support for external cryptographic devices. This code was
14560 previously distributed separately as the "engine" branch.
14562 *Geoff Thorpe, Richard Levitte*
14564 * Rework the filename-translation in the DSO code. It is now possible to
14565 have far greater control over how a "name" is turned into a filename
14566 depending on the operating environment and any oddities about the
14567 different shared library filenames on each system.
14571 * Support threads on FreeBSD-elf in Configure.
14575 * Fix for SHA1 assembly problem with MASM: it produces
14576 warnings about corrupt line number information when assembling
14577 with debugging information. This is caused by the overlapping
14580 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14583 NCONF_get_number() has no error checking at all. As a replacement,
14584 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14585 promoted strongly. The old NCONF_get_number is kept around for
14586 binary backward compatibility.
14587 Make it possible for methods to load from something other than a BIO,
14588 by providing a function pointer that is given a name instead of a BIO.
14589 For example, this could be used to load configuration data from an
14594 * Fix for non blocking accept BIOs. Added new I/O special reason
14595 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14596 with non blocking I/O was not possible because no retry code was
14597 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14602 * Added the beginnings of Rijndael support.
14606 * Fix for bug in DirectoryString mask setting. Add support for
14607 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14608 to allow certificate printing to more controllable, additional
14609 'certopt' option to 'x509' to allow new printing options to be
14614 * Clean old EAY MD5 hack from e_os.h.
14618 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14620 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14621 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14623 *Joe Orton, Steve Henson*
14625 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14627 * Fix additional bug revealed by the NISCC test suite:
14629 Stop bug triggering large recursion when presented with
14630 certain ASN.1 tags ([CVE-2003-0851])
14634 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14636 * Fix various bugs revealed by running the NISCC test suite:
14638 Stop out of bounds reads in the ASN1 code when presented with
14639 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14641 If verify callback ignores invalid public key errors don't try to check
14642 certificate signature with the NULL public key.
14646 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14647 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14652 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14653 extra data after the compression methods not only for TLS 1.0
14654 but also for SSL 3.0 (as required by the specification).
14656 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14658 * Change X509_certificate_type() to mark the key as exported/exportable
14659 when it's 512 *bits* long, not 512 bytes.
14663 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14665 * Countermeasure against the Klima-Pokorny-Rosa extension of
14666 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14667 a protocol version number mismatch like a decryption error
14668 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14672 * Turn on RSA blinding by default in the default implementation
14673 to avoid a timing attack. Applications that don't want it can call
14674 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14675 They would be ill-advised to do so in most cases.
14677 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14679 * Change RSA blinding code so that it works when the PRNG is not
14680 seeded (in this case, the secret RSA exponent is abused as
14681 an unpredictable seed -- if it is not unpredictable, there
14682 is no point in blinding anyway). Make RSA blinding thread-safe
14683 by remembering the creator's thread ID in rsa->blinding and
14684 having all other threads use local one-time blinding factors
14685 (this requires more computation than sharing rsa->blinding, but
14686 avoids excessive locking; and if an RSA object is not shared
14687 between threads, blinding will still be very fast).
14691 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14693 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14694 via timing by performing a MAC computation even if incorrect
14695 block cipher padding has been found. This is a countermeasure
14696 against active attacks where the attacker has to distinguish
14697 between bad padding and a MAC verification error. ([CVE-2003-0078])
14699 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14700 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14701 Martin Vuagnoux (EPFL, Ilion)*
14703 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14705 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14706 memory from its contents. This is done with a counter that will
14707 place alternating values in each byte. This can be used to solve
14708 two issues: 1) the removal of calls to memset() by highly optimizing
14709 compilers, and 2) cleansing with other values than 0, since those can
14710 be read through on certain media, for example a swap space on disk.
14714 * Bugfix: client side session caching did not work with external caching,
14715 because the session->cipher setting was not restored when reloading
14716 from the external cache. This problem was masked, when
14717 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14718 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14722 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14723 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14725 *Zeev Lieber <zeev-l@yahoo.com>*
14727 * Undo an undocumented change introduced in 0.9.6e which caused
14728 repeated calls to OpenSSL_add_all_ciphers() and
14729 OpenSSL_add_all_digests() to be ignored, even after calling
14734 * Change the default configuration reader to deal with last line not
14735 being properly terminated.
14739 * Change X509_NAME_cmp() so it applies the special rules on handling
14740 DN values that are of type PrintableString, as well as RDNs of type
14741 emailAddress where the value has the type ia5String.
14743 *stefank@valicert.com via Richard Levitte*
14745 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14746 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14747 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14748 the bitwise-OR of the two for use by the majority of applications
14749 wanting this behaviour, and update the docs. The documented
14750 behaviour and actual behaviour were inconsistent and had been
14751 changing anyway, so this is more a bug-fix than a behavioural
14754 *Geoff Thorpe, diagnosed by Nadav Har'El*
14756 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14757 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14761 * Fix initialization code race conditions in
14762 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14763 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14764 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14765 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14766 ssl2_get_cipher_by_char(),
14767 ssl3_get_cipher_by_char().
14769 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14771 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14772 the cached sessions are flushed, as the remove_cb() might use ex_data
14773 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14774 (see [openssl.org #212]).
14776 *Geoff Thorpe, Lutz Jaenicke*
14778 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14779 length, instead of the encoding length to d2i_ASN1_OBJECT.
14783 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14785 * [In 0.9.6g-engine release:]
14786 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14788 *Lynn Gazis <lgazis@rainbow.com>*
14790 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14792 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14793 and get fix the header length calculation.
14794 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14795 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14797 * Use proper error handling instead of 'assertions' in buffer
14798 overflow checks added in 0.9.6e. This prevents DoS (the
14799 assertions could call abort()).
14801 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14803 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14805 * Add various sanity checks to asn1_get_length() to reject
14806 the ASN1 length bytes if they exceed sizeof(long), will appear
14807 negative or the content length exceeds the length of the
14810 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14812 * Fix cipher selection routines: ciphers without encryption had no flags
14813 for the cipher strength set and where therefore not handled correctly
14814 by the selection routines (PR #130).
14818 * Fix EVP_dsa_sha macro.
14823 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14824 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14825 that was added in OpenSSL 0.9.6d.
14827 As the countermeasure turned out to be incompatible with some
14828 broken SSL implementations, the new option is part of SSL_OP_ALL.
14829 SSL_OP_ALL is usually employed when compatibility with weird SSL
14830 implementations is desired (e.g. '-bugs' option to 's_client' and
14831 's_server'), so the new option is automatically set in many
14836 * Changes in security patch:
14838 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14839 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14840 Air Force Materiel Command, USAF, under agreement number
14843 * Add various sanity checks to asn1_get_length() to reject
14844 the ASN1 length bytes if they exceed sizeof(long), will appear
14845 negative or the content length exceeds the length of the
14846 supplied buffer. ([CVE-2002-0659])
14848 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14850 * Assertions for various potential buffer overflows, not known to
14851 happen in practice.
14853 *Ben Laurie (CHATS)*
14855 * Various temporary buffers to hold ASCII versions of integers were
14856 too small for 64 bit platforms. ([CVE-2002-0655])
14857 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14859 * Remote buffer overflow in SSL3 protocol - an attacker could
14860 supply an oversized session ID to a client. ([CVE-2002-0656])
14862 *Ben Laurie (CHATS)*
14864 * Remote buffer overflow in SSL2 protocol - an attacker could
14865 supply an oversized client master key. ([CVE-2002-0656])
14867 *Ben Laurie (CHATS)*
14869 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14871 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14872 encoded as NULL) with id-dsa-with-sha1.
14874 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14876 * Check various `X509_...()` return values in `apps/req.c`.
14878 *Nils Larsch <nla@trustcenter.de>*
14880 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14881 an end-of-file condition would erroneously be flagged, when the CRLF
14882 was just at the end of a processed block. The bug was discovered when
14883 processing data through a buffering memory BIO handing the data to a
14884 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14885 <ptsekov@syntrex.com> and Nedelcho Stanev.
14889 * Implement a countermeasure against a vulnerability recently found
14890 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14891 before application data chunks to avoid the use of known IVs
14892 with data potentially chosen by the attacker.
14896 * Fix length checks in ssl3_get_client_hello().
14900 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14901 to prevent ssl3_read_internal() from incorrectly assuming that
14902 ssl3_read_bytes() found application data while handshake
14903 processing was enabled when in fact s->s3->in_read_app_data was
14904 merely automatically cleared during the initial handshake.
14906 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14908 * Fix object definitions for Private and Enterprise: they were not
14909 recognized in their shortname (=lowercase) representation. Extend
14910 obj_dat.pl to issue an error when using undefined keywords instead
14911 of silently ignoring the problem (Svenning Sorensen
14912 <sss@sss.dnsalias.net>).
14916 * Fix DH_generate_parameters() so that it works for 'non-standard'
14917 generators, i.e. generators other than 2 and 5. (Previously, the
14918 code did not properly initialise the 'add' and 'rem' values to
14919 BN_generate_prime().)
14921 In the new general case, we do not insist that 'generator' is
14922 actually a primitive root: This requirement is rather pointless;
14923 a generator of the order-q subgroup is just as good, if not
14928 * Map new X509 verification errors to alerts. Discovered and submitted by
14929 Tom Wu <tom@arcot.com>.
14933 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14934 returning non-zero before the data has been completely received
14935 when using non-blocking I/O.
14937 *Bodo Moeller; problem pointed out by John Hughes*
14939 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14941 *Ben Laurie, Lutz Jaenicke*
14943 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14944 Yoram Zahavi <YoramZ@gilian.com>).
14948 * Add information about CygWin 1.3 and on, and preserve proper
14949 configuration for the versions before that.
14951 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14953 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14954 check whether we deal with a copy of a session and do not delete from
14955 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14956 <izhar@checkpoint.com>.
14960 * Do not store session data into the internal session cache, if it
14961 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14962 flag is set). Proposed by Aslam <aslam@funk.com>.
14966 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14971 * [In 0.9.6d-engine release:]
14972 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14974 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14976 * Add the configuration target linux-s390x.
14978 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14980 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14981 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14982 variable as an indication that a ClientHello message has been
14983 received. As the flag value will be lost between multiple
14984 invocations of ssl3_accept when using non-blocking I/O, the
14985 function may not be aware that a handshake has actually taken
14986 place, thus preventing a new session from being added to the
14989 To avoid this problem, we now set s->new_session to 2 instead of
14990 using a local variable.
14992 *Lutz Jaenicke, Bodo Moeller*
14994 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14995 if the SSL_R_LENGTH_MISMATCH error is detected.
14997 *Geoff Thorpe, Bodo Moeller*
14999 * New 'shared_ldflag' column in Configure platform table.
15003 * Fix EVP_CIPHER_mode macro.
15005 *"Dan S. Camper" <dan@bti.net>*
15007 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15008 type, we must throw them away by setting rr->length to 0.
15010 *D P Chang <dpc@qualys.com>*
15012 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15014 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15015 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15016 worked incorrectly for those cases where range = `10..._2` and
15017 `3*range` is two bits longer than range.)
15021 * Only add signing time to PKCS7 structures if it is not already
15026 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15027 OBJ_ld_ce should be OBJ_id_ce.
15028 Also some ip-pda OIDs in crypto/objects/objects.txt were
15029 incorrect (cf. RFC 3039).
15031 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15033 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15034 returns early because it has nothing to do.
15036 *Andy Schneider <andy.schneider@bjss.co.uk>*
15038 * [In 0.9.6c-engine release:]
15039 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15041 *Andy Schneider <andy.schneider@bjss.co.uk>*
15043 * [In 0.9.6c-engine release:]
15044 Add support for Cryptographic Appliance's keyserver technology.
15045 (Use engine 'keyclient')
15047 *Cryptographic Appliances and Geoff Thorpe*
15049 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15050 is called via tools/c89.sh because arguments have to be
15051 rearranged (all '-L' options must appear before the first object
15054 *Richard Shapiro <rshapiro@abinitio.com>*
15056 * [In 0.9.6c-engine release:]
15057 Add support for Broadcom crypto accelerator cards, backported
15060 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15062 * [In 0.9.6c-engine release:]
15063 Add support for SureWare crypto accelerator cards from
15064 Baltimore Technologies. (Use engine 'sureware')
15066 *Baltimore Technologies and Mark Cox*
15068 * [In 0.9.6c-engine release:]
15069 Add support for crypto accelerator cards from Accelerated
15070 Encryption Processing, www.aep.ie. (Use engine 'aep')
15072 *AEP Inc. and Mark Cox*
15074 * Add a configuration entry for gcc on UnixWare.
15076 *Gary Benson <gbenson@redhat.com>*
15078 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15079 messages are stored in a single piece (fixed-length part and
15080 variable-length part combined) and fix various bugs found on the way.
15084 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15085 instead. BIO_gethostbyname() does not know what timeouts are
15086 appropriate, so entries would stay in cache even when they have
15088 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15090 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15091 faced with a pathologically small ClientHello fragment that does
15092 not contain client_version: Instead of aborting with an error,
15093 simply choose the highest available protocol version (i.e.,
15094 TLS 1.0 unless it is disabled). In practice, ClientHello
15095 messages are never sent like this, but this change gives us
15096 strictly correct behaviour at least for TLS.
15100 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15101 never resets s->method to s->ctx->method when called from within
15102 one of the SSL handshake functions.
15104 *Bodo Moeller; problem pointed out by Niko Baric*
15106 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15107 (sent using the client's version number) if client_version is
15108 smaller than the protocol version in use. Also change
15109 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15110 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15111 the client will at least see that alert.
15115 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15120 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15121 client receives HelloRequest while in a handshake.
15123 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15125 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15126 should end in 'break', not 'goto end' which circumvents various
15127 cleanups done in state SSL_ST_OK. But session related stuff
15128 must be disabled for SSL_ST_OK in the case that we just sent a
15131 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15132 before just sending a HelloRequest.
15134 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15136 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15137 reveal whether illegal block cipher padding was found or a MAC
15138 verification error occurred. (Neither SSLerr() codes nor alerts
15139 are directly visible to potential attackers, but the information
15140 may leak via logfiles.)
15142 Similar changes are not required for the SSL 2.0 implementation
15143 because the number of padding bytes is sent in clear for SSL 2.0,
15144 and the extra bytes are just ignored. However ssl/s2_pkt.c
15145 failed to verify that the purported number of padding bytes is in
15150 * Add OpenUNIX-8 support including shared libraries
15151 (Boyd Lynn Gerber <gerberb@zenez.com>).
15155 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15156 'wristwatch attack' using huge encoding parameters (cf.
15157 James H. Manger's CRYPTO 2001 paper). Note that the
15158 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15159 encoding parameters and hence was not vulnerable.
15163 * BN_sqr() bug fix.
15165 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15167 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15168 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15169 followed by modular reduction.
15171 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15173 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15174 equivalent based on BN_pseudo_rand() instead of BN_rand().
15178 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15179 This function was broken, as the check for a new client hello message
15180 to handle SGC did not allow these large messages.
15181 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15185 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15189 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15190 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15194 * Rework the configuration and shared library support for Tru64 Unix.
15195 The configuration part makes use of modern compiler features and
15196 still retains old compiler behavior for those that run older versions
15197 of the OS. The shared library support part includes a variant that
15198 uses the RPATH feature, and is available through the special
15199 configuration target "alpha-cc-rpath", which will never be selected
15202 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15204 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15205 with the same message size as in ssl3_get_certificate_request().
15206 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15207 messages might inadvertently be reject as too long.
15209 *Petr Lampa <lampa@fee.vutbr.cz>*
15211 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15215 * Modified SSL library such that the verify_callback that has been set
15216 specifically for an SSL object with SSL_set_verify() is actually being
15217 used. Before the change, a verify_callback set with this function was
15218 ignored and the verify_callback() set in the SSL_CTX at the time of
15219 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15220 to allow the necessary settings.
15224 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15225 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15226 done automatically (in contradiction to the requirements of the C
15227 standard). This made problems when used from OpenSSH.
15231 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15232 dh->length and always used
15234 BN_rand_range(priv_key, dh->p).
15236 BN_rand_range() is not necessary for Diffie-Hellman, and this
15237 specific range makes Diffie-Hellman unnecessarily inefficient if
15238 dh->length (recommended exponent length) is much smaller than the
15239 length of dh->p. We could use BN_rand_range() if the order of
15240 the subgroup was stored in the DH structure, but we only have
15245 BN_rand(priv_key, l, ...)
15247 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15254 RSA_eay_public_encrypt
15255 RSA_eay_private_decrypt
15256 RSA_eay_private_encrypt (signing)
15257 RSA_eay_public_decrypt (signature verification)
15259 (default implementations for RSA_public_encrypt,
15260 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15261 always reject numbers >= n.
15265 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15266 to synchronize access to 'locking_thread'. This is necessary on
15267 systems where access to 'locking_thread' (an 'unsigned long'
15268 variable) is not atomic.
15272 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15273 *before* setting the 'crypto_lock_rand' flag. The previous code had
15274 a race condition if 0 is a valid thread ID.
15276 *Travis Vitek <vitek@roguewave.com>*
15278 * Add support for shared libraries under Irix.
15280 *Albert Chin-A-Young <china@thewrittenword.com>*
15282 * Add configuration option to build on Linux on both big-endian and
15283 little-endian MIPS.
15285 *Ralf Baechle <ralf@uni-koblenz.de>*
15287 * Add the possibility to create shared libraries on HP-UX.
15291 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15293 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15294 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15295 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15296 PRNG state recovery was possible based on the output of
15297 one PRNG request appropriately sized to gain knowledge on
15298 'md' followed by enough consecutive 1-byte PRNG requests
15299 to traverse all of 'state'.
15301 1. When updating 'md_local' (the current thread's copy of 'md')
15302 during PRNG output generation, hash all of the previous
15303 'md_local' value, not just the half used for PRNG output.
15305 2. Make the number of bytes from 'state' included into the hash
15306 independent from the number of PRNG bytes requested.
15308 The first measure alone would be sufficient to avoid
15309 Markku-Juhani's attack. (Actually it had never occurred
15310 to me that the half of 'md_local' used for chaining was the
15311 half from which PRNG output bytes were taken -- I had always
15312 assumed that the secret half would be used.) The second
15313 measure makes sure that additional data from 'state' is never
15314 mixed into 'md_local' in small portions; this heuristically
15315 further strengthens the PRNG.
15319 * Fix crypto/bn/asm/mips3.s.
15323 * When only the key is given to "enc", the IV is undefined. Print out
15324 an error message in this case.
15328 * Handle special case when X509_NAME is empty in X509 printing routines.
15332 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15333 positive and less than q.
15337 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15338 used: it isn't thread safe and the add_lock_callback should handle
15341 *Paul Rose <Paul.Rose@bridge.com>*
15343 * Verify that incoming data obeys the block size in
15344 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15350 *Ulf Möller, Bodo Möller*
15352 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15353 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15354 when fixing the server behaviour for backwards-compatible 'client
15355 hello' messages. (Note that the attack is impractical against
15356 SSL 3.0 and TLS 1.0 anyway because length and version checking
15357 means that the probability of guessing a valid ciphertext is
15358 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15361 Before 0.9.5, the countermeasure (hide the error by generating a
15362 random 'decryption result') did not work properly because
15363 ERR_clear_error() was missing, meaning that SSL_get_error() would
15364 detect the supposedly ignored error.
15366 Both problems are now fixed.
15370 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15371 (previously it was 1024).
15375 * Fix for compatibility mode trust settings: ignore trust settings
15376 unless some valid trust or reject settings are present.
15380 * Fix for blowfish EVP: its a variable length cipher.
15384 * Fix various bugs related to DSA S/MIME verification. Handle missing
15385 parameters in DSA public key structures and return an error in the
15386 DSA routines if parameters are absent.
15390 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15391 in the current directory if neither $RANDFILE nor $HOME was set.
15392 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15393 caused some confusion to Windows users who haven't defined $HOME.
15394 Thus RAND_file_name() is changed again: e_os.h can define a
15395 DEFAULT_HOME, which will be used if $HOME is not set.
15396 For Windows, we use "C:"; on other platforms, we still require
15397 environment variables.
15399 * Move 'if (!initialized) RAND_poll()' into regions protected by
15400 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15401 having multiple threads call RAND_poll() concurrently.
15405 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15406 combination of a flag and a thread ID variable.
15407 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15408 flag), *other* threads can enter ssleay_add_bytes without obeying
15409 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15410 that they do not hold after the first thread unsets add_do_not_lock).
15414 * Change bctest again: '-x' expressions are not available in all
15415 versions of 'test'.
15419 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15421 * Fix a couple of memory leaks in PKCS7_dataDecode()
15423 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15425 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15426 the default extension for executables, if any. Also, make the perl
15427 scripts that use symlink() to test if it really exists and use "cp"
15428 if it doesn't. All this made OpenSSL compilable and installable in
15433 * Fix for asn1_GetSequence() for indefinite length constructed data.
15434 If SEQUENCE is length is indefinite just set c->slen to the total
15435 amount of data available.
15437 *Steve Henson, reported by shige@FreeBSD.org*
15439 *This change does not apply to 0.9.7.*
15441 * Change bctest to avoid here-documents inside command substitution
15442 (workaround for FreeBSD /bin/sh bug).
15443 For compatibility with Ultrix, avoid shell functions (introduced
15444 in the bctest version that searches along $PATH).
15448 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15449 with des_encrypt() defined on some operating systems, like Solaris
15454 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15455 On the Importance of Eliminating Errors in Cryptographic
15456 Computations, J. Cryptology 14 (2001) 2, 101-119,
15457 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15461 * MIPS assembler BIGNUM division bug fix.
15465 * Disabled incorrect Alpha assembler code.
15469 * Fix PKCS#7 decode routines so they correctly update the length
15470 after reading an EOC for the EXPLICIT tag.
15474 *This change does not apply to 0.9.7.*
15476 * Fix bug in PKCS#12 key generation routines. This was triggered
15477 if a 3DES key was generated with a 0 initial byte. Include
15478 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15479 (but broken) behaviour.
15483 * Enhance bctest to search for a working bc along $PATH and print
15486 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15488 * Fix memory leaks in err.c: free err_data string if necessary;
15489 don't write to the wrong index in ERR_set_error_data.
15493 * Implement ssl23_peek (analogous to ssl23_read), which previously
15498 * Replace rdtsc with `_emit` statements for VC++ version 5.
15500 *Jeremy Cooper <jeremy@baymoo.org>*
15502 * Make it possible to reuse SSLv2 sessions.
15506 * In copy_email() check for >= 0 as a return value for
15507 X509_NAME_get_index_by_NID() since 0 is a valid index.
15509 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15511 * Avoid coredump with unsupported or invalid public keys by checking if
15512 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15513 PKCS7_verify() fails with non detached data.
15517 * Don't use getenv in library functions when run as setuid/setgid.
15518 New function OPENSSL_issetugid().
15522 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15523 due to incorrect handling of multi-threading:
15525 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15527 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15529 3. Count how many times MemCheck_off() has been called so that
15530 nested use can be treated correctly. This also avoids
15531 inband-signalling in the previous code (which relied on the
15532 assumption that thread ID 0 is impossible).
15536 * Add "-rand" option also to s_client and s_server.
15540 * Fix CPU detection on Irix 6.x.
15541 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15542 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15544 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15549 *This change does not apply to 0.9.7.*
15551 * Use the cached encoding of an X509_NAME structure rather than
15552 copying it. This is apparently the reason for the libsafe "errors"
15553 but the code is actually correct.
15557 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15558 Bleichenbacher's DSA attack.
15559 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15560 to be set and top=0 forces the highest bit to be set; top=-1 is new
15561 and leaves the highest bit random.
15563 *Ulf Moeller, Bodo Moeller*
15565 * In the `NCONF_...`-based implementations for `CONF_...` queries
15566 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15567 a temporary CONF structure with the data component set to NULL
15568 (which gives segmentation faults in lh_retrieve).
15569 Instead, use NULL for the CONF pointer in CONF_get_string and
15570 CONF_get_number (which may use environment variables) and directly
15571 return NULL from CONF_get_section.
15575 * Fix potential buffer overrun for EBCDIC.
15579 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15580 keyUsage if basicConstraints absent for a CA.
15584 * Make SMIME_write_PKCS7() write mail header values with a format that
15585 is more generally accepted (no spaces before the semicolon), since
15586 some programs can't parse those values properly otherwise. Also make
15587 sure BIO's that break lines after each write do not create invalid
15592 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15593 macros previously used would not encode an empty SEQUENCE OF
15594 and break the signature.
15598 *This change does not apply to 0.9.7.*
15600 * Zero the premaster secret after deriving the master secret in
15605 * Add some EVP_add_digest_alias registrations (as found in
15606 OpenSSL_add_all_digests()) to SSL_library_init()
15607 aka OpenSSL_add_ssl_algorithms(). This provides improved
15608 compatibility with peers using X.509 certificates
15609 with unconventional AlgorithmIdentifier OIDs.
15613 * Fix for Irix with NO_ASM.
15615 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15617 * ./config script fixes.
15619 *Ulf Moeller, Richard Levitte*
15621 * Fix 'openssl passwd -1'.
15625 * Change PKCS12_key_gen_asc() so it can cope with non null
15626 terminated strings whose length is passed in the passlen
15627 parameter, for example from PEM callbacks. This was done
15628 by adding an extra length parameter to asc2uni().
15630 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15632 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15633 call failed, free the DSA structure.
15637 * Fix to uni2asc() to cope with zero length Unicode strings.
15638 These are present in some PKCS#12 files.
15642 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15643 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15644 when writing a 32767 byte record.
15646 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15648 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15649 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15651 (RSA objects have a reference count access to which is protected
15652 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15653 so they are meant to be shared between threads.)
15654 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15655 "Reddie, Steven" <Steven.Reddie@ca.com>*
15657 * Fix a deadlock in CRYPTO_mem_leaks().
15661 * Use better test patterns in bntest.
15665 * rand_win.c fix for Borland C.
15669 * BN_rshift bugfix for n == 0.
15673 * Add a 'bctest' script that checks for some known 'bc' bugs
15674 so that 'make test' does not abort just because 'bc' is broken.
15678 * Store verify_result within SSL_SESSION also for client side to
15679 avoid potential security hole. (Re-used sessions on the client side
15680 always resulted in verify_result==X509_V_OK, not using the original
15681 result of the server certificate verification.)
15685 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15686 SSL3_RT_APPLICATION_DATA, return 0.
15687 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15692 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15693 releases, have been re-implemented by renaming the previous
15694 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15695 and ssl3_read_internal, respectively, and adding 'peek' parameters
15696 to them. The new ssl[23]_{read,peek} functions are calls to
15697 ssl[23]_read_internal with the 'peek' flag set appropriately.
15698 A 'peek' parameter has also been added to ssl3_read_bytes, which
15699 does the actual work for ssl3_read_internal.
15703 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15704 the method-specific "init()" handler. Also clean up ex_data after
15705 calling the method-specific "finish()" handler. Previously, this was
15706 happening the other way round.
15710 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15711 The previous value, 12, was not always sufficient for BN_mod_exp().
15715 * Make sure that shared libraries get the internal name engine with
15716 the full version number and not just 0. This should mark the
15717 shared libraries as not backward compatible. Of course, this should
15718 be changed again when we can guarantee backward binary compatibility.
15722 * Fix typo in get_cert_by_subject() in by_dir.c
15724 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15726 * Rework the system to generate shared libraries:
15728 - Make note of the expected extension for the shared libraries and
15729 if there is a need for symbolic links from for example libcrypto.so.0
15730 to libcrypto.so.0.9.7. There is extended info in Configure for
15733 - Make as few rebuilds of the shared libraries as possible.
15735 - Still avoid linking the OpenSSL programs with the shared libraries.
15737 - When installing, install the shared libraries separately from the
15742 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15744 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15745 and not in SSL_clear because the latter is also used by the
15746 accept/connect functions; previously, the settings made by
15747 SSL_set_read_ahead would be lost during the handshake.
15749 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15751 * Correct util/mkdef.pl to be selective about disabled algorithms.
15752 Previously, it would create entries for disabled algorithms no
15757 * Added several new manual pages for SSL_* function.
15761 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15763 * In ssl23_get_client_hello, generate an error message when faced
15764 with an initial SSL 3.0/TLS record that is too small to contain the
15765 first two bytes of the ClientHello message, i.e. client_version.
15766 (Note that this is a pathologic case that probably has never happened
15767 in real life.) The previous approach was to use the version number
15768 from the record header as a substitute; but our protocol choice
15769 should not depend on that one because it is not authenticated
15770 by the Finished messages.
15774 * More robust randomness gathering functions for Windows.
15776 *Jeffrey Altman <jaltman@columbia.edu>*
15778 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15779 not set then we don't setup the error code for issuer check errors
15780 to avoid possibly overwriting other errors which the callback does
15781 handle. If an application does set the flag then we assume it knows
15782 what it is doing and can handle the new informational codes
15787 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15788 a general "ANY" type, as such it should be able to decode anything
15789 including tagged types. However it didn't check the class so it would
15790 wrongly interpret tagged types in the same way as their universal
15791 counterpart and unknown types were just rejected. Changed so that the
15792 tagged and unknown types are handled in the same way as a SEQUENCE:
15793 that is the encoding is stored intact. There is also a new type
15794 "V_ASN1_OTHER" which is used when the class is not universal, in this
15795 case we have no idea what the actual type is so we just lump them all
15800 * On VMS, stdout may very well lead to a file that is written to
15801 in a record-oriented fashion. That means that every write() will
15802 write a separate record, which will be read separately by the
15803 programs trying to read from it. This can be very confusing.
15805 The solution is to put a BIO filter in the way that will buffer
15806 text until a linefeed is reached, and then write everything a
15807 line at a time, so every record written will be an actual line,
15808 not chunks of lines and not (usually doesn't happen, but I've
15809 seen it once) several lines in one record. BIO_f_linebuffer() is
15812 Currently, it's a VMS-only method, because that's where it has
15813 been tested well enough.
15817 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15818 it can return incorrect results.
15819 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15820 but it was in 0.9.6-beta[12].)
15824 * Disable the check for content being present when verifying detached
15825 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15826 include zero length content when signing messages.
15830 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15831 BIO_ctrl (for BIO pairs).
15835 * Add DSO method for VMS.
15839 * Bug fix: Montgomery multiplication could produce results with the
15844 * Add RPM specification openssl.spec and modify it to build three
15845 packages. The default package contains applications, application
15846 documentation and run-time libraries. The devel package contains
15847 include files, static libraries and function documentation. The
15848 doc package contains the contents of the doc directory. The original
15849 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15853 * Add a large number of documentation files for many SSL routines.
15855 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15857 * Add a configuration entry for Sony News 4.
15859 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15861 * Don't set the two most significant bits to one when generating a
15862 random number < q in the DSA library.
15866 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15867 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15868 the underlying transport is blocking) if a handshake took place.
15869 (The default behaviour is needed by applications such as s_client
15870 and s_server that use select() to determine when to use SSL_read;
15871 but for applications that know in advance when to expect data, it
15872 just makes things more complicated.)
15876 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15881 * Add a few more EBCDIC conditionals that make `req` and `x509`
15882 work better on such systems.
15884 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15886 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15887 Update PKCS12_parse() so it copies the friendlyName and the
15888 keyid to the certificates aux info.
15892 * Fix bug in PKCS7_verify() which caused an infinite loop
15893 if there was more than one signature.
15895 *Sven Uszpelkat <su@celocom.de>*
15897 * Major change in util/mkdef.pl to include extra information
15898 about each symbol, as well as presenting variables as well
15899 as functions. This change means that there's n more need
15900 to rebuild the .num files when some algorithms are excluded.
15904 * Allow the verify time to be set by an application,
15905 rather than always using the current time.
15909 * Phase 2 verify code reorganisation. The certificate
15910 verify code now looks up an issuer certificate by a
15911 number of criteria: subject name, authority key id
15912 and key usage. It also verifies self signed certificates
15913 by the same criteria. The main comparison function is
15914 X509_check_issued() which performs these checks.
15916 Lot of changes were necessary in order to support this
15917 without completely rewriting the lookup code.
15919 Authority and subject key identifier are now cached.
15921 The LHASH 'certs' is X509_STORE has now been replaced
15922 by a STACK_OF(X509_OBJECT). This is mainly because an
15923 LHASH can't store or retrieve multiple objects with
15924 the same hash value.
15926 As a result various functions (which were all internal
15927 use only) have changed to handle the new X509_STORE
15928 structure. This will break anything that messed round
15929 with X509_STORE internally.
15931 The functions X509_STORE_add_cert() now checks for an
15932 exact match, rather than just subject name.
15934 The X509_STORE API doesn't directly support the retrieval
15935 of multiple certificates matching a given criteria, however
15936 this can be worked round by performing a lookup first
15937 (which will fill the cache with candidate certificates)
15938 and then examining the cache for matches. This is probably
15939 the best we can do without throwing out X509_LOOKUP
15940 entirely (maybe later...).
15942 The X509_VERIFY_CTX structure has been enhanced considerably.
15944 All certificate lookup operations now go via a get_issuer()
15945 callback. Although this currently uses an X509_STORE it
15946 can be replaced by custom lookups. This is a simple way
15947 to bypass the X509_STORE hackery necessary to make this
15948 work and makes it possible to use more efficient techniques
15949 in future. A very simple version which uses a simple
15950 STACK for its trusted certificate store is also provided
15951 using X509_STORE_CTX_trusted_stack().
15953 The verify_cb() and verify() callbacks now have equivalents
15954 in the X509_STORE_CTX structure.
15956 X509_STORE_CTX also has a 'flags' field which can be used
15957 to customise the verify behaviour.
15961 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15962 excludes S/MIME capabilities.
15966 * When a certificate request is read in keep a copy of the
15967 original encoding of the signed data and use it when outputting
15968 again. Signatures then use the original encoding rather than
15969 a decoded, encoded version which may cause problems if the
15970 request is improperly encoded.
15974 * For consistency with other BIO_puts implementations, call
15975 buffer_write(b, ...) directly in buffer_puts instead of calling
15978 In BIO_puts, increment b->num_write as in BIO_write.
15980 *Peter.Sylvester@EdelWeb.fr*
15982 * Fix BN_mul_word for the case where the word is 0. (We have to use
15983 BN_zero, we may not return a BIGNUM with an array consisting of
15984 words set to zero.)
15988 * Avoid calling abort() from within the library when problems are
15989 detected, except if preprocessor symbols have been defined
15990 (such as REF_CHECK, BN_DEBUG etc.).
15994 * New openssl application 'rsautl'. This utility can be
15995 used for low-level RSA operations. DER public key
15996 BIO/fp routines also added.
16000 * New Configure entry and patches for compiling on QNX 4.
16002 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16004 * A demo state-machine implementation was sponsored by
16005 Nuron (<http://www.nuron.com/>) and is now available in
16006 demos/state_machine.
16010 * New options added to the 'dgst' utility for signature
16011 generation and verification.
16015 * Unrecognized PKCS#7 content types are now handled via a
16016 catch all ASN1_TYPE structure. This allows unsupported
16017 types to be stored as a "blob" and an application can
16018 encode and decode it manually.
16022 * Fix various signed/unsigned issues to make a_strex.c
16023 compile under VC++.
16025 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16027 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16028 length if passed a buffer. ASN1_INTEGER_to_BN failed
16029 if passed a NULL BN and its argument was negative.
16031 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16033 * Modification to PKCS#7 encoding routines to output definite
16034 length encoding. Since currently the whole structures are in
16035 memory there's not real point in using indefinite length
16036 constructed encoding. However if OpenSSL is compiled with
16037 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16041 * Added BIO_vprintf() and BIO_vsnprintf().
16045 * Added more prefixes to parse for in the strings written
16046 through a logging bio, to cover all the levels that are available
16047 through syslog. The prefixes are now:
16049 PANIC, EMERG, EMR => LOG_EMERG
16050 ALERT, ALR => LOG_ALERT
16051 CRIT, CRI => LOG_CRIT
16052 ERROR, ERR => LOG_ERR
16053 WARNING, WARN, WAR => LOG_WARNING
16054 NOTICE, NOTE, NOT => LOG_NOTICE
16055 INFO, INF => LOG_INFO
16056 DEBUG, DBG => LOG_DEBUG
16058 and as before, if none of those prefixes are present at the
16059 beginning of the string, LOG_ERR is chosen.
16061 On Win32, the `LOG_*` levels are mapped according to this:
16063 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16064 LOG_WARNING => EVENTLOG_WARNING_TYPE
16065 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16069 * Made it possible to reconfigure with just the configuration
16070 argument "reconf" or "reconfigure". The command line arguments
16071 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16072 and are retrieved from there when reconfiguring.
16078 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16080 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16084 * The obj_dat.pl script was messing up the sorting of object
16085 names. The reason was that it compared the quoted version
16086 of strings as a result "OCSP" > "OCSP Signing" because
16087 " > SPACE. Changed script to store unquoted versions of
16088 names and add quotes on output. It was also omitting some
16089 names from the lookup table if they were given a default
16090 value (that is if SN is missing it is given the same
16091 value as LN and vice versa), these are now added on the
16092 grounds that if an object has a name we should be able to
16093 look it up. Finally added warning output when duplicate
16094 short or long names are found.
16098 * Changes needed for Tandem NSK.
16100 *Scott Uroff <scott@xypro.com>*
16102 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16103 RSA_padding_check_SSLv23(), special padding was never detected
16104 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16105 version rollback attacks was not effective.
16107 In s23_clnt.c, don't use special rollback-attack detection padding
16108 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16109 client; similarly, in s23_srvr.c, don't do the rollback check if
16110 SSL 2.0 is the only protocol enabled in the server.
16114 * Make it possible to get hexdumps of unprintable data with 'openssl
16115 asn1parse'. By implication, the functions ASN1_parse_dump() and
16116 BIO_dump_indent() are added.
16120 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16121 these print out strings and name structures based on various
16122 flags including RFC2253 support and proper handling of
16123 multibyte characters. Added options to the 'x509' utility
16124 to allow the various flags to be set.
16128 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16129 Also change the functions X509_cmp_current_time() and
16130 X509_gmtime_adj() work with an ASN1_TIME structure,
16131 this will enable certificates using GeneralizedTime in validity
16132 dates to be checked.
16136 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16137 negative public key encodings) on by default,
16138 NO_NEG_PUBKEY_BUG can be set to disable it.
16142 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16143 content octets. An i2c_ASN1_OBJECT is unnecessary because
16144 the encoding can be trivially obtained from the structure.
16148 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16149 not read locks (`CRYPTO_r_[un]lock`).
16153 * A first attempt at creating official support for shared
16154 libraries through configuration. I've kept it so the
16155 default is static libraries only, and the OpenSSL programs
16156 are always statically linked for now, but there are
16157 preparations for dynamic linking in place.
16158 This has been tested on Linux and Tru64.
16162 * Randomness polling function for Win9x, as described in:
16163 Peter Gutmann, Software Generation of Practically Strong
16168 * Fix so PRNG is seeded in req if using an already existing
16173 * New options to smime application. -inform and -outform
16174 allow alternative formats for the S/MIME message including
16175 PEM and DER. The -content option allows the content to be
16176 specified separately. This should allow things like Netscape
16177 form signing output easier to verify.
16181 * Fix the ASN1 encoding of tags using the 'long form'.
16185 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16186 STRING types. These convert content octets to and from the
16187 underlying type. The actual tag and length octets are
16188 already assumed to have been read in and checked. These
16189 are needed because all other string types have virtually
16190 identical handling apart from the tag. By having versions
16191 of the ASN1 functions that just operate on content octets
16192 IMPLICIT tagging can be handled properly. It also allows
16193 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16194 and ASN1_INTEGER are identical apart from the tag.
16198 * Change the handling of OID objects as follows:
16200 - New object identifiers are inserted in objects.txt, following
16201 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16202 - objects.pl is used to process obj_mac.num and create a new
16204 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16207 This is currently kind of a hack, and the perl code in objects.pl
16208 isn't very elegant, but it works as I intended. The simplest way
16209 to check that it worked correctly is to look in obj_dat.h and
16210 check the array nid_objs and make sure the objects haven't moved
16211 around (this is important!). Additions are OK, as well as
16212 consistent name changes.
16216 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16220 * Addition of the command line parameter '-rand file' to 'openssl req'.
16221 The given file adds to whatever has already been seeded into the
16222 random pool through the RANDFILE configuration file option or
16223 environment variable, or the default random state file.
16227 * mkstack.pl now sorts each macro group into lexical order.
16228 Previously the output order depended on the order the files
16229 appeared in the directory, resulting in needless rewriting
16234 * Patches to make OpenSSL compile under Win32 again. Mostly
16235 work arounds for the VC++ problem that it treats func() as
16236 func(void). Also stripped out the parts of mkdef.pl that
16237 added extra typesafe functions: these no longer exist.
16241 * Reorganisation of the stack code. The macros are now all
16242 collected in safestack.h . Each macro is defined in terms of
16243 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16244 DEBUG_SAFESTACK is now handled in terms of function casts,
16245 this has the advantage of retaining type safety without the
16246 use of additional functions. If DEBUG_SAFESTACK is not defined
16247 then the non typesafe macros are used instead. Also modified the
16248 mkstack.pl script to handle the new form. Needs testing to see
16249 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16250 the default if no major problems. Similar behaviour for ASN1_SET_OF
16251 and PKCS12_STACK_OF.
16255 * When some versions of IIS use the 'NET' form of private key the
16256 key derivation algorithm is different. Normally MD5(password) is
16257 used as a 128 bit RC4 key. In the modified case
16258 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16259 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16260 as the old Netscape_RSA functions except they have an additional
16261 'sgckey' parameter which uses the modified algorithm. Also added
16262 an -sgckey command line option to the rsa utility. Thanks to
16263 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16264 algorithm to openssl-dev.
16268 * The evp_local.h macros were using 'c.##kname' which resulted in
16269 invalid expansion on some systems (SCO 5.0.5 for example).
16270 Corrected to 'c.kname'.
16272 *Phillip Porch <root@theporch.com>*
16274 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16275 a STACK of email addresses from a certificate or request, these look
16276 in the subject name and the subject alternative name extensions and
16277 omit any duplicate addresses.
16281 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16282 This makes DSA verification about 2 % faster.
16286 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16287 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16288 plus overhead for 1024 bit moduli).
16289 This makes exponentiations about 0.5 % faster for 1024 bit
16290 exponents (as measured by "openssl speed rsa2048").
16294 * Rename memory handling macros to avoid conflicts with other
16296 Malloc => OPENSSL_malloc
16297 Malloc_locked => OPENSSL_malloc_locked
16298 Realloc => OPENSSL_realloc
16299 Free => OPENSSL_free
16303 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16304 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16308 * CygWin32 support.
16310 *John Jarvie <jjarvie@newsguy.com>*
16312 * The type-safe stack code has been rejigged. It is now only compiled
16313 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16314 by default all type-specific stack functions are "#define"d back to
16315 standard stack functions. This results in more streamlined output
16316 but retains the type-safety checking possibilities of the original
16321 * The STACK code has been cleaned up, and certain type declarations
16322 that didn't make a lot of sense have been brought in line. This has
16323 also involved a cleanup of sorts in safestack.h to more correctly
16324 map type-safe stack functions onto their plain stack counterparts.
16325 This work has also resulted in a variety of "const"ifications of
16326 lots of the code, especially `_cmp` operations which should normally
16327 be prototyped with "const" parameters anyway.
16331 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16332 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16333 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16334 where all of 'md' is used each time the PRNG is used, but 'state'
16335 is used only indexed by a cyclic counter. As entropy may not be
16336 well distributed from the beginning, 'md' is important as a
16337 chaining variable. However, the output function chains only half
16338 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16339 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16340 in all of 'state' being rewritten, with the new values depending
16341 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16345 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16346 the handshake is continued after ssl_verify_cert_chain();
16347 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16348 can lead to 'unexplainable' connection aborts later.
16350 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16352 * Major EVP API cipher revision.
16353 Add hooks for extra EVP features. This allows various cipher
16354 parameters to be set in the EVP interface. Support added for variable
16355 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16356 setting of RC2 and RC5 parameters.
16358 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16361 Remove lots of duplicated code from the EVP library. For example *every*
16362 cipher init() function handles the 'iv' in the same way according to the
16363 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16364 for CFB and OFB modes they zero ctx->num.
16366 New functionality allows removal of S/MIME code RC2 hack.
16368 Most of the routines have the same form and so can be declared in terms
16371 By shifting this to the top level EVP_CipherInit() it can be removed from
16372 all individual ciphers. If the cipher wants to handle IVs or keys
16373 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16376 Change lots of functions like EVP_EncryptUpdate() to now return a
16377 value: although software versions of the algorithms cannot fail
16378 any installed hardware versions can.
16382 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16383 this option is set, tolerate broken clients that send the negotiated
16384 protocol version number instead of the requested protocol version
16389 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16390 i.e. non-zero for export ciphersuites, zero otherwise.
16391 Previous versions had this flag inverted, inconsistent with
16392 rsa_tmp_cb (..._TMP_RSA_CB).
16394 *Bodo Moeller; problem reported by Amit Chopra*
16396 * Add missing DSA library text string. Work around for some IIS
16397 key files with invalid SEQUENCE encoding.
16401 * Add a document (doc/standards.txt) that list all kinds of standards
16402 and so on that are implemented in OpenSSL.
16406 * Enhance c_rehash script. Old version would mishandle certificates
16407 with the same subject name hash and wouldn't handle CRLs at all.
16408 Added -fingerprint option to crl utility, to support new c_rehash
16413 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16417 * Fix for SSL server purpose checking. Server checking was
16418 rejecting certificates which had extended key usage present
16419 but no ssl client purpose.
16421 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16423 * Make PKCS#12 code work with no password. The PKCS#12 spec
16424 is a little unclear about how a blank password is handled.
16425 Since the password in encoded as a BMPString with terminating
16426 double NULL a zero length password would end up as just the
16427 double NULL. However no password at all is different and is
16428 handled differently in the PKCS#12 key generation code. NS
16429 treats a blank password as zero length. MSIE treats it as no
16430 password on export: but it will try both on import. We now do
16431 the same: PKCS12_parse() tries zero length and no password if
16432 the password is set to "" or NULL (NULL is now a valid password:
16433 it wasn't before) as does the pkcs12 application.
16437 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16438 perror when PEM_read_bio_X509_REQ fails, the error message must
16439 be obtained from the error queue.
16443 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16444 it in ERR_remove_state if appropriate, and change ERR_get_state
16445 accordingly to avoid race conditions (this is necessary because
16446 thread_hash is no longer constant once set).
16450 * Bugfix for linux-elf makefile.one.
16454 * RSA_get_default_method() will now cause a default
16455 RSA_METHOD to be chosen if one doesn't exist already.
16456 Previously this was only set during a call to RSA_new()
16457 or RSA_new_method(NULL) meaning it was possible for
16458 RSA_get_default_method() to return NULL.
16462 * Added native name translation to the existing DSO code
16463 that will convert (if the flag to do so is set) filenames
16464 that are sufficiently small and have no path information
16465 into a canonical native form. Eg. "blah" converted to
16466 "libblah.so" or "blah.dll" etc.
16470 * New function ERR_error_string_n(e, buf, len) which is like
16471 ERR_error_string(e, buf), but writes at most 'len' bytes
16472 including the 0 terminator. For ERR_error_string_n, 'buf'
16475 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16477 * CONF library reworked to become more general. A new CONF
16478 configuration file reader "class" is implemented as well as a
16479 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16480 old `CONF_*` functions are still there, but are reimplemented to
16481 work in terms of the new functions. Also, a set of functions
16482 to handle the internal storage of the configuration data is
16483 provided to make it easier to write new configuration file
16484 reader "classes" (I can definitely see something reading a
16485 configuration file in XML format, for example), called `_CONF_*`,
16486 or "the configuration storage API"...
16488 The new configuration file reading functions are:
16490 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16491 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16493 NCONF_default, NCONF_WIN32
16495 NCONF_dump_fp, NCONF_dump_bio
16497 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16498 NCONF_new creates a new CONF object. This works in the same way
16499 as other interfaces in OpenSSL, like the BIO interface.
16500 `NCONF_dump_*` dump the internal storage of the configuration file,
16501 which is useful for debugging. All other functions take the same
16502 arguments as the old `CONF_*` functions with the exception of the
16503 first that must be a `CONF *` instead of a `LHASH *`.
16505 To make it easier to use the new classes with the old `CONF_*` functions,
16506 the function CONF_set_default_method is provided.
16510 * Add '-tls1' option to 'openssl ciphers', which was already
16511 mentioned in the documentation but had not been implemented.
16512 (This option is not yet really useful because even the additional
16513 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16517 * Initial DSO code added into libcrypto for letting OpenSSL (and
16518 OpenSSL-based applications) load shared libraries and bind to
16519 them in a portable way.
16521 *Geoff Thorpe, with contributions from Richard Levitte*
16523 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16525 * Make sure _lrotl and _lrotr are only used with MSVC.
16527 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16528 (the default implementation of RAND_status).
16530 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16531 to '-clrext' (= clear extensions), as intended and documented.
16532 *Bodo Moeller; inconsistency pointed out by Michael Attili
16533 <attili@amaxo.com>*
16535 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16536 was larger than the MD block size.
16538 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16540 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16541 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16542 using the passed key: if the passed key was a private key the result
16543 of X509_print(), for example, would be to print out all the private key
16548 * des_quad_cksum() byte order bug fix.
16549 *Ulf Möller, using the problem description in krb4-0.9.7, where
16550 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16552 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16555 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16557 * For easily testing in shell scripts whether some command
16558 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16559 returns with exit code 0 iff no command of the given name is available.
16560 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16561 the output goes to stdout and nothing is printed to stderr.
16562 Additional arguments are always ignored.
16564 Since for each cipher there is a command of the same name,
16565 the 'no-cipher' compilation switches can be tested this way.
16567 ('openssl no-XXX' is not able to detect pseudo-commands such
16568 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16572 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16576 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16577 is set; it will be thrown away anyway because each handshake creates
16579 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16580 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16581 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16582 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16586 * New s_client option -ign_eof: EOF at stdin is ignored, and
16587 'Q' and 'R' lose their special meanings (quit/renegotiate).
16588 This is part of what -quiet does; unlike -quiet, -ign_eof
16589 does not suppress any output.
16593 * Add compatibility options to the purpose and trust code. The
16594 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16595 accepts a certificate or CA, this was the previous behaviour,
16596 with all the associated security issues.
16598 X509_TRUST_COMPAT is the old trust behaviour: only and
16599 automatically trust self signed roots in certificate store. A
16600 new trust setting X509_TRUST_DEFAULT is used to specify that
16601 a purpose has no associated trust setting and it should instead
16602 use the value in the default purpose.
16606 * Fix the PKCS#8 DSA private key code so it decodes keys again
16607 and fix a memory leak.
16611 * In util/mkerr.pl (which implements 'make errors'), preserve
16612 reason strings from the previous version of the .c file, as
16613 the default to have only downcase letters (and digits) in
16614 automatically generated reasons codes is not always appropriate.
16618 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16619 using strerror. Previously, ERR_reason_error_string() returned
16620 library names as reason strings for SYSerr; but SYSerr is a special
16621 case where small numbers are errno values, not library numbers.
16625 * Add '-dsaparam' option to 'openssl dhparam' application. This
16626 converts DSA parameters into DH parameters. (When creating parameters,
16627 DSA_generate_parameters is used.)
16631 * Include 'length' (recommended exponent length) in C code generated
16632 by 'openssl dhparam -C'.
16636 * The second argument to set_label in perlasm was already being used
16637 so couldn't be used as a "file scope" flag. Moved to third argument
16642 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16643 instead of RAND_bytes for encryption IVs and salts.
16647 * Include RAND_status() into RAND_METHOD instead of implementing
16648 it only for md_rand.c Otherwise replacing the PRNG by calling
16649 RAND_set_rand_method would be impossible.
16653 * Don't let DSA_generate_key() enter an infinite loop if the random
16654 number generation fails.
16658 * New 'rand' application for creating pseudo-random output.
16662 * Added configuration support for Linux/IA64
16664 *Rolf Haberrecker <rolf@suse.de>*
16666 * Assembler module support for Mingw32.
16670 * Shared library support for HPUX (in shlib/).
16672 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16674 * Shared library support for Solaris gcc.
16676 *Lutz Behnke <behnke@trustcenter.de>*
16678 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16680 * PKCS7_encrypt() was adding text MIME headers twice because they
16681 were added manually and by SMIME_crlf_copy().
16685 * In bntest.c don't call BN_rand with zero bits argument.
16687 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16689 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16690 case was implemented. This caused BN_div_recp() to fail occasionally.
16694 * Add an optional second argument to the set_label() in the perl
16695 assembly language builder. If this argument exists and is set
16696 to 1 it signals that the assembler should use a symbol whose
16697 scope is the entire file, not just the current function. This
16698 is needed with MASM which uses the format label:: for this scope.
16700 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16702 * Change the ASN1 types so they are typedefs by default. Before
16703 almost all types were #define'd to ASN1_STRING which was causing
16704 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16709 * Change names of new functions to the new get1/get0 naming
16710 convention: After 'get1', the caller owns a reference count
16711 and has to call `..._free`; 'get0' returns a pointer to some
16712 data structure without incrementing reference counters.
16713 (Some of the existing 'get' functions increment a reference
16714 counter, some don't.)
16715 Similarly, 'set1' and 'add1' functions increase reference
16716 counters or duplicate objects.
16720 * Allow for the possibility of temp RSA key generation failure:
16721 the code used to assume it always worked and crashed on failure.
16725 * Fix potential buffer overrun problem in BIO_printf().
16726 *Ulf Möller, using public domain code by Patrick Powell; problem
16727 pointed out by David Sacerdote <das33@cornell.edu>*
16729 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16730 RAND_egd() and RAND_status(). In the command line application,
16731 the EGD socket can be specified like a seed file using RANDFILE
16736 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16737 Some CAs (e.g. Verisign) distribute certificates in this form.
16741 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16742 list to exclude them. This means that no special compilation option
16743 is needed to use anonymous DH: it just needs to be included in the
16748 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16749 EVP_MD_type. The old functionality is available in a new macro called
16750 EVP_MD_md(). Change code that uses it and update docs.
16754 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16755 where the `void *` argument is replaced by a function pointer argument.
16756 Previously `void *` was abused to point to functions, which works on
16757 many platforms, but is not correct. As these functions are usually
16758 called by macros defined in OpenSSL header files, most source code
16759 should work without changes.
16763 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16764 sections with information on -D... compiler switches used for
16765 compiling the library so that applications can see them. To enable
16766 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16767 must be defined. E.g.,
16768 #define OPENSSL_ALGORITHM_DEFINES
16769 #include <openssl/opensslconf.h>
16770 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16772 *Richard Levitte, Ulf and Bodo Möller*
16774 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16779 * Change the 'other' type in certificate aux info to a STACK_OF
16780 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16781 the required ASN1 format: arbitrary types determined by an OID.
16785 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16786 argument to 'req'. This is not because the function is newer or
16787 better than others it just uses the work 'NEW' in the certificate
16788 request header lines. Some software needs this.
16792 * Reorganise password command line arguments: now passwords can be
16793 obtained from various sources. Delete the PEM_cb function and make
16794 it the default behaviour: i.e. if the callback is NULL and the
16795 usrdata argument is not NULL interpret it as a null terminated pass
16796 phrase. If usrdata and the callback are NULL then the pass phrase
16797 is prompted for as usual.
16801 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16802 the support is automatically enabled. The resulting binaries will
16803 autodetect the card and use it if present.
16805 *Ben Laurie and Compaq Inc.*
16807 * Work around for Netscape hang bug. This sends certificate request
16808 and server done in one record. Since this is perfectly legal in the
16809 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16810 the bugs/SSLv3 entry for more info.
16814 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16818 * Add -rand argument to smime and pkcs12 applications and read/write
16823 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16827 * Add command line password options to the remaining applications.
16831 * Bug fix for BN_div_recp() for numerators with an even number of
16836 * More tests in bntest.c, and changed test_bn output.
16840 * ./config recognizes MacOS X now.
16844 * Bug fix for BN_div() when the first words of num and divisor are
16845 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16849 * Add support for various broken PKCS#8 formats, and command line
16850 options to produce them.
16854 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16855 get temporary BIGNUMs from a BN_CTX.
16859 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16864 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16865 include a #define from the old name to the new. The original intent
16866 was that statically linked binaries could for example just call
16867 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16868 link with digests. This never worked because SSLeay_add_all_digests()
16869 and SSLeay_add_all_ciphers() were in the same source file so calling
16870 one would link with the other. They are now in separate source files.
16874 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16878 * Use a less unusual form of the Miller-Rabin primality test (it used
16879 a binary algorithm for exponentiation integrated into the Miller-Rabin
16880 loop, our standard modexp algorithms are faster).
16884 * Support for the EBCDIC character set completed.
16886 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16888 * Source code cleanups: use const where appropriate, eliminate casts,
16889 use `void *` instead of `char *` in lhash.
16893 * Bugfix: ssl3_send_server_key_exchange was not restartable
16894 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16895 this the server could overwrite ephemeral keys that the client
16900 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16901 using 50 iterations of the Rabin-Miller test.
16903 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16904 iterations of the Rabin-Miller test as required by the appendix
16905 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16906 As BN_is_prime_fasttest includes trial division, DSA parameter
16907 generation becomes much faster.
16909 This implies a change for the callback functions in DSA_is_prime
16910 and DSA_generate_parameters: The callback function is called once
16911 for each positive witness in the Rabin-Miller test, not just
16912 occasionally in the inner loop; and the parameters to the
16913 callback function now provide an iteration count for the outer
16914 loop rather than for the current invocation of the inner loop.
16915 DSA_generate_parameters additionally can call the callback
16916 function with an 'iteration count' of -1, meaning that a
16917 candidate has passed the trial division test (when q is generated
16918 from an application-provided seed, trial division is skipped).
16922 * New function BN_is_prime_fasttest that optionally does trial
16923 division before starting the Rabin-Miller test and has
16924 an additional BN_CTX * argument (whereas BN_is_prime always
16925 has to allocate at least one BN_CTX).
16926 'callback(1, -1, cb_arg)' is called when a number has passed the
16927 trial division stage.
16931 * Fix for bug in CRL encoding. The validity dates weren't being handled
16936 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16940 * New function BN_pseudo_rand().
16944 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16945 bignum version of BN_from_montgomery() with the working code from
16946 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16951 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16952 made it impossible to use the same SSL_SESSION data structure in
16953 SSL2 clients in multiple threads.
16957 * The return value of RAND_load_file() no longer counts bytes obtained
16958 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16959 to seed the PRNG (previously an explicit byte count was required).
16961 *Ulf Möller, Bodo Möller*
16963 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16964 used `char *` instead of `void *` and had casts all over the place.
16968 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16972 * Retain source code compatibility for BN_prime_checks macro:
16973 BN_is_prime(..., BN_prime_checks, ...) now uses
16974 BN_prime_checks_for_size to determine the appropriate number of
16975 Rabin-Miller iterations.
16979 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16980 DH_CHECK_P_NOT_SAFE_PRIME.
16981 (Check if this is true? OpenPGP calls them "strong".)
16985 * Merge the functionality of "dh" and "gendh" programs into a new program
16986 "dhparam". The old programs are retained for now but will handle DH keys
16987 (instead of parameters) in future.
16991 * Make the ciphers, s_server and s_client programs check the return values
16992 when a new cipher list is set.
16996 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16997 ciphers. Before when the 56bit ciphers were enabled the sorting was
17000 The syntax for the cipher sorting has been extended to support sorting by
17001 cipher-strength (using the strength_bits hard coded in the tables).
17002 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17004 Fix a bug in the cipher-command parser: when supplying a cipher command
17005 string with an "undefined" symbol (neither command nor alphanumeric
17006 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17007 an error is flagged.
17009 Due to the strength-sorting extension, the code of the
17010 ssl_create_cipher_list() function was completely rearranged. I hope that
17011 the readability was also increased :-)
17013 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17015 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17016 for the first serial number and places 2 in the serial number file. This
17017 avoids problems when the root CA is created with serial number zero and
17018 the first user certificate has the same issuer name and serial number
17023 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17024 the new code. Add documentation for this stuff.
17028 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17029 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17030 structures and behave in an analogous way to the X509v3 functions:
17031 they shouldn't be called directly but wrapper functions should be used
17034 So we also now have some wrapper functions that call the X509at functions
17035 when passed certificate requests. (TO DO: similar things can be done with
17036 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17037 things. Some of these need some d2i or i2d and print functionality
17038 because they handle more complex structures.)
17042 * Add missing #ifndefs that caused missing symbols when building libssl
17043 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17044 NO_RSA in `ssl/s2*.c`.
17046 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17048 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17049 has a return value which indicates the quality of the random data
17050 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17051 error queue. New function RAND_pseudo_bytes() generates output that is
17052 guaranteed to be unique but not unpredictable. RAND_add is like
17053 RAND_seed, but takes an extra argument for an entropy estimate
17054 (RAND_seed always assumes full entropy).
17058 * Do more iterations of Rabin-Miller probable prime test (specifically,
17059 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17060 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17061 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17062 false-positive rate of at most 2^-80 for random input.
17066 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17070 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17071 in the 0.9.5 release), this returns the chain
17072 from an X509_CTX structure with a dup of the stack and all
17073 the X509 reference counts upped: so the stack will exist
17074 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17077 Also make SSL_SESSION_print() print out the verify return
17082 * Add manpage for the pkcs12 command. Also change the default
17083 behaviour so MAC iteration counts are used unless the new
17084 -nomaciter option is used. This improves file security and
17085 only older versions of MSIE (4.0 for example) need it.
17089 * Honor the no-xxx Configure options when creating .DEF files.
17093 * Add PKCS#10 attributes to field table: challengePassword,
17094 unstructuredName and unstructuredAddress. These are taken from
17095 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17096 international characters are used.
17098 More changes to X509_ATTRIBUTE code: allow the setting of types
17099 based on strings. Remove the 'loc' parameter when adding
17100 attributes because these will be a SET OF encoding which is sorted
17105 * Initial changes to the 'req' utility to allow request generation
17106 automation. This will allow an application to just generate a template
17107 file containing all the field values and have req construct the
17110 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17111 used all over the place including certificate requests and PKCS#7
17112 structures. They are currently handled manually where necessary with
17113 some primitive wrappers for PKCS#7. The new functions behave in a
17114 manner analogous to the X509 extension functions: they allow
17115 attributes to be looked up by NID and added.
17117 Later something similar to the X509V3 code would be desirable to
17118 automatically handle the encoding, decoding and printing of the
17119 more complex types. The string types like challengePassword can
17120 be handled by the string table functions.
17122 Also modified the multi byte string table handling. Now there is
17123 a 'global mask' which masks out certain types. The table itself
17124 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17125 is useful when for example there is only one permissible type
17126 (as in countryName) and using the mask might result in no valid
17131 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17132 SSL_get_peer_finished to allow applications to obtain the latest
17133 Finished messages sent to the peer or expected from the peer,
17134 respectively. (SSL_get_peer_finished is usually the Finished message
17135 actually received from the peer, otherwise the protocol will be aborted.)
17137 As the Finished message are message digests of the complete handshake
17138 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17139 be used for external authentication procedures when the authentication
17140 provided by SSL/TLS is not desired or is not enough.
17144 * Enhanced support for Alpha Linux is added. Now ./config checks if
17145 the host supports BWX extension and if Compaq C is present on the
17146 $PATH. Just exploiting of the BWX extension results in 20-30%
17147 performance kick for some algorithms, e.g. DES and RC4 to mention
17148 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17153 * Add support for MS "fast SGC". This is arguably a violation of the
17154 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17155 weak crypto and after checking the certificate is SGC a second one
17156 with strong crypto. MS SGC stops the first handshake after receiving
17157 the server certificate message and sends a second client hello. Since
17158 a server will typically do all the time consuming operations before
17159 expecting any further messages from the client (server key exchange
17160 is the most expensive) there is little difference between the two.
17162 To get OpenSSL to support MS SGC we have to permit a second client
17163 hello message after we have sent server done. In addition we have to
17164 reset the MAC if we do get this second client hello.
17168 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17169 if a DER encoded private key is RSA or DSA traditional format. Changed
17170 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17171 format DER encoded private key. Newer code should use PKCS#8 format which
17172 has the key type encoded in the ASN1 structure. Added DER private key
17173 support to pkcs8 application.
17177 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17178 ciphersuites has been selected (as required by the SSL 3/TLS 1
17179 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17180 is set, we interpret this as a request to violate the specification
17181 (the worst that can happen is a handshake failure, and 'correct'
17182 behaviour would result in a handshake failure anyway).
17186 * In SSL_CTX_add_session, take into account that there might be multiple
17187 SSL_SESSION structures with the same session ID (e.g. when two threads
17188 concurrently obtain them from an external cache).
17189 The internal cache can handle only one SSL_SESSION with a given ID,
17190 so if there's a conflict, we now throw out the old one to achieve
17195 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17196 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17197 some routines that use cipher OIDs: some ciphers do not have OIDs
17198 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17203 * Simplify the trust setting structure and code. Now we just have
17204 two sequences of OIDs for trusted and rejected settings. These will
17205 typically have values the same as the extended key usage extension
17206 and any application specific purposes.
17208 The trust checking code now has a default behaviour: it will just
17209 check for an object with the same NID as the passed id. Functions can
17210 be provided to override either the default behaviour or the behaviour
17211 for a given id. SSL client, server and email already have functions
17212 in place for compatibility: they check the NID and also return "trusted"
17213 if the certificate is self signed.
17217 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17218 traditional format into an EVP_PKEY structure.
17222 * Add a password callback function PEM_cb() which either prompts for
17223 a password if usr_data is NULL or otherwise assumes it is a null
17224 terminated password. Allow passwords to be passed on command line
17225 environment or config files in a few more utilities.
17229 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17230 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17231 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17232 Update documentation.
17236 * Support for ASN1 "NULL" type. This could be handled before by using
17237 ASN1_TYPE but there wasn't any function that would try to read a NULL
17238 and produce an error if it couldn't. For compatibility we also have
17239 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17240 don't allocate anything because they don't need to.
17244 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17247 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17249 * Rebuild of the memory allocation routines used by OpenSSL code and
17250 possibly others as well. The purpose is to make an interface that
17251 provide hooks so anyone can build a separate set of allocation and
17252 deallocation routines to be used by OpenSSL, for example memory
17253 pool implementations, or something else, which was previously hard
17254 since Malloc(), Realloc() and Free() were defined as macros having
17255 the values malloc, realloc and free, respectively (except for Win32
17256 compilations). The same is provided for memory debugging code.
17257 OpenSSL already comes with functionality to find memory leaks, but
17258 this gives people a chance to debug other memory problems.
17260 With these changes, a new set of functions and macros have appeared:
17262 CRYPTO_set_mem_debug_functions() [F]
17263 CRYPTO_get_mem_debug_functions() [F]
17264 CRYPTO_dbg_set_options() [F]
17265 CRYPTO_dbg_get_options() [F]
17266 CRYPTO_malloc_debug_init() [M]
17268 The memory debug functions are NULL by default, unless the library
17269 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17270 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17271 gives the standard debugging functions that come with OpenSSL) or
17272 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17273 provided by the library user) must be used. When the standard
17274 debugging functions are used, CRYPTO_dbg_set_options can be used to
17275 request additional information:
17276 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17277 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17279 Also, things like CRYPTO_set_mem_functions will always give the
17280 expected result (the new set of functions is used for allocation
17281 and deallocation) at all times, regardless of platform and compiler
17284 To finish it up, some functions that were never use in any other
17285 way than through macros have a new API and new semantic:
17287 CRYPTO_dbg_malloc()
17288 CRYPTO_dbg_realloc()
17291 All macros of value have retained their old syntax.
17293 *Richard Levitte and Bodo Moeller*
17295 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17296 ordering of SMIMECapabilities wasn't in "strength order" and there
17297 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17302 * Some ASN1 types with illegal zero length encoding (INTEGER,
17303 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17305 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17307 * Merge in my S/MIME library for OpenSSL. This provides a simple
17308 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17309 functionality to handle multipart/signed properly) and a utility
17310 called 'smime' to call all this stuff. This is based on code I
17311 originally wrote for Celo who have kindly allowed it to be
17312 included in OpenSSL.
17316 * Add variants des_set_key_checked and des_set_key_unchecked of
17317 des_set_key (aka des_key_sched). Global variable des_check_key
17318 decides which of these is called by des_set_key; this way
17319 des_check_key behaves as it always did, but applications and
17320 the library itself, which was buggy for des_check_key == 1,
17321 have a cleaner way to pick the version they need.
17325 * New function PKCS12_newpass() which changes the password of a
17330 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17331 dynamic mix. In both cases the ids can be used as an index into the
17332 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17333 functions so they accept a list of the field values and the
17334 application doesn't need to directly manipulate the X509_TRUST
17339 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17344 * Modify the way the V3 extension code looks up extensions. This now
17345 works in a similar way to the object code: we have some "standard"
17346 extensions in a static table which is searched with OBJ_bsearch()
17347 and the application can add dynamic ones if needed. The file
17348 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17349 updated whenever a new extension is added to the core code and kept
17350 in ext_nid order. There is a simple program 'tabtest.c' which checks
17351 this. New extensions are not added too often so this file can readily
17352 be maintained manually.
17354 There are two big advantages in doing things this way. The extensions
17355 can be looked up immediately and no longer need to be "added" using
17356 X509V3_add_standard_extensions(): this function now does nothing.
17357 Side note: I get *lots* of email saying the extension code doesn't
17358 work because people forget to call this function.
17359 Also no dynamic allocation is done unless new extensions are added:
17360 so if we don't add custom extensions there is no need to call
17361 X509V3_EXT_cleanup().
17365 * Modify enc utility's salting as follows: make salting the default. Add a
17366 magic header, so unsalted files fail gracefully instead of just decrypting
17367 to garbage. This is because not salting is a big security hole, so people
17368 should be discouraged from doing it.
17372 * Fixes and enhancements to the 'x509' utility. It allowed a message
17373 digest to be passed on the command line but it only used this
17374 parameter when signing a certificate. Modified so all relevant
17375 operations are affected by the digest parameter including the
17376 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17377 DSA key was used because it didn't fix the digest.
17381 * Initial certificate chain verify code. Currently tests the untrusted
17382 certificates for consistency with the verify purpose (which is set
17383 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17385 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17386 this is because it will reject chains with invalid extensions whereas
17387 every previous version of OpenSSL and SSLeay made no checks at all.
17389 Trust code: checks the root CA for the relevant trust settings. Trust
17390 settings have an initial value consistent with the verify purpose: e.g.
17391 if the verify purpose is for SSL client use it expects the CA to be
17392 trusted for SSL client use. However the default value can be changed to
17393 permit custom trust settings: one example of this would be to only trust
17394 certificates from a specific "secure" set of CAs.
17396 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17397 which should be used for version portability: especially since the
17398 verify structure is likely to change more often now.
17400 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17401 to set them. If not set then assume SSL clients will verify SSL servers
17404 Two new options to the verify program: -untrusted allows a set of
17405 untrusted certificates to be passed in and -purpose which sets the
17406 intended purpose of the certificate. If a purpose is set then the
17407 new chain verify code is used to check extension consistency.
17411 * Support for the authority information access extension.
17415 * Modify RSA and DSA PEM read routines to transparently handle
17416 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17417 public keys in a format compatible with certificate
17418 SubjectPublicKeyInfo structures. Unfortunately there were already
17419 functions called *_PublicKey_* which used various odd formats so
17420 these are retained for compatibility: however the DSA variants were
17421 never in a public release so they have been deleted. Changed dsa/rsa
17422 utilities to handle the new format: note no releases ever handled public
17423 keys so we should be OK.
17425 The primary motivation for this change is to avoid the same fiasco
17426 that dogs private keys: there are several incompatible private key
17427 formats some of which are standard and some OpenSSL specific and
17428 require various evil hacks to allow partial transparent handling and
17429 even then it doesn't work with DER formats. Given the option anything
17430 other than PKCS#8 should be dumped: but the other formats have to
17431 stay in the name of compatibility.
17433 With public keys and the benefit of hindsight one standard format
17434 is used which works with EVP_PKEY, RSA or DSA structures: though
17435 it clearly returns an error if you try to read the wrong kind of key.
17437 Added a -pubkey option to the 'x509' utility to output the public key.
17438 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17439 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17440 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17441 that do the same as the `EVP_PKEY_assign_*()` except they up the
17442 reference count of the added key (they don't "swallow" the
17447 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17448 CRLs would fail if the file contained no certificates or no CRLs:
17449 added a new function to read in both types and return the number
17450 read: this means that if none are read it will be an error. The
17451 DER versions of the certificate and CRL reader would always fail
17452 because it isn't possible to mix certificates and CRLs in DER format
17453 without choking one or the other routine. Changed this to just read
17454 a certificate: this is the best we can do. Also modified the code
17455 in `apps/verify.c` to take notice of return codes: it was previously
17456 attempting to read in certificates from NULL pointers and ignoring
17457 any errors: this is one reason why the cert and CRL reader seemed
17458 to work. It doesn't check return codes from the default certificate
17459 routines: these may well fail if the certificates aren't installed.
17463 * Code to support otherName option in GeneralName.
17467 * First update to verify code. Change the verify utility
17468 so it warns if it is passed a self signed certificate:
17469 for consistency with the normal behaviour. X509_verify
17470 has been modified to it will now verify a self signed
17471 certificate if *exactly* the same certificate appears
17472 in the store: it was previously impossible to trust a
17473 single self signed certificate. This means that:
17474 openssl verify ss.pem
17475 now gives a warning about a self signed certificate but
17476 openssl verify -CAfile ss.pem ss.pem
17481 * For servers, store verify_result in SSL_SESSION data structure
17482 (and add it to external session representation).
17483 This is needed when client certificate verifications fails,
17484 but an application-provided verification callback (set by
17485 SSL_CTX_set_cert_verify_callback) allows accepting the session
17486 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17487 but returns 1): When the session is reused, we have to set
17488 ssl->verify_result to the appropriate error code to avoid
17491 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17493 * Fix a bug in the new PKCS#7 code: it didn't consider the
17494 case in PKCS7_dataInit() where the signed PKCS7 structure
17495 didn't contain any existing data because it was being created.
17497 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17499 * Add a salt to the key derivation routines in enc.c. This
17500 forms the first 8 bytes of the encrypted file. Also add a
17501 -S option to allow a salt to be input on the command line.
17505 * New function X509_cmp(). Oddly enough there wasn't a function
17506 to compare two certificates. We do this by working out the SHA1
17507 hash and comparing that. X509_cmp() will be needed by the trust
17512 * SSL_get1_session() is like SSL_get_session(), but increments
17513 the reference count in the SSL_SESSION returned.
17515 *Geoff Thorpe <geoff@eu.c2.net>*
17517 * Fix for 'req': it was adding a null to request attributes.
17518 Also change the X509_LOOKUP and X509_INFO code to handle
17519 certificate auxiliary information.
17523 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17528 * Add the possibility to add extra information to the memory leak
17529 detecting output, to form tracebacks, showing from where each
17530 allocation was originated: CRYPTO_push_info("constant string") adds
17531 the string plus current file name and line number to a per-thread
17532 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17533 is like calling CYRPTO_pop_info() until the stack is empty.
17534 Also updated memory leak detection code to be multi-thread-safe.
17538 * Add options -text and -noout to pkcs7 utility and delete the
17539 encryption options which never did anything. Update docs.
17543 * Add options to some of the utilities to allow the pass phrase
17544 to be included on either the command line (not recommended on
17545 OSes like Unix) or read from the environment. Update the
17546 manpages and fix a few bugs.
17550 * Add a few manpages for some of the openssl commands.
17554 * Fix the -revoke option in ca. It was freeing up memory twice,
17555 leaking and not finding already revoked certificates.
17559 * Extensive changes to support certificate auxiliary information.
17560 This involves the use of X509_CERT_AUX structure and X509_AUX
17561 functions. An X509_AUX function such as PEM_read_X509_AUX()
17562 can still read in a certificate file in the usual way but it
17563 will also read in any additional "auxiliary information". By
17564 doing things this way a fair degree of compatibility can be
17565 retained: existing certificates can have this information added
17566 using the new 'x509' options.
17568 Current auxiliary information includes an "alias" and some trust
17569 settings. The trust settings will ultimately be used in enhanced
17570 certificate chain verification routines: currently a certificate
17571 can only be trusted if it is self signed and then it is trusted
17576 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17577 The problem was that one of the replacement routines had not been working
17578 since SSLeay releases. For now the offending routine has been replaced
17579 with non-optimised assembler. Even so, this now gives around 95%
17580 performance improvement for 1024 bit RSA signs.
17584 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17585 handling. Most clients have the effective key size in bits equal to
17586 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17587 A few however don't do this and instead use the size of the decrypted key
17588 to determine the RC2 key length and the AlgorithmIdentifier to determine
17589 the effective key length. In this case the effective key length can still
17590 be 40 bits but the key length can be 168 bits for example. This is fixed
17591 by manually forcing an RC2 key into the EVP_PKEY structure because the
17592 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17593 the key length and effective key length are equal.
17597 * Add a bunch of functions that should simplify the creation of
17598 X509_NAME structures. Now you should be able to do:
17599 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17600 and have it automatically work out the correct field type and fill in
17601 the structures. The more adventurous can try:
17602 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17603 and it will (hopefully) work out the correct multibyte encoding.
17607 * Change the 'req' utility to use the new field handling and multibyte
17608 copy routines. Before the DN field creation was handled in an ad hoc
17609 way in req, ca, and x509 which was rather broken and didn't support
17610 BMPStrings or UTF8Strings. Since some software doesn't implement
17611 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17612 using the dirstring_type option. See the new comment in the default
17613 openssl.cnf for more info.
17617 * Make crypto/rand/md_rand.c more robust:
17618 - Assure unique random numbers after fork().
17619 - Make sure that concurrent threads access the global counter and
17620 md serializably so that we never lose entropy in them
17621 or use exactly the same state in multiple threads.
17622 Access to the large state is not always serializable because
17623 the additional locking could be a performance killer, and
17624 md should be large enough anyway.
17628 * New file `apps/app_rand.c` with commonly needed functionality
17629 for handling the random seed file.
17631 Use the random seed file in some applications that previously did not:
17633 dsaparam -genkey (which also ignored its '-rand' option),
17636 x509 (when signing).
17637 Except on systems with /dev/urandom, it is crucial to have a random
17638 seed file at least for key creation, DSA signing, and for DH exchanges;
17639 for RSA signatures we could do without one.
17641 gendh and gendsa (unlike genrsa) used to read only the first byte
17642 of each file listed in the '-rand' option. The function as previously
17643 found in genrsa is now in app_rand.c and is used by all programs
17644 that support '-rand'.
17648 * In RAND_write_file, use mode 0600 for creating files;
17649 don't just chmod when it may be too late.
17653 * Report an error from X509_STORE_load_locations
17654 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17658 * New function ASN1_mbstring_copy() this copies a string in either
17659 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17660 into an ASN1_STRING type. A mask of permissible types is passed
17661 and it chooses the "minimal" type to use or an error if not type
17666 * Add function equivalents to the various macros in asn1.h. The old
17667 macros are retained with an `M_` prefix. Code inside the library can
17668 use the `M_` macros. External code (including the openssl utility)
17669 should *NOT* in order to be "shared library friendly".
17673 * Add various functions that can check a certificate's extensions
17674 to see if it usable for various purposes such as SSL client,
17675 server or S/MIME and CAs of these types. This is currently
17676 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17677 verification. Also added a -purpose flag to x509 utility to
17678 print out all the purposes.
17682 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17687 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17688 for, obtain and decode and extension and obtain its critical flag.
17689 This allows all the necessary extension code to be handled in a
17690 single function call.
17694 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17695 platforms. See crypto/rc4/rc4_enc.c for further details.
17699 * New -noout option to asn1parse. This causes no output to be produced
17700 its main use is when combined with -strparse and -out to extract data
17701 from a file (which may not be in ASN.1 format).
17705 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17706 when producing the local key id.
17708 *Richard Levitte <levitte@stacken.kth.se>*
17710 * New option -dhparam in s_server. This allows a DH parameter file to be
17711 stated explicitly. If it is not stated then it tries the first server
17712 certificate file. The previous behaviour hard coded the filename
17717 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17718 a public key to be input or output. For example:
17719 openssl rsa -in key.pem -pubout -out pubkey.pem
17720 Also added necessary DSA public key functions to handle this.
17724 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17725 in the message. This was handled by allowing
17726 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17728 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17730 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17731 to the end of the strings whereas this didn't. This would cause problems
17732 if strings read with d2i_ASN1_bytes() were later modified.
17734 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17736 * Fix for base64 decode bug. When a base64 bio reads only one line of
17737 data and it contains EOF it will end up returning an error. This is
17738 caused by input 46 bytes long. The cause is due to the way base64
17739 BIOs find the start of base64 encoded data. They do this by trying a
17740 trial decode on each line until they find one that works. When they
17741 do a flag is set and it starts again knowing it can pass all the
17742 data directly through the decoder. Unfortunately it doesn't reset
17743 the context it uses. This means that if EOF is reached an attempt
17744 is made to pass two EOFs through the context and this causes the
17745 resulting error. This can also cause other problems as well. As is
17746 usual with these problems it takes *ages* to find and the fix is
17747 trivial: move one line.
17749 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17751 * Ugly workaround to get s_client and s_server working under Windows. The
17752 old code wouldn't work because it needed to select() on sockets and the
17753 tty (for keypresses and to see if data could be written). Win32 only
17754 supports select() on sockets so we select() with a 1s timeout on the
17755 sockets and then see if any characters are waiting to be read, if none
17756 are present then we retry, we also assume we can always write data to
17757 the tty. This isn't nice because the code then blocks until we've
17758 received a complete line of data and it is effectively polling the
17759 keyboard at 1s intervals: however it's quite a bit better than not
17760 working at all :-) A dedicated Windows application might handle this
17761 with an event loop for example.
17765 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17766 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17767 will be called when RSA_sign() and RSA_verify() are used. This is useful
17768 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17769 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17770 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17771 This necessitated the support of an extra signature type NID_md5_sha1
17772 for SSL signatures and modifications to the SSL library to use it instead
17773 of calling RSA_public_decrypt() and RSA_private_encrypt().
17777 * Add new -verify -CAfile and -CApath options to the crl program, these
17778 will lookup a CRL issuers certificate and verify the signature in a
17779 similar way to the verify program. Tidy up the crl program so it
17780 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17781 less strict. It will now permit CRL extensions even if it is not
17782 a V2 CRL: this will allow it to tolerate some broken CRLs.
17786 * Initialize all non-automatic variables each time one of the openssl
17787 sub-programs is started (this is necessary as they may be started
17788 multiple times from the "OpenSSL>" prompt).
17790 *Lennart Bang, Bodo Moeller*
17792 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17793 removing all other RSA functionality (this is what NO_RSA does). This
17794 is so (for example) those in the US can disable those operations covered
17795 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17800 * Non-copying interface to BIO pairs.
17801 (still largely untested)
17805 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17806 ASCII string. This was handled independently in various places before.
17810 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17811 UTF8 strings a character at a time.
17815 * Use client_version from client hello to select the protocol
17816 (s23_srvr.c) and for RSA client key exchange verification
17817 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17821 * Add various utility functions to handle SPKACs, these were previously
17822 handled by poking round in the structure internals. Added new function
17823 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17824 print, verify and generate SPKACs. Based on an original idea from
17825 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17829 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17833 * Allow the config file extension section to be overwritten on the
17834 command line. Based on an original idea from Massimiliano Pala
17835 <madwolf@comune.modena.it>. The new option is called -extensions
17836 and can be applied to ca, req and x509. Also -reqexts to override
17837 the request extensions in req and -crlexts to override the crl extensions
17842 * Add new feature to the SPKAC handling in ca. Now you can include
17843 the same field multiple times by preceding it by "XXXX." for example:
17846 this is the same syntax as used in the req config file.
17850 * Allow certificate extensions to be added to certificate requests. These
17851 are specified in a 'req_extensions' option of the req section of the
17852 config file. They can be printed out with the -text option to req but
17853 are otherwise ignored at present.
17857 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17858 data read consists of only the final block it would not decrypted because
17859 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17860 A misplaced 'break' also meant the decrypted final block might not be
17861 copied until the next read.
17865 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17866 a few extra parameters to the DH structure: these will be useful if
17867 for example we want the value of 'q' or implement X9.42 DH.
17871 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17872 provides hooks that allow the default DSA functions or functions on a
17873 "per key" basis to be replaced. This allows hardware acceleration and
17874 hardware key storage to be handled without major modification to the
17875 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17876 associated functions.
17880 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17881 as "read only": it can't be written to and the buffer it points to will
17882 not be freed. Reading from a read only BIO is much more efficient than
17883 a normal memory BIO. This was added because there are several times when
17884 an area of memory needs to be read from a BIO. The previous method was
17885 to create a memory BIO and write the data to it, this results in two
17886 copies of the data and an O(n^2) reading algorithm. There is a new
17887 function BIO_new_mem_buf() which creates a read only memory BIO from
17888 an area of memory. Also modified the PKCS#7 routines to use read only
17893 * Bugfix: ssl23_get_client_hello did not work properly when called in
17894 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17895 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17896 but a retry condition occurred while trying to read the rest.
17900 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17901 NID_pkcs7_encrypted by default: this was wrong since this should almost
17902 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17903 the encrypted data type: this is a more sensible place to put it and it
17904 allows the PKCS#12 code to be tidied up that duplicated this
17909 * Changed obj_dat.pl script so it takes its input and output files on
17910 the command line. This should avoid shell escape redirection problems
17915 * Initial support for certificate extension requests, these are included
17916 in things like Xenroll certificate requests. Included functions to allow
17917 extensions to be obtained and added.
17921 * -crlf option to s_client and s_server for sending newlines as
17922 CRLF (as required by many protocols).
17926 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17928 * Install libRSAglue.a when OpenSSL is built with RSAref.
17930 *Ralf S. Engelschall*
17932 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17934 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17936 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17941 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17942 DH parameters/keys (q is lost during that conversion, but the resulting
17943 DH parameters contain its length).
17945 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17946 much faster than DH_generate_parameters (which creates parameters
17947 where `p = 2*q + 1`), and also the smaller q makes DH computations
17948 much more efficient (160-bit exponentiation instead of 1024-bit
17949 exponentiation); so this provides a convenient way to support DHE
17950 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17951 utter importance to use
17952 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17954 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17955 when such DH parameters are used, because otherwise small subgroup
17956 attacks may become possible!
17960 * Avoid memory leak in i2d_DHparams.
17964 * Allow the -k option to be used more than once in the enc program:
17965 this allows the same encrypted message to be read by multiple recipients.
17969 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17970 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17971 it will always use the numerical form of the OID, even if it has a short
17976 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17977 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17978 otherwise bn_mod_exp was called. In the case of hardware keys for example
17979 no private key components need be present and it might store extra data
17980 in the RSA structure, which cannot be accessed from bn_mod_exp.
17981 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17982 private key operations.
17986 * Added support for SPARC Linux.
17990 * pem_password_cb function type incompatibly changed from
17991 typedef int pem_password_cb(char *buf, int size, int rwflag);
17993 ....(char *buf, int size, int rwflag, void *userdata);
17994 so that applications can pass data to their callbacks:
17995 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17996 additional void * argument, which is just handed through whenever
17997 the password callback is called.
17999 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18001 New function SSL_CTX_set_default_passwd_cb_userdata.
18003 Compatibility note: As many C implementations push function arguments
18004 onto the stack in reverse order, the new library version is likely to
18005 interoperate with programs that have been compiled with the old
18006 pem_password_cb definition (PEM_whatever takes some data that
18007 happens to be on the stack as its last argument, and the callback
18008 just ignores this garbage); but there is no guarantee whatsoever that
18011 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18012 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18013 problems not only on Windows, but also on some Unix platforms.
18014 To avoid problematic command lines, these definitions are now in an
18015 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18016 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18020 * MIPS III/IV assembler module is reimplemented.
18024 * More DES library cleanups: remove references to srand/rand and
18025 delete an unused file.
18029 * Add support for the free Netwide assembler (NASM) under Win32,
18030 since not many people have MASM (ml) and it can be hard to obtain.
18031 This is currently experimental but it seems to work OK and pass all
18032 the tests. Check out INSTALL.W32 for info.
18036 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18037 without temporary keys kept an extra copy of the server key,
18038 and connections with temporary keys did not free everything in case
18043 * New function RSA_check_key and new openssl rsa option -check
18044 for verifying the consistency of RSA keys.
18046 *Ulf Moeller, Bodo Moeller*
18048 * Various changes to make Win32 compile work:
18049 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18050 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18051 comparison" warnings.
18052 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18056 * Add a debugging option to PKCS#5 v2 key generation function: when
18057 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18058 derived keys are printed to stderr.
18062 * Copy the flags in ASN1_STRING_dup().
18064 *Roman E. Pavlov <pre@mo.msk.ru>*
18066 * The x509 application mishandled signing requests containing DSA
18067 keys when the signing key was also DSA and the parameters didn't match.
18069 It was supposed to omit the parameters when they matched the signing key:
18070 the verifying software was then supposed to automatically use the CA's
18071 parameters if they were absent from the end user certificate.
18073 Omitting parameters is no longer recommended. The test was also
18074 the wrong way round! This was probably due to unusual behaviour in
18075 EVP_cmp_parameters() which returns 1 if the parameters match.
18076 This meant that parameters were omitted when they *didn't* match and
18077 the certificate was useless. Certificates signed with 'ca' didn't have
18080 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18082 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18083 The interface is as follows:
18084 Applications can use
18085 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18086 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18087 "off" is now the default.
18088 The library internally uses
18089 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18090 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18091 to disable memory-checking temporarily.
18093 Some inconsistent states that previously were possible (and were
18094 even the default) are now avoided.
18096 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18097 with each memory chunk allocated; this is occasionally more helpful
18098 than just having a counter.
18100 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18102 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18107 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18108 which largely parallels "options", but is for changing API behaviour,
18109 whereas "options" are about protocol behaviour.
18110 Initial "mode" flags are:
18112 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18113 a single record has been written.
18114 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18115 retries use the same buffer location.
18116 (But all of the contents must be
18121 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18124 * Fix problems with no-hmac etc.
18126 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18128 * New functions RSA_get_default_method(), RSA_set_method() and
18129 RSA_get_method(). These allows replacement of RSA_METHODs without having
18130 to mess around with the internals of an RSA structure.
18134 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18135 Also really enable memory leak checks in openssl.c and in some
18138 *Chad C. Mulligan, Bodo Moeller*
18140 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18141 up the length of negative integers. This has now been simplified to just
18142 store the length when it is first determined and use it later, rather
18143 than trying to keep track of where data is copied and updating it to
18145 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18147 * Add a new function PKCS7_signatureVerify. This allows the verification
18148 of a PKCS#7 signature but with the signing certificate passed to the
18149 function itself. This contrasts with PKCS7_dataVerify which assumes the
18150 certificate is present in the PKCS#7 structure. This isn't always the
18151 case: certificates can be omitted from a PKCS#7 structure and be
18152 distributed by "out of band" means (such as a certificate database).
18156 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18157 function prototypes in pem.h, also change util/mkdef.pl to add the
18158 necessary function names.
18162 * mk1mf.pl (used by Windows builds) did not properly read the
18163 options set by Configure in the top level Makefile, and Configure
18164 was not even able to write more than one option correctly.
18165 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18169 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18170 file to be loaded from a BIO or FILE pointer. The BIO version will
18171 for example allow memory BIOs to contain config info.
18175 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18176 Whoever hopes to achieve shared-library compatibility across versions
18177 must use this, not the compile-time macro.
18178 (Exercise 0.9.4: Which is the minimum library version required by
18180 Note: All this applies only to multi-threaded programs, others don't
18185 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18186 through a BIO pair triggered the default case, i.e.
18187 SSLerr(...,SSL_R_UNKNOWN_STATE).
18191 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18192 can use the SSL library even if none of the specific BIOs is
18197 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18198 for the encoded length.
18200 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18202 * Add initial documentation of the X509V3 functions.
18206 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18207 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18208 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18209 secure PKCS#8 private key format with a high iteration count.
18213 * Fix determination of Perl interpreter: A perl or perl5
18214 *directory* in $PATH was also accepted as the interpreter.
18216 *Ralf S. Engelschall*
18218 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18219 wrong with it but it was very old and did things like calling
18220 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18221 unusual formatting.
18225 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18226 to use the new extension code.
18230 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18231 with macros. This should make it easier to change their form, add extra
18232 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18237 * Add to configuration table a new entry that can specify an alternative
18238 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18239 according to Mark Crispin <MRC@Panda.COM>.
18243 * DES CBC did not update the IV. Weird.
18247 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18248 Changing the behaviour of the former might break existing programs --
18249 where IV updating is needed, des_ncbc_encrypt can be used.
18252 * When bntest is run from "make test" it drives bc to check its
18253 calculations, as well as internally checking them. If an internal check
18254 fails, it needs to cause bc to give a non-zero result or make test carries
18255 on without noticing the failure. Fixed.
18259 * DES library cleanups.
18263 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18264 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18265 ciphers. NOTE: although the key derivation function has been verified
18266 against some published test vectors it has not been extensively tested
18267 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18272 * Instead of "mkdir -p", which is not fully portable, use new
18273 Perl script "util/mkdir-p.pl".
18277 * Rewrite the way password based encryption (PBE) is handled. It used to
18278 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18279 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18280 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18281 the 'parameter' field of the AlgorithmIdentifier is passed to the
18282 underlying key generation function so it must do its own ASN1 parsing.
18283 This has also changed the EVP_PBE_CipherInit() function which now has a
18284 'parameter' argument instead of literal salt and iteration count values
18285 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18289 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18290 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18291 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18292 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18293 value was just used as a "magic string" and not used directly its
18294 value doesn't matter.
18298 * Introduce some semblance of const correctness to BN. Shame C doesn't
18303 * "linux-sparc64" configuration (ultrapenguin).
18305 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18306 "linux-sparc" configuration.
18308 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18310 * config now generates no-xxx options for missing ciphers.
18314 * Support the EBCDIC character set (work in progress).
18315 File ebcdic.c not yet included because it has a different license.
18317 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18319 * Support BS2000/OSD-POSIX.
18321 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18323 * Make callbacks for key generation use `void *` instead of `char *`.
18327 * Make S/MIME samples compile (not yet tested).
18331 * Additional typesafe stacks.
18335 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18339 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18341 * New configuration variant "sco5-gcc".
18343 * Updated some demos.
18345 *Sean O Riordain, Wade Scholine*
18347 * Add missing BIO_free at exit of pkcs12 application.
18351 * Fix memory leak in conf.c.
18355 * Updates for Win32 to assembler version of MD5.
18359 * Set #! path to perl in `apps/der_chop` to where we found it
18360 instead of using a fixed path.
18364 * SHA library changes for irix64-mips4-cc.
18368 * Improvements for VMS support.
18372 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18374 * Bignum library bug fix. IRIX 6 passes "make test" now!
18375 This also avoids the problems with SC4.2 and unpatched SC5.
18377 *Andy Polyakov <appro@fy.chalmers.se>*
18379 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18380 These are required because of the typesafe stack would otherwise break
18381 existing code. If old code used a structure member which used to be STACK
18382 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18383 sk_num or sk_value it would produce an error because the num, data members
18384 are not present in STACK_OF. Now it just produces a warning. sk_set
18385 replaces the old method of assigning a value to sk_value
18386 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18387 that does this will no longer work (and should use sk_set instead) but
18388 this could be regarded as a "questionable" behaviour anyway.
18392 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18393 correctly handle encrypted S/MIME data.
18397 * Change type of various DES function arguments from des_cblock
18398 (which means, in function argument declarations, pointer to char)
18399 to des_cblock * (meaning pointer to array with 8 char elements),
18400 which allows the compiler to do more typechecking; it was like
18401 that back in SSLeay, but with lots of ugly casts.
18403 Introduce new type const_des_cblock.
18407 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18408 problems: find RecipientInfo structure that matches recipient certificate
18409 and initialise the ASN1 structures properly based on passed cipher.
18413 * Belatedly make the BN tests actually check the results.
18417 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18418 to and from BNs: it was completely broken. New compilation option
18419 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18420 key elements as negative integers.
18424 * Reorganize and speed up MD5.
18426 *Andy Polyakov <appro@fy.chalmers.se>*
18430 *Richard Levitte <richard@levitte.org>*
18432 * New option -out to asn1parse to allow the parsed structure to be
18433 output to a file. This is most useful when combined with the -strparse
18434 option to examine the output of things like OCTET STRINGS.
18438 * Make SSL library a little more fool-proof by not requiring any longer
18439 that `SSL_set_{accept,connect}_state` be called before
18440 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18441 in many applications because usually everything *appeared* to work as
18442 intended anyway -- now it really works as intended).
18446 * Move openssl.cnf out of lib/.
18450 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18451 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18452 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18454 *Ralf S. Engelschall*
18456 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18457 handle PKCS#7 enveloped data properly.
18459 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18461 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18462 copying pointers. The cert_st handling is changed by this in
18463 various ways (and thus what used to be known as ctx->default_cert
18464 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18465 any longer when s->cert does not give us what we need).
18466 ssl_cert_instantiate becomes obsolete by this change.
18467 As soon as we've got the new code right (possibly it already is?),
18468 we have solved a couple of bugs of the earlier code where s->cert
18469 was used as if it could not have been shared with other SSL structures.
18471 Note that using the SSL API in certain dirty ways now will result
18472 in different behaviour than observed with earlier library versions:
18473 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18474 does not influence s as it used to.
18476 In order to clean up things more thoroughly, inside SSL_SESSION
18477 we don't use CERT any longer, but a new structure SESS_CERT
18478 that holds per-session data (if available); currently, this is
18479 the peer's certificate chain and, for clients, the server's certificate
18480 and temporary key. CERT holds only those values that can have
18481 meaningful defaults in an SSL_CTX.
18485 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18486 from the internal representation. Various PKCS#7 fixes: remove some
18487 evil casts and set the enc_dig_alg field properly based on the signing
18492 * Allow PKCS#12 password to be set from the command line or the
18493 environment. Let 'ca' get its config file name from the environment
18494 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18499 * Allow certificate policies extension to use an IA5STRING for the
18500 organization field. This is contrary to the PKIX definition but
18501 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18506 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18507 without disallowing inline assembler and the like for non-pedantic builds.
18511 * Support Borland C++ builder.
18513 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18519 * SHA-1 cleanups and performance enhancements.
18521 *Andy Polyakov <appro@fy.chalmers.se>*
18523 * Sparc v8plus assembler for the bignum library.
18525 *Andy Polyakov <appro@fy.chalmers.se>*
18527 * Accept any -xxx and +xxx compiler options in Configure.
18531 * Update HPUX configuration.
18535 * Add missing `sk_<type>_unshift()` function to safestack.h
18537 *Ralf S. Engelschall*
18539 * New function SSL_CTX_use_certificate_chain_file that sets the
18540 "extra_cert"s in addition to the certificate. (This makes sense
18541 only for "PEM" format files, as chains as a whole are not
18546 * Support verify_depth from the SSL API.
18547 x509_vfy.c had what can be considered an off-by-one-error:
18548 Its depth (which was not part of the external interface)
18549 was actually counting the number of certificates in a chain;
18550 now it really counts the depth.
18554 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18555 instead of X509err, which often resulted in confusing error
18556 messages since the error codes are not globally unique
18557 (e.g. an alleged error in ssl3_accept when a certificate
18558 didn't match the private key).
18560 * New function SSL_CTX_set_session_id_context that allows to set a default
18561 value (so that you don't need SSL_set_session_id_context for each
18562 connection using the SSL_CTX).
18566 * OAEP decoding bug fix.
18570 * Support INSTALL_PREFIX for package builders, as proposed by
18575 * New Configure options "threads" and "no-threads". For systems
18576 where the proper compiler options are known (currently Solaris
18577 and Linux), "threads" is the default.
18581 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18585 * Install various scripts to $(OPENSSLDIR)/misc, not to
18586 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18587 such as /usr/local/bin.
18591 * "make linux-shared" to build shared libraries.
18593 *Niels Poppe <niels@netbox.org>*
18595 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18599 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18600 extension adding in x509 utility.
18604 * Remove NOPROTO sections and error code comments.
18608 * Partial rewrite of the DEF file generator to now parse the ANSI
18613 * New Configure options --prefix=DIR and --openssldir=DIR.
18617 * Complete rewrite of the error code script(s). It is all now handled
18618 by one script at the top level which handles error code gathering,
18619 header rewriting and C source file generation. It should be much better
18620 than the old method: it now uses a modified version of Ulf's parser to
18621 read the ANSI prototypes in all header files (thus the old K&R definitions
18622 aren't needed for error creation any more) and do a better job of
18623 translating function codes into names. The old 'ASN1 error code embedded
18624 in a comment' is no longer necessary and it doesn't use .err files which
18625 have now been deleted. Also the error code call doesn't have to appear all
18626 on one line (which resulted in some large lines...).
18630 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18634 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18635 0 (which usually indicates a closed connection), but continue reading.
18639 * Fix some race conditions.
18643 * Add support for CRL distribution points extension. Add Certificate
18644 Policies and CRL distribution points documentation.
18648 * Move the autogenerated header file parts to crypto/opensslconf.h.
18652 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18653 8 of keying material. Merlin has also confirmed interop with this fix
18654 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18656 *Merlin Hughes <merlin@baltimore.ie>*
18658 * Fix lots of warnings.
18660 *Richard Levitte <levitte@stacken.kth.se>*
18662 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18663 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18665 *Richard Levitte <levitte@stacken.kth.se>*
18667 * Fix problems with sizeof(long) == 8.
18669 *Andy Polyakov <appro@fy.chalmers.se>*
18671 * Change functions to ANSI C.
18675 * Fix typos in error codes.
18677 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18679 * Remove defunct assembler files from Configure.
18683 * SPARC v8 assembler BIGNUM implementation.
18685 *Andy Polyakov <appro@fy.chalmers.se>*
18687 * Support for Certificate Policies extension: both print and set.
18688 Various additions to support the r2i method this uses.
18692 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18693 return a const string when you are expecting an allocated buffer.
18697 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18698 types DirectoryString and DisplayText.
18702 * Add code to allow r2i extensions to access the configuration database,
18703 add an LHASH database driver and add several ctx helper functions.
18707 * Fix an evil bug in bn_expand2() which caused various BN functions to
18708 fail when they extended the size of a BIGNUM.
18712 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18713 support typesafe stack.
18717 * Fix typo in SSL_[gs]et_options().
18719 *Nils Frostberg <nils@medcom.se>*
18721 * Delete various functions and files that belonged to the (now obsolete)
18722 old X509V3 handling code.
18726 * New Configure option "rsaref".
18730 * Don't auto-generate pem.h.
18734 * Introduce type-safe ASN.1 SETs.
18738 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18740 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18742 * Introduce type-safe STACKs. This will almost certainly break lots of code
18743 that links with OpenSSL (well at least cause lots of warnings), but fear
18744 not: the conversion is trivial, and it eliminates loads of evil casts. A
18745 few STACKed things have been converted already. Feel free to convert more.
18746 In the fullness of time, I'll do away with the STACK type altogether.
18750 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18751 specified in `<certfile>` by updating the entry in the index.txt file.
18752 This way one no longer has to edit the index.txt file manually for
18753 revoking a certificate. The -revoke option does the gory details now.
18755 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18757 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18758 `-text` option at all and this way the `-noout -text` combination was
18759 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18761 *Ralf S. Engelschall*
18763 * Make sure a corresponding plain text error message exists for the
18764 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18765 verify callback function determined that a certificate was revoked.
18767 *Ralf S. Engelschall*
18769 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18770 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18771 all available ciphers including rc5, which was forgotten until now.
18772 In order to let the testing shell script know which algorithms
18773 are available, a new (up to now undocumented) command
18774 `openssl list-cipher-commands` is used.
18778 * Bugfix: s_client occasionally would sleep in select() when
18779 it should have checked SSL_pending() first.
18783 * New functions DSA_do_sign and DSA_do_verify to provide access to
18784 the raw DSA values prior to ASN.1 encoding.
18788 * Tweaks to Configure
18790 *Niels Poppe <niels@netbox.org>*
18792 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18797 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18801 * New config option to avoid instructions that are illegal on the 80386.
18802 The default code is faster, but requires at least a 486.
18806 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18807 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18808 same as SSL2_VERSION anyway.
18812 * New "-showcerts" option for s_client.
18816 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18817 application. Various cleanups and fixes.
18821 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18822 modify error routines to work internally. Add error codes and PBE init
18823 to library startup routines.
18827 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18828 packing functions to asn1 and evp. Changed function names and error
18829 codes along the way.
18833 * PKCS12 integration: and so it begins... First of several patches to
18834 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18835 objects to objects.h
18839 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18840 and display support for Thawte strong extranet extension.
18844 * Add LinuxPPC support.
18846 *Jeff Dubrule <igor@pobox.org>*
18848 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18849 bn_div_words in alpha.s.
18851 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18853 * Make sure the RSA OAEP test is skipped under -DRSAref because
18854 OAEP isn't supported when OpenSSL is built with RSAref.
18856 *Ulf Moeller <ulf@fitug.de>*
18858 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18859 so they no longer are missing under -DNOPROTO.
18861 *Soren S. Jorvang <soren@t.dk>*
18863 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18865 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18866 doesn't work when the session is reused. Coming soon!
18870 * Fix a security hole, that allows sessions to be reused in the wrong
18871 context thus bypassing client cert protection! All software that uses
18872 client certs and session caches in multiple contexts NEEDS PATCHING to
18873 allow session reuse! A fuller solution is in the works.
18875 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18877 * Some more source tree cleanups (removed obsolete files
18878 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18879 permission on "config" script to be executable) and a fix for the INSTALL
18882 *Ulf Moeller <ulf@fitug.de>*
18884 * Remove some legacy and erroneous uses of malloc, free instead of
18887 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18889 * Make rsa_oaep_test return non-zero on error.
18891 *Ulf Moeller <ulf@fitug.de>*
18893 * Add support for native Solaris shared libraries. Configure
18894 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18895 if someone would make that last step automatic.
18897 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18899 * ctx_size was not built with the right compiler during "make links". Fixed.
18903 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18904 except NULL ciphers". This means the default cipher list will no longer
18905 enable NULL ciphers. They need to be specifically enabled e.g. with
18906 the string "DEFAULT:eNULL".
18910 * Fix to RSA private encryption routines: if p < q then it would
18911 occasionally produce an invalid result. This will only happen with
18912 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18916 * Be less restrictive and allow also `perl util/perlpath.pl
18917 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18918 because this way one can also use an interpreter named `perl5` (which is
18919 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18920 installed as `perl`).
18922 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18924 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18926 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18928 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18929 advapi32.lib to Win32 build and change the pem test comparison
18930 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18931 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18932 and crypto/des/ede_cbcm_enc.c.
18936 * DES quad checksum was broken on big-endian architectures. Fixed.
18940 * Comment out two functions in bio.h that aren't implemented. Fix up the
18941 Win32 test batch file so it (might) work again. The Win32 test batch file
18942 is horrible: I feel ill....
18946 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18947 in e_os.h. Audit of header files to check ANSI and non ANSI
18948 sections: 10 functions were absent from non ANSI section and not exported
18949 from Windows DLLs. Fixed up libeay.num for new functions.
18953 * Make `openssl version` output lines consistent.
18955 *Ralf S. Engelschall*
18957 * Fix Win32 symbol export lists for BIO functions: Added
18958 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18959 to ms/libeay{16,32}.def.
18961 *Ralf S. Engelschall*
18963 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18964 fine under Unix and passes some trivial tests I've now added. But the
18965 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18966 added to make sure no one expects that this stuff really works in the
18967 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18968 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18971 *Ralf S. Engelschall*
18973 * Fix the generation of two part addresses in perl.
18975 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18977 * Add config entry for Linux on MIPS.
18979 *John Tobey <jtobey@channel1.com>*
18981 * Make links whenever Configure is run, unless we are on Windoze.
18985 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18986 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18991 * Add a useful kludge to allow package maintainers to specify compiler and
18992 other platforms details on the command line without having to patch the
18993 Configure script every time: One now can use
18994 `perl Configure <id>:<details>`,
18995 i.e. platform ids are allowed to have details appended
18996 to them (separated by colons). This is treated as there would be a static
18997 pre-configured entry in Configure's %table under key `<id>` with value
18998 `<details>` and `perl Configure <id>` is called. So, when you want to
18999 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19000 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19001 now, which overrides the FreeBSD-elf entry on-the-fly.
19003 *Ralf S. Engelschall*
19005 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19009 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19010 on the `perl Configure ...` command line. This way one can compile
19011 OpenSSL libraries with Position Independent Code (PIC) which is needed
19012 for linking it into DSOs.
19014 *Ralf S. Engelschall*
19016 * Remarkably, export ciphers were totally broken and no-one had noticed!
19021 * Cleaned up the LICENSE document: The official contact for any license
19022 questions now is the OpenSSL core team under openssl-core@openssl.org.
19023 And add a paragraph about the dual-license situation to make sure people
19024 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19025 to the OpenSSL toolkit.
19027 *Ralf S. Engelschall*
19029 * General source tree makefile cleanups: Made `making xxx in yyy...`
19030 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19031 Additionally cleaned up the `make links` target: Remove unnecessary
19032 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19033 to speed processing and no longer clutter the display with confusing
19034 stuff. Instead only the actually done links are displayed.
19036 *Ralf S. Engelschall*
19038 * Permit null encryption ciphersuites, used for authentication only. It used
19039 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19040 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19045 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19046 signed attributes when verifying signatures (this would break them),
19047 the detached data encoding was wrong and public keys obtained using
19048 X509_get_pubkey() weren't freed.
19052 * Add text documentation for the BUFFER functions. Also added a work around
19053 to a Win95 console bug. This was triggered by the password read stuff: the
19054 last character typed gets carried over to the next fread(). If you were
19055 generating a new cert request using 'req' for example then the last
19056 character of the passphrase would be CR which would then enter the first
19061 * Added the new 'Includes OpenSSL Cryptography Software' button as
19062 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19063 button and can be used by applications based on OpenSSL to show the
19064 relationship to the OpenSSL project.
19066 *Ralf S. Engelschall*
19068 * Remove confusing variables in function signatures in files
19069 ssl/ssl_lib.c and ssl/ssl.h.
19071 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19073 * Don't install bss_file.c under PREFIX/include/
19075 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19077 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19078 functions that return function pointers and has support for NT specific
19079 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19080 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19081 unsigned to signed types: this was killing the Win32 compile.
19085 * Add new certificate file to stack functions,
19086 SSL_add_dir_cert_subjects_to_stack() and
19087 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19088 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19089 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19090 This means that Apache-SSL and similar packages don't have to mess around
19091 to add as many CAs as they want to the preferred list.
19095 * Experiment with doxygen documentation. Currently only partially applied to
19097 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19098 openssl.doxy as the configuration file.
19102 * Get rid of remaining C++-style comments which strict C compilers hate.
19104 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19106 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19107 compiled in by default: it has problems with large keys.
19111 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19112 DH private keys and/or callback functions which directly correspond to
19113 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19114 is needed for applications which have to configure certificates on a
19115 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19117 For the RSA certificate situation is makes no difference, but
19118 for the DSA certificate situation this fixes the "no shared cipher"
19119 problem where the OpenSSL cipher selection procedure failed because the
19120 temporary keys were not overtaken from the context and the API provided
19121 no way to reconfigure them.
19122 The new functions now let applications reconfigure the stuff and they
19123 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19124 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19125 non-public-API function ssl_cert_instantiate() is used as a helper
19126 function and also to reduce code redundancy inside ssl_rsa.c.
19128 *Ralf S. Engelschall*
19130 * Move s_server -dcert and -dkey options out of the undocumented feature
19131 area because they are useful for the DSA situation and should be
19132 recognized by the users.
19134 *Ralf S. Engelschall*
19136 * Fix the cipher decision scheme for export ciphers: the export bits are
19137 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19138 SSL_EXP_MASK. So, the original variable has to be used instead of the
19139 already masked variable.
19141 *Richard Levitte <levitte@stacken.kth.se>*
19143 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19145 *Richard Levitte <levitte@stacken.kth.se>*
19147 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19148 from `int` to `unsigned int` because it is a length and initialized by
19149 EVP_DigestFinal() which expects an `unsigned int *`.
19151 *Richard Levitte <levitte@stacken.kth.se>*
19153 * Don't hard-code path to Perl interpreter on shebang line of Configure
19154 script. Instead use the usual Shell->Perl transition trick.
19156 *Ralf S. Engelschall*
19158 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19159 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19160 -noout -modulus` as it's already the case for `openssl rsa -noout
19161 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19162 currently the public key is printed (a decision which was already done by
19163 `openssl dsa -modulus` in the past) which serves a similar purpose.
19164 Additionally the NO_RSA no longer completely removes the whole -modulus
19165 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19168 *Ralf S. Engelschall*
19170 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19171 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19173 *Arne Ansper <arne@ats.cyber.ee>*
19175 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19176 to be added. Now both 'req' and 'ca' can use new objects defined in the
19181 * Add cool BIO that does syslog (or event log on NT).
19183 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19185 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19186 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19187 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19188 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19192 * Add preliminary config info for new extension code.
19196 * Make RSA_NO_PADDING really use no padding.
19198 *Ulf Moeller <ulf@fitug.de>*
19200 * Generate errors when private/public key check is done.
19204 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19205 for some CRL extensions and new objects added.
19209 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19210 key usage extension and fuller support for authority key id.
19214 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19215 padding method for RSA, which is recommended for new applications in PKCS
19216 #1 v2.0 (RFC 2437, October 1998).
19217 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19218 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19219 against Bleichbacher's attack on RSA.
19220 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19223 * Updates to the new SSL compression code
19225 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19227 * Fix so that the version number in the master secret, when passed
19228 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19229 (because the server will not accept higher), that the version number
19230 is 0x03,0x01, not 0x03,0x00
19232 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19234 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19235 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19236 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19240 * Support for RAW extensions where an arbitrary extension can be
19241 created by including its DER encoding. See `apps/openssl.cnf` for
19246 * Make sure latest Perl versions don't interpret some generated C array
19247 code as Perl array code in the crypto/err/err_genc.pl script.
19249 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19251 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19252 not many people have the assembler. Various Win32 compilation fixes and
19253 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19254 build instructions.
19258 * Modify configure script 'Configure' to automatically create crypto/date.h
19259 file under Win32 and also build pem.h from pem.org. New script
19260 util/mkfiles.pl to create the MINFO file on environments that can't do a
19261 'make files': perl util/mkfiles.pl >MINFO should work.
19265 * Major rework of DES function declarations, in the pursuit of correctness
19266 and purity. As a result, many evil casts evaporated, and some weirdness,
19267 too. You may find this causes warnings in your code. Zapping your evil
19268 casts will probably fix them. Mostly.
19272 * Fix for a typo in asn1.h. Bug fix to object creation script
19273 obj_dat.pl. It considered a zero in an object definition to mean
19274 "end of object": none of the objects in objects.h have any zeros
19275 so it wasn't spotted.
19277 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19279 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19280 Masking (CBCM). In the absence of test vectors, the best I have been able
19281 to do is check that the decrypt undoes the encrypt, so far. Send me test
19282 vectors if you have them.
19286 * Correct calculation of key length for export ciphers (too much space was
19287 allocated for null ciphers). This has not been tested!
19291 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19292 message is now correct (it understands "crypto" and "ssl" on its
19293 command line). There is also now an "update" option. This will update
19294 the util/ssleay.num and util/libeay.num files with any new functions.
19296 perl util/mkdef.pl crypto ssl update
19297 it will update them.
19301 * Overhauled the Perl interface:
19302 - ported BN stuff to OpenSSL's different BN library
19303 - made the perl/ source tree CVS-aware
19304 - renamed the package from SSLeay to OpenSSL (the files still contain
19305 their history because I've copied them in the repository)
19306 - removed obsolete files (the test scripts will be replaced
19307 by better Test::Harness variants in the future)
19309 *Ralf S. Engelschall*
19311 * First cut for a very conservative source tree cleanup:
19312 1. merge various obsolete readme texts into doc/ssleay.txt
19313 where we collect the old documents and readme texts.
19314 2. remove the first part of files where I'm already sure that we no
19315 longer need them because of three reasons: either they are just temporary
19316 files which were left by Eric or they are preserved original files where
19317 I've verified that the diff is also available in the CVS via "cvs diff
19318 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19319 the crypto/md/ stuff).
19321 *Ralf S. Engelschall*
19323 * More extension code. Incomplete support for subject and issuer alt
19324 name, issuer and authority key id. Change the i2v function parameters
19325 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19326 what that's for :-) Fix to ASN1 macro which messed up
19327 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19331 * Preliminary support for ENUMERATED type. This is largely copied from the
19336 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19338 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19340 * Make sure `make rehash` target really finds the `openssl` program.
19342 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19344 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19345 like to hear about it if this slows down other processors.
19349 * Add CygWin32 platform information to Configure script.
19351 *Alan Batie <batie@aahz.jf.intel.com>*
19353 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19355 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19357 * New program nseq to manipulate netscape certificate sequences
19361 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19366 * Fixes to BN code. Previously the default was to define BN_RECURSION
19367 but the BN code had some problems that would cause failures when
19368 doing certificate verification and some other functions.
19370 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19372 * Add ASN1 and PEM code to support netscape certificate sequences.
19376 * Add ASN1 and PEM code to support netscape certificate sequences.
19380 * Add several PKIX and private extended key usage OIDs.
19384 * Modify the 'ca' program to handle the new extension code. Modify
19385 openssl.cnf for new extension format, add comments.
19389 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19390 and add a sample to openssl.cnf so req -x509 now adds appropriate
19395 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19396 error code, add initial support to X509_print() and x509 application.
19400 * Takes a deep breath and start adding X509 V3 extension support code. Add
19401 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19402 stuff is currently isolated and isn't even compiled yet.
19406 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19407 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19408 Removed the versions check from X509 routines when loading extensions:
19409 this allows certain broken certificates that don't set the version
19410 properly to be processed.
19414 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19415 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19416 can still be regenerated with "make depend".
19420 * Spelling mistake in C version of CAST-128.
19422 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19424 * Changes to the error generation code. The perl script err-code.pl
19425 now reads in the old error codes and retains the old numbers, only
19426 adding new ones if necessary. It also only changes the .err files if new
19427 codes are added. The makefiles have been modified to only insert errors
19428 when needed (to avoid needlessly modifying header files). This is done
19429 by only inserting errors if the .err file is newer than the auto generated
19430 C file. To rebuild all the error codes from scratch (the old behaviour)
19431 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19432 or delete all the .err files.
19436 * CAST-128 was incorrectly implemented for short keys. The C version has
19437 been fixed, but is untested. The assembler versions are also fixed, but
19438 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19439 to regenerate it if needed.
19440 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19441 Hagino <itojun@kame.net>*
19443 * File was opened incorrectly in randfile.c.
19445 *Ulf Möller <ulf@fitug.de>*
19447 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19448 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19449 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19450 al: it's just almost always a UTCTime. Note this patch adds new error
19451 codes so do a "make errors" if there are problems.
19455 * Correct Linux 1 recognition in config.
19457 *Ulf Möller <ulf@fitug.de>*
19459 * Remove pointless MD5 hash when using DSA keys in ca.
19461 *Anonymous <nobody@replay.com>*
19463 * Generate an error if given an empty string as a cert directory. Also
19464 generate an error if handed NULL (previously returned 0 to indicate an
19465 error, but didn't set one).
19467 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19469 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19473 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19474 parameters. This was causing a warning which killed off the Win32 compile.
19478 * Remove C++ style comments from crypto/bn/bn_local.h.
19480 *Neil Costigan <neil.costigan@celocom.com>*
19482 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19483 based on a text string, looking up short and long names and finally
19484 "dot" format. The "dot" format stuff didn't work. Added new function
19485 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19486 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19487 OID is not part of the table.
19491 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19492 X509_LOOKUP_by_alias().
19496 * Sort openssl functions by name.
19500 * Get the `gendsa` command working and add it to the `list` command. Remove
19501 encryption from sample DSA keys (in case anyone is interested the password
19506 * Make *all* `*_free` functions accept a NULL pointer.
19508 *Frans Heymans <fheymans@isaserver.be>*
19510 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19513 *Anonymous <nobody@replay.com>*
19515 * s_server should send the CAfile as acceptable CAs, not its own cert.
19517 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19519 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19521 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19523 * Temp key "for export" tests were wrong in s3_srvr.c.
19525 *Anonymous <nobody@replay.com>*
19527 * Add prototype for temp key callback functions
19528 SSL_CTX_set_tmp_{rsa,dh}_callback().
19532 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19533 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19537 * X509_name_add_entry() freed the wrong thing after an error.
19539 *Arne Ansper <arne@ats.cyber.ee>*
19541 * rsa_eay.c would attempt to free a NULL context.
19543 *Arne Ansper <arne@ats.cyber.ee>*
19545 * BIO_s_socket() had a broken should_retry() on Windoze.
19547 *Arne Ansper <arne@ats.cyber.ee>*
19549 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19551 *Arne Ansper <arne@ats.cyber.ee>*
19553 * Make sure the already existing X509_STORE->depth variable is initialized
19554 in X509_STORE_new(), but document the fact that this variable is still
19555 unused in the certificate verification process.
19557 *Ralf S. Engelschall*
19559 * Fix the various library and `apps/` files to free up pkeys obtained from
19560 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19564 * Fix reference counting in X509_PUBKEY_get(). This makes
19565 demos/maurice/example2.c work, amongst others, probably.
19567 *Steve Henson and Ben Laurie*
19569 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19570 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19571 are no longer created. This way we have a single and consistent command
19572 line interface `openssl <command>`, similar to `cvs <command>`.
19574 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19576 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19577 BIT STRING wrapper always have zero unused bits.
19581 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19585 * Make the top-level INSTALL documentation easier to understand.
19589 * Makefiles updated to exit if an error occurs in a sub-directory
19590 make (including if user presses ^C) [Paul Sutton]
19592 * Make Montgomery context stuff explicit in RSA data structure.
19596 * Fix build order of pem and err to allow for generated pem.h.
19600 * Fix renumbering bug in X509_NAME_delete_entry().
19604 * Enhanced the err-ins.pl script so it makes the error library number
19605 global and can add a library name. This is needed for external ASN1 and
19606 other error libraries.
19610 * Fixed sk_insert which never worked properly.
19614 * Fix ASN1 macros so they can handle indefinite length constructed
19615 EXPLICIT tags. Some non standard certificates use these: they can now
19620 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19621 into a single doc/ssleay.txt bundle. This way the information is still
19622 preserved but no longer messes up this directory. Now it's new room for
19623 the new set of documentation files.
19625 *Ralf S. Engelschall*
19627 * SETs were incorrectly DER encoded. This was a major pain, because they
19628 shared code with SEQUENCEs, which aren't coded the same. This means that
19629 almost everything to do with SETs or SEQUENCEs has either changed name or
19630 number of arguments.
19632 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19634 * Fix test data to work with the above.
19638 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19639 was already fixed by Eric for 0.9.1 it seems.
19641 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19643 * Autodetect FreeBSD3.
19647 * Fix various bugs in Configure. This affects the following platforms:
19651 unixware-2.0-pentium
19656 * Eliminate generated files from CVS. Reorder tests to regenerate files
19657 before they are needed.
19661 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19665 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19667 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19668 changed SSLeay to OpenSSL in version strings.
19670 *Ralf S. Engelschall*
19672 * Some fixups to the top-level documents.
19676 * Fixed the nasty bug where rsaref.h was not found under compile-time
19677 because the symlink to include/ was missing.
19679 *Ralf S. Engelschall*
19681 * Incorporated the popular no-RSA/DSA-only patches
19682 which allow to compile an RSA-free SSLeay.
19684 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19686 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19687 when "ssleay" is still not found.
19689 *Ralf S. Engelschall*
19691 * Added more platforms to Configure: Cray T3E, HPUX 11,
19693 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19695 * Updated the README file.
19697 *Ralf S. Engelschall*
19699 * Added various .cvsignore files in the CVS repository subdirs
19700 to make a "cvs update" really silent.
19702 *Ralf S. Engelschall*
19704 * Recompiled the error-definition header files and added
19705 missing symbols to the Win32 linker tables.
19707 *Ralf S. Engelschall*
19709 * Cleaned up the top-level documents;
19710 o new files: CHANGES and LICENSE
19711 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19712 o merged COPYRIGHT into LICENSE
19713 o removed obsolete TODO file
19714 o renamed MICROSOFT to INSTALL.W32
19716 *Ralf S. Engelschall*
19718 * Removed dummy files from the 0.9.1b source tree:
19719 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19720 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19721 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19722 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19723 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19725 *Ralf S. Engelschall*
19727 * Added various platform portability fixes.
19731 * The Genesis of the OpenSSL rpject:
19732 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19733 Young and Tim J. Hudson created while they were working for C2Net until
19736 *The OpenSSL Project*
19738 ### Changes between 0.9.0b and 0.9.1b [not released]
19740 * Updated a few CA certificates under certs/
19744 * Changed some BIGNUM api stuff.
19748 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19749 DGUX x86, Linux Alpha, etc.
19753 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19754 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19759 * Add -strparse option to asn1pars program which parses nested
19762 *Dr Stephen Henson <shenson@bigfoot.com>*
19764 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19768 * DSA fix for "ca" program.
19772 * Added "-genkey" option to "dsaparam" program.
19776 * Added RIPE MD160 (rmd160) message digest.
19780 * Added -a (all) option to "ssleay version" command.
19784 * Added PLATFORM define which is the id given to Configure.
19788 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19792 * Extended the ASN.1 parser routines.
19796 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19800 * Added a BN_CTX to the BN library.
19804 * Fixed the weak key values in DES library
19808 * Changed API in EVP library for cipher aliases.
19812 * Added support for RC2/64bit cipher.
19816 * Converted the lhash library to the crypto/mem.c functions.
19820 * Added more recognized ASN.1 object ids.
19824 * Added more RSA padding checks for SSL/TLS.
19828 * Added BIO proxy/filter functionality.
19832 * Added extra_certs to SSL_CTX which can be used
19833 send extra CA certificates to the client in the CA cert chain sending
19834 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19838 * Now Fortezza is denied in the authentication phase because
19839 this is key exchange mechanism is not supported by SSLeay at all.
19843 * Additional PKCS1 checks.
19847 * Support the string "TLSv1" for all TLS v1 ciphers.
19851 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19852 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19856 * Fixed a few memory leaks.
19860 * Fixed various code and comment typos.
19864 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19865 bytes sent in the client random.
19867 *Edward Bishop <ebishop@spyglass.com>*
19871 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
19872 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
19873 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
19874 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
19875 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
19876 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
19877 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
19878 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
19879 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
19880 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19881 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
19882 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19883 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19884 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19885 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19886 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19887 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19888 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19889 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19890 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19891 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19892 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19893 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19894 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19895 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19896 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19897 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19898 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19899 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19900 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19901 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19902 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19903 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19904 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19905 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19906 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19907 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19908 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19909 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19910 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19911 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19912 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19913 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19914 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19915 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19916 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19917 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19918 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19919 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19920 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19921 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19922 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19923 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19924 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19925 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19926 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19927 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19928 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19929 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19930 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19931 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19932 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19933 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19934 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19935 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19936 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19937 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19938 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19939 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19940 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19941 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19942 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19943 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19944 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19945 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19946 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19947 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19948 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19949 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19950 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19951 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19952 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19953 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19954 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19955 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19956 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19957 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19958 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19959 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19960 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19961 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19962 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19963 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19964 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19965 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19966 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19967 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19968 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19969 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19970 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19971 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19972 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19973 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19974 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19975 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19976 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19977 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19978 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19979 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19980 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19981 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19982 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19983 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19984 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19985 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19986 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19987 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19988 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19989 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19990 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19991 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19992 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19993 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19994 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19995 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19996 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19997 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19998 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19999 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20000 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20001 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20002 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20003 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20004 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20005 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20006 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20007 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20008 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20009 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20010 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20011 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20012 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20013 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20014 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20015 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20016 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20017 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20018 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20019 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20020 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20021 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20022 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20023 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20024 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20025 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20026 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20027 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20028 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20029 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20030 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20031 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20032 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20033 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20034 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20035 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20036 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20037 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20038 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20039 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20040 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20041 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20042 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20043 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655