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 Hybrid Public Key Encryption (HPKE) as defined
34 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
35 Message Layer Security (MLS) and other IETF specifications.
36 HPKE can also be used by other applications that require
37 encrypting "to" an ECDH public key. External APIs are defined in
38 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
42 * Add support for certificate compression (RFC8879), including
43 library support for Brotli and Zstandard compression.
47 * Add the ability to add custom attributes to PKCS12 files. Add a new API
48 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
49 for a user specified callback and optional argument.
50 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
51 added to the existing STACK_OF attrs.
55 * Major refactor of the libssl record layer
59 * Added a new BIO_s_dgram_mem() to read/write datagrams to memory
63 * Add a mac salt length option for the pkcs12 command.
67 * Add more SRTP protection profiles from RFC8723 and RFC8269.
71 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
73 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
75 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
76 supported and enabled.
80 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
81 to the list of ciphersuites providing Perfect Forward Secrecy as
82 required by SECLEVEL >= 3.
84 *Dmitry Belyavskiy, Nicola Tuveri*
86 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
87 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
88 SSL_get0_iana_groups() function-like macro, retrieves the list of
89 supported groups sent by the peer.
90 The function SSL_client_hello_get_extension_order() populates
91 a caller-supplied array with the list of extension types present in the
92 ClientHello, in order of appearance.
96 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
97 to make it possible to use empty passphrase strings.
101 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
102 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
103 of 160 bits and above and less than 224 bits were previously accepted by
104 default but are now no longer allowed. By default TLS compression was
105 already disabled in previous OpenSSL versions. At security level 2 it cannot
110 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
115 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
116 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
117 will need to load the legacy crypto provider.
120 * CCM8 cipher suites in TLS have been downgraded to security level zero
121 because they use a short authentication tag which lowers their strength.
125 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
130 * Add X.509 certificate codeSigning purpose and related checks on key usage and
131 extended key usage of the leaf certificate according to the CA/Browser Forum.
135 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
136 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
137 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
138 X.509 version 3 if the certificate information includes X.509 extensions.
142 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
143 such as adding a trace facility for debugging certificate chain building.
147 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
148 in particular supporting requests for central key generation, generalized
149 polling, and various types of genm/genp exchanges defined in CMP Updates.
153 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
154 like correcting the TLS and proxy support and adding tracing for debugging.
158 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
162 * Fixed and extended `util/check-format.pl` for checking adherence to the
163 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
164 The checks are meanwhile more complete and yield fewer false positives.
168 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
169 sending and receiving multiple messages in a single call. An implementation
170 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
174 * The `SSL_CERT_PATH` and `SSL_CERT_URI` environment variables are introduced.
175 `SSL_CERT_URI` can be used to specify a URI for a root certificate store. The
176 `SSL_CERT_PATH` environment variable specifies a delimiter-separated list of
177 paths which are searched for root certificates.
179 The existing `SSL_CERT_DIR` environment variable is deprecated.
180 `SSL_CERT_DIR` was previously used to specify either a delimiter-separated
181 list of paths or an URI, which is ambiguous. Setting `SSL_CERT_PATH` causes
182 `SSL_CERT_DIR` to be ignored for the purposes of determining root certificate
183 directories, and setting `SSL_CERT_URI` causes `SSL_CERT_DIR` to be ignored
184 for the purposes of determining root certificate stores.
188 * Support for loading root certificates from the Windows certificate store
189 has been added. The support is in the form of a store which recognises the
190 URI string of `org.openssl.winstore://`. This store is enabled by default and
191 can be disabled using the new compile-time option `no-winstore`.
195 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
196 kernel versions that support KTLS have a known bug in CCM processing. That
197 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
198 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
203 * Zerocopy KTLS sendfile() support on Linux.
207 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
208 decryption as a protection against Bleichenbacher-like attacks.
209 The RSA decryption API will now return a randomly generated deterministic
210 message instead of an error in case it detects an error when checking
211 padding during PKCS#1 v1.5 decryption. This is a general protection against
212 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
214 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
215 on the RSA decryption context.
222 ### Changes between 3.0 and 3.1.0 [xx XXX xxxx]
224 * The FIPS provider includes a few non-approved algorithms for
225 backward compatibility purposes and the "fips=yes" property query
226 must be used for all algorithm fetches to ensure FIPS compliance.
228 The algorithms that are included but not approved are Triple DES ECB,
229 Triple DES CBC and EdDSA.
233 * Added support for KMAC in KBKDF.
237 * RNDR and RNDRRS support in provider functions to provide
238 random number generation for Arm CPUs (aarch64).
242 * s_client and s_server apps now explicitly say when the TLS version
243 does not include the renegotiation mechanism. This avoids confusion
244 between that scenario versus when the TLS version includes secure
245 renegotiation but the peer lacks support for it.
249 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
251 *Tomasz Kantecki, Andrey Matyukov*
253 * The various OBJ_* functions have been made thread safe.
257 * Parallel dual-prime 1536/2048-bit modular exponentiation for
258 AVX512_IFMA capable processors.
260 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
262 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
263 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
264 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
265 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
266 `OPENSSL_NO_DEPRECATED_3_1`.
268 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
269 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
270 definitions for these functions regardless of whether
271 `OPENSSL_NO_DEPRECATED_3_1` is defined.
273 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
274 functions regardless of whether they are using them. It is recommended that
275 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
279 * When generating safe-prime DH parameters set the recommended private key
280 length equivalent to minimum key lengths as in RFC 7919.
284 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
285 maximum size that is smaller or equal to the digest length to comply with
286 FIPS 186-4 section 5. This is implemented by a new option
287 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
288 `rsa_pss_saltlen` parameter, which is now the default. Signature
289 verification is not affected by this change and continues to work as before.
296 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
297 listed here are only a brief description.
298 The migration guide contains more detailed information related to new features,
299 breaking changes, and mappings for the large list of deprecated functions.
301 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
303 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
305 * Fixed NULL dereference during PKCS7 data verification.
307 A NULL pointer can be dereferenced when signatures are being
308 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
309 algorithm used for the signature is known to the OpenSSL library but
310 the implementation of the hash algorithm is not available the digest
311 initialization will fail. There is a missing check for the return
312 value from the initialization function which later leads to invalid
313 usage of the digest API most likely leading to a crash.
316 PKCS7 data is processed by the SMIME library calls and also by the
317 time stamp (TS) library calls. The TLS implementation in OpenSSL does
318 not call these functions however third party applications would be
319 affected if they call these functions to verify signatures on untrusted
324 * Fixed X.400 address type confusion in X.509 GeneralName.
326 There is a type confusion vulnerability relating to X.400 address processing
327 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
328 but the public structure definition for GENERAL_NAME incorrectly specified
329 the type of the x400Address field as ASN1_TYPE. This field is subsequently
330 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
333 When CRL checking is enabled (i.e. the application sets the
334 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
335 pass arbitrary pointers to a memcmp call, enabling them to read memory
336 contents or enact a denial of service.
341 * Fixed NULL dereference validating DSA public key.
343 An invalid pointer dereference on read can be triggered when an
344 application tries to check a malformed DSA public key by the
345 EVP_PKEY_public_check() function. This will most likely lead
346 to an application crash. This function can be called on public
347 keys supplied from untrusted sources which could allow an attacker
348 to cause a denial of service attack.
350 The TLS implementation in OpenSSL does not call this function
351 but applications might call the function if there are additional
352 security requirements imposed by standards such as FIPS 140-3.
355 *Shane Lontis, Tomáš Mráz*
357 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
359 An invalid pointer dereference on read can be triggered when an
360 application tries to load malformed PKCS7 data with the
361 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
363 The result of the dereference is an application crash which could
364 lead to a denial of service attack. The TLS implementation in OpenSSL
365 does not call this function however third party applications might
366 call these functions on untrusted data.
371 * Fixed Use-after-free following BIO_new_NDEF.
373 The public API function BIO_new_NDEF is a helper function used for
374 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
375 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
376 be called directly by end user applications.
378 The function receives a BIO from the caller, prepends a new BIO_f_asn1
379 filter BIO onto the front of it to form a BIO chain, and then returns
380 the new head of the BIO chain to the caller. Under certain conditions,
381 for example if a CMS recipient public key is invalid, the new filter BIO
382 is freed and the function returns a NULL result indicating a failure.
383 However, in this case, the BIO chain is not properly cleaned up and the
384 BIO passed by the caller still retains internal pointers to the previously
385 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
386 then a use-after-free will occur. This will most likely result in a crash.
389 *Viktor Dukhovni, Matt Caswell*
391 * Fixed Double free after calling PEM_read_bio_ex.
393 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
394 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
395 data. If the function succeeds then the "name_out", "header" and "data"
396 arguments are populated with pointers to buffers containing the relevant
397 decoded data. The caller is responsible for freeing those buffers. It is
398 possible to construct a PEM file that results in 0 bytes of payload data.
399 In this case PEM_read_bio_ex() will return a failure code but will populate
400 the header argument with a pointer to a buffer that has already been freed.
401 If the caller also frees this buffer then a double free will occur. This
402 will most likely lead to a crash.
404 The functions PEM_read_bio() and PEM_read() are simple wrappers around
405 PEM_read_bio_ex() and therefore these functions are also directly affected.
407 These functions are also called indirectly by a number of other OpenSSL
408 functions including PEM_X509_INFO_read_bio_ex() and
409 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
410 internal uses of these functions are not vulnerable because the caller does
411 not free the header argument if PEM_read_bio_ex() returns a failure code.
414 *Kurt Roeckx, Matt Caswell*
416 * Fixed Timing Oracle in RSA Decryption.
418 A timing based side channel exists in the OpenSSL RSA Decryption
419 implementation which could be sufficient to recover a plaintext across
420 a network in a Bleichenbacher style attack. To achieve a successful
421 decryption an attacker would have to be able to send a very large number
422 of trial messages for decryption. The vulnerability affects all RSA padding
423 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
426 *Dmitry Belyavsky, Hubert Kario*
428 * Fixed X.509 Name Constraints Read Buffer Overflow.
430 A read buffer overrun can be triggered in X.509 certificate verification,
431 specifically in name constraint checking. The read buffer overrun might
432 result in a crash which could lead to a denial of service attack.
433 In a TLS client, this can be triggered by connecting to a malicious
434 server. In a TLS server, this can be triggered if the server requests
435 client authentication and a malicious client connects.
440 * Fixed X.509 Policy Constraints Double Locking security issue.
442 If an X.509 certificate contains a malformed policy constraint and
443 policy processing is enabled, then a write lock will be taken twice
444 recursively. On some operating systems (most widely: Windows) this
445 results in a denial of service when the affected process hangs. Policy
446 processing being enabled on a publicly facing server is not considered
447 to be a common setup.
452 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
453 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
454 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
455 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
456 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
457 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
458 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
459 for legacy EC and SM2 keys is also changed similarly to honor the
460 equivalent conversion format flag as specified in the underlying
461 `EC_KEY` object being exported to a provider, when this function is
462 called through `EVP_PKEY_export()`.
466 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
468 * Fixed two buffer overflows in punycode decoding functions.
470 A buffer overrun can be triggered in X.509 certificate verification,
471 specifically in name constraint checking. Note that this occurs after
472 certificate chain signature verification and requires either a CA to
473 have signed the malicious certificate or for the application to continue
474 certificate verification despite failure to construct a path to a trusted
477 In a TLS client, this can be triggered by connecting to a malicious
478 server. In a TLS server, this can be triggered if the server requests
479 client authentication and a malicious client connects.
481 An attacker can craft a malicious email address to overflow
482 an arbitrary number of bytes containing the `.` character (decimal 46)
483 on the stack. This buffer overflow could result in a crash (causing a
487 An attacker can craft a malicious email address to overflow four
488 attacker-controlled bytes on the stack. This buffer overflow could
489 result in a crash (causing a denial of service) or potentially remote code
490 execution depending on stack layout for any given platform/compiler.
495 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
496 parameters in OpenSSL code.
497 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
498 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
499 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
500 Using these invalid names may cause algorithms to use slower methods
501 that ignore the CRT parameters.
505 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
510 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
511 data to be signed before signing the certificate.
515 * Added RIPEMD160 to the default provider.
519 * Ensured that the key share group sent or accepted for the key exchange
520 is allowed for the protocol version.
524 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
526 * OpenSSL supports creating a custom cipher via the legacy
527 EVP_CIPHER_meth_new() function and associated function calls. This function
528 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
529 to use the new provider mechanism in order to implement custom ciphers.
531 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
532 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
533 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
534 and decryption initialisation functions). Instead of using the custom cipher
535 directly it incorrectly tries to fetch an equivalent cipher from the
536 available providers. An equivalent cipher is found based on the NID passed to
537 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
538 given cipher. However it is possible for an application to incorrectly pass
539 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
540 is used in this way the OpenSSL encryption/decryption initialisation function
541 will match the NULL cipher as being equivalent and will fetch this from the
542 available providers. This will succeed if the default provider has been
543 loaded (or if a third party provider has been loaded that offers this
544 cipher). Using the NULL cipher means that the plaintext is emitted as the
547 Applications are only affected by this issue if they call
548 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
549 encryption/decryption initialisation function. Applications that only use
550 SSL/TLS are not impacted by this issue.
555 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
560 * Fixed the linux-mips64 Configure target which was missing the
561 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
566 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
571 * Correctly handle a retransmitted ClientHello in DTLS
575 * Fixed detection of ktls support in cross-compile environment on Linux
579 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
584 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
585 report correct results in some cases
589 * Fix UWP builds by defining VirtualLock
593 * For known safe primes use the minimum key length according to RFC 7919.
594 Longer private key sizes unnecessarily raise the cycles needed to compute the
595 shared secret without any increase of the real security. This fixes a
596 regression from 1.1.1 where these shorter keys were generated for the known
601 * Added the loongarch64 target
605 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
606 only passed to the FIPS provider and not to the default or legacy provider.
610 * Fixed reported performance degradation on aarch64. Restored the
611 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
612 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
613 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
614 The new algorithm is still used for 32 bit targets.
618 * Added a missing header for memcmp that caused compilation failure on some
623 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
625 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
626 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
627 This issue makes the RSA implementation with 2048 bit private keys
628 incorrect on such machines and memory corruption will happen during
629 the computation. As a consequence of the memory corruption an attacker
630 may be able to trigger a remote code execution on the machine performing
633 SSL/TLS servers or other servers using 2048 bit RSA private keys running
634 on machines supporting AVX512IFMA instructions of the X86_64 architecture
635 are affected by this issue.
640 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
641 implementation would not encrypt the entirety of the data under some
642 circumstances. This could reveal sixteen bytes of data that was
643 preexisting in the memory that wasn't written. In the special case of
644 "in place" encryption, sixteen bytes of the plaintext would be revealed.
646 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
647 they are both unaffected.
650 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
652 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
654 * In addition to the c_rehash shell command injection identified in
655 CVE-2022-1292, further bugs where the c_rehash script does not
656 properly sanitise shell metacharacters to prevent command injection have been
659 When the CVE-2022-1292 was fixed it was not discovered that there
660 are other places in the script where the file names of certificates
661 being hashed were possibly passed to a command executed through the shell.
663 This script is distributed by some operating systems in a manner where
664 it is automatically executed. On such operating systems, an attacker
665 could execute arbitrary commands with the privileges of the script.
667 Use of the c_rehash script is considered obsolete and should be replaced
668 by the OpenSSL rehash command line tool.
671 *Daniel Fiala, Tomáš Mráz*
673 * Case insensitive string comparison no longer uses locales. It has instead
674 been directly implemented.
678 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
680 * Case insensitive string comparison is reimplemented via new locale-agnostic
681 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
682 comparison. The previous implementation had problems when the Turkish locale
687 * Fixed a bug in the c_rehash script which was not properly sanitising shell
688 metacharacters to prevent command injection. This script is distributed by
689 some operating systems in a manner where it is automatically executed. On
690 such operating systems, an attacker could execute arbitrary commands with the
691 privileges of the script.
693 Use of the c_rehash script is considered obsolete and should be replaced
694 by the OpenSSL rehash command line tool.
699 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
700 certificate on an OCSP response. The bug caused the function in the case
701 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
702 response (meaning a successful verification) even in the case where the
703 response signing certificate fails to verify.
705 It is anticipated that most users of `OCSP_basic_verify` will not use the
706 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
707 a negative value (indicating a fatal error) in the case of a certificate
708 verification failure. The normal expected return value in this case would be
711 This issue also impacts the command line OpenSSL "ocsp" application. When
712 verifying an ocsp response with the "-no_cert_checks" option the command line
713 application will report that the verification is successful even though it
714 has in fact failed. In this case the incorrect successful response will also
715 be accompanied by error messages showing the failure and contradicting the
716 apparently successful result.
721 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
722 AAD data as the MAC key. This made the MAC key trivially predictable.
724 An attacker could exploit this issue by performing a man-in-the-middle attack
725 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
726 that the modified data would still pass the MAC integrity check.
728 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
729 endpoint will always be rejected by the recipient and the connection will
730 fail at that point. Many application protocols require data to be sent from
731 the client to the server first. Therefore, in such a case, only an OpenSSL
732 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
734 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
735 sent in both directions. In this case both clients and servers could be
736 affected, regardless of the application protocol.
738 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
739 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
740 the handshake when using this ciphersuite.
742 The confidentiality of data is not impacted by this issue, i.e. an attacker
743 cannot decrypt data that has been encrypted using this ciphersuite - they can
746 In order for this attack to work both endpoints must legitimately negotiate
747 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
748 OpenSSL 3.0, and is not available within the default provider or the default
749 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
750 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
751 following must have occurred:
753 1) OpenSSL must have been compiled with the (non-default) compile time option
754 enable-weak-ssl-ciphers
756 2) OpenSSL must have had the legacy provider explicitly loaded (either
757 through application code or via configuration)
759 3) The ciphersuite must have been explicitly added to the ciphersuite list
761 4) The libssl security level must have been set to 0 (default is 1)
763 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
765 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
766 others that both endpoints have in common
771 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
772 occupied by the removed hash table entries.
774 This function is used when decoding certificates or keys. If a long lived
775 process periodically decodes certificates or keys its memory usage will
776 expand without bounds and the process might be terminated by the operating
777 system causing a denial of service. Also traversing the empty hash table
778 entries will take increasingly more time.
780 Typically such long lived processes might be TLS clients or TLS servers
781 configured to accept client certificate authentication.
784 *Hugo Landau, Aliaksei Levin*
786 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
787 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
788 statistics are no longer supported. For compatibility, these statistics are
789 still listed in the output but are now always reported as zero.
793 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
795 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
796 for non-prime moduli.
798 Internally this function is used when parsing certificates that contain
799 elliptic curve public keys in compressed form or explicit elliptic curve
800 parameters with a base point encoded in compressed form.
802 It is possible to trigger the infinite loop by crafting a certificate that
803 has invalid explicit curve parameters.
805 Since certificate parsing happens prior to verification of the certificate
806 signature, any process that parses an externally supplied certificate may thus
807 be subject to a denial of service attack. The infinite loop can also be
808 reached when parsing crafted private keys as they can contain explicit
809 elliptic curve parameters.
811 Thus vulnerable situations include:
813 - TLS clients consuming server certificates
814 - TLS servers consuming client certificates
815 - Hosting providers taking certificates or private keys from customers
816 - Certificate authorities parsing certification requests from subscribers
817 - Anything else which parses ASN.1 elliptic curve parameters
819 Also any other applications that use the BN_mod_sqrt() where the attacker
820 can control the parameter values are vulnerable to this DoS issue.
825 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
826 to the list of ciphersuites providing Perfect Forward Secrecy as
827 required by SECLEVEL >= 3.
829 *Dmitry Belyavskiy, Nicola Tuveri*
831 * Made the AES constant time code for no-asm configurations
832 optional due to the resulting 95% performance degradation.
833 The AES constant time code can be enabled, for no assembly
834 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
838 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
843 * The negative return value handling of the certificate verification callback
844 was reverted. The replacement is to set the verification retry state with
845 the SSL_set_retry_verify() function.
849 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
851 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
852 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
853 verify a certificate supplied by a server. That function may return a
854 negative return value to indicate an internal error (for example out of
855 memory). Such a negative return value is mishandled by OpenSSL and will cause
856 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
857 success and a subsequent call to SSL_get_error() to return the value
858 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
859 returned by OpenSSL if the application has previously called
860 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
861 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
862 totally unexpected and applications may not behave correctly as a result. The
863 exact behaviour will depend on the application but it could result in
864 crashes, infinite loops or other similar incorrect responses.
866 This issue is made more serious in combination with a separate bug in OpenSSL
867 3.0 that will cause X509_verify_cert() to indicate an internal error when
868 processing a certificate chain. This will occur where a certificate does not
869 include the Subject Alternative Name extension but where a Certificate
870 Authority has enforced name constraints. This issue can occur even with valid
876 * Corrected a few file name and file reference bugs in the build,
877 installation and setup scripts, which lead to installation verification
878 failures. Slightly enhanced the installation verification script.
882 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
887 * Fixed PVK encoder to properly query for the passphrase.
891 * Multiple fixes in the OSSL_HTTP API functions.
895 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
896 OSSL_PARAM_INTEGER data type and return error on negative numbers
897 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
898 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
902 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
906 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
910 * Multiple threading fixes.
914 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
918 * Allow fetching an operation from the provider that owns an unexportable key
919 as a fallback if that is still allowed by the property query.
923 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
925 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
930 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
931 S390X capability vector to zero. This simplifies testing of different code
932 paths on S390X architecture.
936 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
937 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
938 SP 800-38D". The communication will fail at this point.
942 * The EC_GROUP_clear_free() function is deprecated as there is nothing
943 confidential in EC_GROUP data.
947 * The byte order mark (BOM) character is ignored if encountered at the
948 beginning of a PEM-formatted file.
952 * Added CMS support for the Russian GOST algorithms.
956 * Due to move of the implementation of cryptographic operations
957 to the providers, validation of various operation parameters can
958 be postponed until the actual operation is executed where previously
959 it happened immediately when an operation parameter was set.
961 For example when setting an unsupported curve with
962 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
963 fail but later keygen operations with the EVP_PKEY_CTX will fail.
965 *OpenSSL team members and many third party contributors*
967 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
968 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
969 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
970 instead to retrieve these algorithms from a provider.
974 * On build targets where the multilib postfix is set in the build
975 configuration the libdir directory was changing based on whether
976 the lib directory with the multilib postfix exists on the system
977 or not. This unpredictable behavior was removed and eventual
978 multilib postfix is now always added to the default libdir. Use
979 `--libdir=lib` to override the libdir if adding the postfix is
984 * The triple DES key wrap functionality now conforms to RFC 3217 but is
985 no longer interoperable with OpenSSL 1.1.1.
989 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
990 function codes, this function can only cause problems for calling
995 * Add a configurable flag to output date formats as ISO 8601. Does not
996 change the default date format.
1000 * Version of MSVC earlier than 1300 could get link warnings, which could
1001 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1002 Support for this flag has been removed.
1006 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1007 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1008 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1009 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1010 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1014 * The signatures of the functions to get and set options on SSL and
1015 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1016 Some source code changes may be required.
1020 * The public definitions of conf_method_st and conf_st have been
1021 deprecated. They will be made opaque in a future release.
1023 *Rich Salz and Tomáš Mráz*
1025 * Client-initiated renegotiation is disabled by default. To allow it, use
1026 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1027 flag, or the "ClientRenegotiation" config parameter as appropriate.
1031 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1032 or modify relative pathname inclusion.
1036 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1037 validated. Please consult the README-FIPS and
1038 README-PROVIDERS files, as well as the migration guide.
1040 *OpenSSL team members and many third party contributors*
1042 * For the key types DH and DHX the allowed settable parameters are now different.
1046 * The openssl commands that read keys, certificates, and CRLs now
1047 automatically detect the PEM or DER format of the input files.
1049 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1051 * Added enhanced PKCS#12 APIs which accept a library context.
1055 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1059 * Added support for Kernel TLS (KTLS).
1061 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1063 * Support for RFC 5746 secure renegotiation is now required by default for
1064 SSL or TLS connections to succeed.
1068 * The signature of the `copy` functional parameter of the
1069 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1070 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1071 the signature of the `pub_decode` functional parameter of the
1072 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1073 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1077 * The error return values from some control calls (ctrl) have changed.
1081 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1085 * Many functions in the EVP_ namespace that are getters of values from
1086 implementations or contexts were renamed to include get or get0 in their
1087 names. Old names are provided as macro aliases for compatibility and
1092 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1093 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1094 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1099 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1102 * The output from the command line applications may have minor
1107 * The output from numerous "printing" may have minor changes.
1111 * Windows thread synchronization uses read/write primitives (SRWLock) when
1112 supported by the OS, otherwise CriticalSection continues to be used.
1116 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1117 work on read only BIO source/sinks that do not support these functions.
1118 This allows piping or redirection of a file BIO using stdin to be buffered
1119 into memory. This is used internally in OSSL_DECODER_from_bio().
1123 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1124 this function would return one of the values OSSL_STORE_INFO_NAME,
1125 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1126 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1127 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1128 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1129 using this function should be amended to handle the changed return value.
1133 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1134 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1135 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1136 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1137 contains more than one certificate identifier: This means that all
1138 certificates referenced there MUST be part of the validation chain.
1142 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1143 RC5, DESX and DES have been moved to the legacy provider.
1147 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1148 RIPEMD-160 have been moved to the legacy provider.
1152 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1157 * The deprecated functions EVP_PKEY_get0_RSA(),
1158 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1159 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1160 well as the similarly named "get1" functions behave differently in
1165 * A number of functions handling low-level keys or engines were deprecated
1166 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1167 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1168 EVP_PKEY_get0_siphash().
1172 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1173 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1174 will need to load the legacy crypto provider. This includes these PBE
1175 algorithms which use this KDF:
1176 - NID_pbeWithMD2AndDES_CBC
1177 - NID_pbeWithMD5AndDES_CBC
1178 - NID_pbeWithSHA1AndRC2_CBC
1179 - NID_pbeWithMD2AndRC2_CBC
1180 - NID_pbeWithMD5AndRC2_CBC
1181 - NID_pbeWithSHA1AndDES_CBC
1185 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1186 BIO_debug_callback() functions.
1190 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1191 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1195 * The RAND_METHOD APIs have been deprecated.
1199 * The SRP APIs have been deprecated.
1203 * Add a compile time option to prevent the caching of provider fetched
1204 algorithms. This is enabled by including the no-cached-fetch option
1205 at configuration time.
1209 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1210 count of PKCS12_DEFAULT_ITER.
1212 *Tomáš Mráz and Sahana Prasad*
1214 * The openssl speed command does not use low-level API calls anymore.
1218 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1221 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1223 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1227 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1228 providers may supply their own group implementations (using either the "key
1229 exchange" or the "key encapsulation" methods) which will automatically be
1230 detected and used by libssl.
1232 *Matt Caswell, Nicola Tuveri*
1234 * The undocumented function X509_certificate_type() has been deprecated;
1238 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1242 * Removed RSA padding mode for SSLv23 (which was only used for
1243 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1244 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1249 * Deprecated the obsolete X9.31 RSA key generation related functions.
1251 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1252 is not allowed to return a value > 1, this is no more taken as failure.
1254 *Viktor Dukhovni and David von Oheimb*
1256 * Deprecated the obsolete X9.31 RSA key generation related functions
1257 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1258 BN_X931_generate_prime_ex().
1262 * The default key generation method for the regular 2-prime RSA keys was
1263 changed to the FIPS 186-4 B.3.6 method.
1267 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1271 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1275 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1276 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1278 *Rich Salz, Richard Levitte, and David von Oheimb*
1280 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1284 * Deprecated `OCSP_parse_url()`.
1288 * Validation of SM2 keys has been separated from the validation of regular EC
1293 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1294 switches: a validation failure triggers an early exit, returning a failure
1295 exit status to the parent process.
1299 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1300 to ignore unknown ciphers.
1304 * The `-cipher-commands` and `-digest-commands` options
1305 of the command line utility `list` have been deprecated.
1306 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1310 * Added convenience functions for generating asymmetric key pairs:
1311 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1312 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1316 * All of the low-level EC_KEY functions have been deprecated.
1318 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1320 * Deprecated all the libcrypto and libssl error string loading
1325 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1326 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1331 * The `-crypt` option to the `passwd` command line tool has been removed.
1335 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1340 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1344 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1345 EVP_PKEY_get1_tls_encodedpoint().
1349 * The security callback, which can be customised by application code, supports
1350 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1351 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1355 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1356 interface. Their functionality remains unchanged.
1360 * Added new option for 'openssl list', '-providers', which will display the
1361 list of loaded providers, their names, version and status. It optionally
1362 displays their gettable parameters.
1366 * Removed EVP_PKEY_set_alias_type().
1370 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1371 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1375 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1376 parameters are still checked for type safety at compile time via helper
1381 * Remove the RAND_DRBG API
1383 *Paul Dale and Matthias St. Pierre*
1385 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1386 as well as actual hostnames.
1390 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1391 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1392 conversely, silently ignore DTLS protocol version bounds when configuring
1393 TLS-based contexts. The commands can be repeated to set bounds of both
1394 types. The same applies with the corresponding "min_protocol" and
1395 "max_protocol" command-line switches, in case some application uses both TLS
1398 SSL_CTX instances that are created for a fixed protocol version (e.g.
1399 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1400 attempts to apply bounds to these protocol versions would result in an
1401 error. Now only the "version-flexible" SSL_CTX instances are subject to
1402 limits in configuration files in command-line options.
1406 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1411 * Reworked the recorded ERR codes to make better space for system errors.
1412 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1413 given code is a system error (true) or an OpenSSL error (false).
1417 * Reworked the test perl framework to better allow parallel testing.
1419 *Nicola Tuveri and David von Oheimb*
1421 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1422 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1426 * 'Configure' has been changed to figure out the configuration target if
1427 none is given on the command line. Consequently, the 'config' script is
1428 now only a mere wrapper. All documentation is changed to only mention
1431 *Rich Salz and Richard Levitte*
1433 * Added a library context `OSSL_LIB_CTX` that applications as well as
1434 other libraries can use to form a separate context within which
1435 libcrypto operations are performed.
1439 * Added various `_ex` functions to the OpenSSL API that support using
1440 a non-default `OSSL_LIB_CTX`.
1444 * Handshake now fails if Extended Master Secret extension is dropped
1449 * Dropped interactive mode from the `openssl` program.
1453 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1455 *David von Oheimb and Shane Lontis*
1457 * Deprecated `EC_METHOD_get_field_type()`.
1461 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1462 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1463 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1467 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1471 * Add CAdES-BES signature verification support, mostly derived
1472 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1474 *Filipe Raimundo da Silva*
1476 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1480 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1481 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1485 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1489 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1490 EC_KEY_precompute_mult().
1494 * Deprecated EC_POINTs_mul().
1498 * Removed FIPS_mode() and FIPS_mode_set().
1502 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1506 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1507 EC_POINT_get_Jprojective_coordinates_GFp().
1511 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1512 arrays to be more easily constructed via a series of utility functions.
1513 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1514 the various push functions and finally convert to a passable OSSL_PARAM
1515 array using OSSL_PARAM_BLD_to_param().
1519 * The security strength of SHA1 and MD5 based signatures in TLS has been
1524 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1525 contain a provider side internal key.
1529 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1533 * Project text documents not yet having a proper file name extension
1534 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1535 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1539 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1540 have been converted to Markdown with the goal to produce documents
1541 which not only look pretty when viewed online in the browser, but
1542 remain well readable inside a plain text editor.
1544 To achieve this goal, a 'minimalistic' Markdown style has been applied
1545 which avoids formatting elements that interfere too much with the
1546 reading flow in the text file. For example, it
1548 * avoids [ATX headings][] and uses [setext headings][] instead
1549 (which works for `<h1>` and `<h2>` headings only).
1550 * avoids [inline links][] and uses [reference links][] instead.
1551 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1553 [ATX headings]: https://github.github.com/gfm/#atx-headings
1554 [setext headings]: https://github.github.com/gfm/#setext-headings
1555 [inline links]: https://github.github.com/gfm/#inline-link
1556 [reference links]: https://github.github.com/gfm/#reference-link
1557 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1558 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1560 *Matthias St. Pierre*
1562 * The test suite is changed to preserve results of each test recipe.
1563 A new directory test-runs/ with subdirectories named like the
1564 test recipes are created in the build tree for this purpose.
1568 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1569 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1570 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1572 *David von Oheimb, Martin Peylo*
1574 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1575 It supports arbitrary request and response content types, GET redirection,
1576 TLS, connections via HTTP(S) proxies, connections and exchange via
1577 user-defined BIOs (allowing implicit connections), persistent connections,
1578 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1579 The legacy OCSP-focused (and only partly documented) API
1580 is retained for backward compatibility, while most of it is deprecated.
1584 * Added `util/check-format.pl`, a tool for checking adherence to the
1585 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1586 The checks performed are incomplete and yield some false positives.
1587 Still the tool should be useful for detecting most typical glitches.
1591 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1592 If domain name resolution yields multiple IP addresses all of them are tried
1593 after `connect()` failures.
1597 * All of the low-level RSA functions have been deprecated.
1601 * X509 certificates signed using SHA1 are no longer allowed at security
1606 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1607 modified to use PKEY APIs. These commands are now in maintenance mode
1608 and no new features will be added to them.
1612 * The command line utility rsautl has been deprecated.
1616 * The command line utilities genrsa and rsa have been modified to use PKEY
1617 APIs. They now write PKCS#8 keys by default. These commands are now in
1618 maintenance mode and no new features will be added to them.
1622 * All of the low-level DH functions have been deprecated.
1624 *Paul Dale and Matt Caswell*
1626 * All of the low-level DSA functions have been deprecated.
1630 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1631 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1635 * Deprecated low-level ECDH and ECDSA functions.
1639 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1643 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1644 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1645 a new formulation to include all the things it can be used for,
1646 as well as words of caution.
1650 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1654 * All of the low-level HMAC functions have been deprecated.
1656 *Paul Dale and David von Oheimb*
1658 * Over two thousand fixes were made to the documentation, including:
1659 - Common options (such as -rand/-writerand, TLS version control, etc)
1660 were refactored and point to newly-enhanced descriptions in openssl.pod.
1661 - Added style conformance for all options (with help from Richard Levitte),
1662 documented all reported missing options, added a CI build to check
1663 that all options are documented and that no unimplemented options
1665 - Documented some internals, such as all use of environment variables.
1666 - Addressed all internal broken L<> references.
1670 * All of the low-level CMAC functions have been deprecated.
1674 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1675 functions have been deprecated.
1677 *Paul Dale and David von Oheimb*
1679 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1680 set of functions. The documentation mentioned negative values for some
1681 errors, but this was never the case, so the mention of negative values
1684 Code that followed the documentation and thereby check with something
1685 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1689 * All of the low-level cipher functions have been deprecated.
1691 *Matt Caswell and Paul Dale*
1693 * Removed include/openssl/opensslconf.h.in and replaced it with
1694 include/openssl/configuration.h.in, which differs in not including
1695 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1696 was added to include both.
1698 This allows internal hacks where one might need to modify the set
1699 of configured macros, for example this if deprecated symbols are
1700 still supposed to be available internally:
1702 #include <openssl/configuration.h>
1704 #undef OPENSSL_NO_DEPRECATED
1705 #define OPENSSL_SUPPRESS_DEPRECATED
1707 #include <openssl/macros.h>
1709 This should not be used by applications that use the exported
1710 symbols, as that will lead to linking errors.
1714 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1715 used in exponentiation with 512-bit moduli. No EC algorithms are
1716 affected. Analysis suggests that attacks against 2-prime RSA1024,
1717 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1718 difficult to perform and are not believed likely. Attacks against DH512
1719 are considered just feasible. However, for an attack the target would
1720 have to re-use the DH512 private key, which is not recommended anyway.
1721 Also applications directly using the low-level API BN_mod_exp may be
1722 affected if they use BN_FLG_CONSTTIME.
1727 * Most memory-debug features have been deprecated, and the functionality
1728 replaced with no-ops.
1732 * Added documentation for the STACK API.
1736 * Introduced a new method type and API, OSSL_ENCODER, to represent
1737 generic encoders. These do the same sort of job that PEM writers
1738 and d2i functions do, but with support for methods supplied by
1739 providers, and the possibility for providers to support other
1744 * Introduced a new method type and API, OSSL_DECODER, to represent
1745 generic decoders. These do the same sort of job that PEM readers
1746 and i2d functions do, but with support for methods supplied by
1747 providers, and the possibility for providers to support other
1752 * Added a .pragma directive to the syntax of configuration files, to
1753 allow varying behavior in a supported and predictable manner.
1754 Currently added pragma:
1758 This allows dollar signs to be a keyword character unless it's
1759 followed by a opening brace or parenthesis. This is useful for
1760 platforms where dollar signs are commonly used in names, such as
1761 volume names and system directory names on VMS.
1765 * Added functionality to create an EVP_PKEY from user data.
1769 * Change the interpretation of the '--api' configuration option to
1770 mean that this is a desired API compatibility level with no
1771 further meaning. The previous interpretation, that this would
1772 also mean to remove all deprecated symbols up to and including
1773 the given version, no requires that 'no-deprecated' is also used
1774 in the configuration.
1776 When building applications, the desired API compatibility level
1777 can be set with the OPENSSL_API_COMPAT macro like before. For
1778 API compatibility version below 3.0, the old style numerical
1779 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1780 For version 3.0 and on, the value is expected to be the decimal
1781 value calculated from the major and minor version like this:
1783 MAJOR * 10000 + MINOR * 100
1787 -DOPENSSL_API_COMPAT=30000 For 3.0
1788 -DOPENSSL_API_COMPAT=30200 For 3.2
1790 To hide declarations that are deprecated up to and including the
1791 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1792 given when building the application as well.
1796 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1797 access to certificate and CRL stores via URIs and OSSL_STORE
1800 This adds the following functions:
1802 - X509_LOOKUP_store()
1803 - X509_STORE_load_file()
1804 - X509_STORE_load_path()
1805 - X509_STORE_load_store()
1806 - SSL_add_store_cert_subjects_to_stack()
1807 - SSL_CTX_set_default_verify_store()
1808 - SSL_CTX_load_verify_file()
1809 - SSL_CTX_load_verify_dir()
1810 - SSL_CTX_load_verify_store()
1814 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1815 The presence of this system service is determined at run-time.
1819 * Added functionality to create an EVP_PKEY context based on data
1820 for methods from providers. This takes an algorithm name and a
1821 property query string and simply stores them, with the intent
1822 that any operation that uses this context will use those strings
1823 to fetch the needed methods implicitly, thereby making the port
1824 of application written for pre-3.0 OpenSSL easier.
1828 * The undocumented function NCONF_WIN32() has been deprecated; for
1829 conversion details see the HISTORY section of doc/man5/config.pod
1833 * Introduced the new functions EVP_DigestSignInit_ex() and
1834 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1835 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1836 pages for further details.
1840 * Over two thousand fixes were made to the documentation, including:
1841 adding missing command flags, better style conformance, documentation
1844 *Rich Salz, Richard Levitte*
1846 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1847 X25519, X448, Ed25519 and Ed448.
1851 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1856 * Deprecated the public definition of `ERR_STATE` as well as the function
1857 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1862 * Added ERR functionality to give callers access to the stored function
1863 names that have replaced the older function code based functions.
1865 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1866 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1867 ERR_peek_error_all() and ERR_peek_last_error_all().
1869 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1870 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1871 ERR_func_error_string().
1875 * Extended testing to be verbose for failing tests only. The make variables
1876 VERBOSE_FAILURE or VF can be used to enable this:
1878 $ make VF=1 test # Unix
1879 $ mms /macro=(VF=1) test ! OpenVMS
1880 $ nmake VF=1 test # Windows
1884 * Added the `-copy_extensions` option to the `x509` command for use with
1885 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1886 all extensions in the request are copied to the certificate or vice versa.
1888 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1890 * Added the `-copy_extensions` option to the `req` command for use with
1891 `-x509`. When given with the `copy` or `copyall` argument,
1892 all extensions in the certification request are copied to the certificate.
1896 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1897 they generate are by default RFC 5280 compliant in the following sense:
1898 There is a subjectKeyIdentifier extension with a hash value of the public key
1899 and for not self-signed certs there is an authorityKeyIdentifier extension
1900 with a keyIdentifier field or issuer information identifying the signing key.
1901 This is done unless some configuration overrides the new default behavior,
1902 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1906 * Added several checks to `X509_verify_cert()` according to requirements in
1907 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1908 (which may be done by using the CLI option `-x509_strict`):
1909 * The basicConstraints of CA certificates must be marked critical.
1910 * CA certificates must explicitly include the keyUsage extension.
1911 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1912 * The issuer name of any certificate must not be empty.
1913 * The subject name of CA certs, certs with keyUsage crlSign,
1914 and certs without subjectAlternativeName must not be empty.
1915 * If a subjectAlternativeName extension is given it must not be empty.
1916 * The signatureAlgorithm field and the cert signature must be consistent.
1917 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1918 must not be marked critical.
1919 * The authorityKeyIdentifier must be given for X.509v3 certs
1920 unless they are self-signed.
1921 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1925 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1926 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1930 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1931 used even when parsing explicit parameters, when loading a encoded key
1932 or calling `EC_GROUP_new_from_ecpkparameters()`/
1933 `EC_GROUP_new_from_ecparameters()`.
1934 This prevents bypass of security hardening and performance gains,
1935 especially for curves with specialized EC_METHODs.
1936 By default, if a key encoded with explicit parameters is loaded and later
1937 encoded, the output is still encoded with explicit parameters, even if
1938 internally a "named" EC_GROUP is used for computation.
1942 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1943 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1944 NULL. After this change, only the cofactor parameter can be NULL. It also
1945 does some minimal sanity checks on the passed order.
1950 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1951 An attack is simple, if the first CMS_recipientInfo is valid but the
1952 second CMS_recipientInfo is chosen ciphertext. If the second
1953 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1954 encryption key will be replaced by garbage, and the message cannot be
1955 decoded, but if the RSA decryption fails, the correct encryption key is
1956 used and the recipient will not notice the attack.
1957 As a work around for this potential attack the length of the decrypted
1958 key must be equal to the cipher default key length, in case the
1959 certificate is not given and all recipientInfo are tried out.
1960 The old behaviour can be re-enabled in the CMS code by setting the
1961 CMS_DEBUG_DECRYPT flag.
1965 * Early start up entropy quality from the DEVRANDOM seed source has been
1966 improved for older Linux systems. The RAND subsystem will wait for
1967 /dev/random to be producing output before seeding from /dev/urandom.
1968 The seeded state is stored for future library initialisations using
1969 a system global shared memory segment. The shared memory identifier
1970 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1971 the desired value. The default identifier is 114.
1975 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1976 when primes for RSA keys are computed.
1977 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1978 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1979 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
1980 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1981 This avoids possible fingerprinting of newly generated RSA modules.
1985 * Correct the extended master secret constant on EBCDIC systems. Without this
1986 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1987 negotiate EMS will fail. Unfortunately this also means that TLS connections
1988 between EBCDIC systems with this fix, and EBCDIC systems without this
1989 fix will fail if they negotiate EMS.
1993 * Changed the library initialisation so that the config file is now loaded
1994 by default. This was already the case for libssl. It now occurs for both
1995 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1996 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2000 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2001 where the former acts as a replacement for `ERR_put_error()`, and the
2002 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2003 `ERR_raise_data()` adds more flexibility by taking a format string and
2004 an arbitrary number of arguments following it, to be processed with
2009 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2010 to check if a named provider is loaded and available. When called, it
2011 will also activate all fallback providers if such are still present.
2015 * Enforce a minimum DH modulus size of 512 bits.
2019 * Changed DH parameters to generate the order q subgroup instead of 2q.
2020 Previously generated DH parameters are still accepted by DH_check
2021 but DH_generate_key works around that by clearing bit 0 of the
2022 private key for those. This avoids leaking bit 0 of the private key.
2026 * Significantly reduce secure memory usage by the randomness pools.
2030 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2035 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2036 algorithms. An implementation of a key exchange algorithm can be obtained
2037 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2038 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2039 the older EVP_PKEY_derive_init() function. See the man pages for the new
2040 functions for further details.
2044 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2048 * Removed the function names from error messages and deprecated the
2053 * Removed NextStep support and the macro OPENSSL_UNISTD
2057 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2058 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2059 Also removed "export var as function" capability; we do not export
2060 variables, only functions.
2064 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2065 an error and 1 indicating success. In previous versions of OpenSSL this
2066 was a void type. If a key was set longer than the maximum possible this
2071 * Support SM2 signing and verification schemes with X509 certificate.
2075 * Use SHA256 as the default digest for TS query in the `ts` app.
2079 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2083 * Default cipher lists/suites are now available via a function, the
2084 #defines are deprecated.
2088 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2089 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2090 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2094 * Join the directories crypto/x509 and crypto/x509v3
2098 * Added command 'openssl kdf' that uses the EVP_KDF API.
2102 * Added command 'openssl mac' that uses the EVP_MAC API.
2106 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2107 as default directories. Also added the command 'openssl info'
2108 for scripting purposes.
2112 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2117 * Add prediction resistance to the DRBG reseeding process.
2121 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2122 mandated by IEEE Std 1619-2018.
2126 * Added newline escaping functionality to a filename when using openssl dgst.
2127 This output format is to replicate the output format found in the `*sum`
2128 checksum programs. This aims to preserve backward compatibility.
2130 *Matt Eaton, Richard Levitte, and Paul Dale*
2132 * Removed the heartbeat message in DTLS feature, as it has very
2133 little usage and doesn't seem to fulfill a valuable purpose.
2134 The configuration option is now deprecated.
2138 * Changed the output of 'openssl {digestname} < file' to display the
2139 digest name in its output.
2143 * Added a new generic trace API which provides support for enabling
2144 instrumentation through trace output.
2146 *Richard Levitte & Matthias St. Pierre*
2148 * Added build tests for C++. These are generated files that only do one
2149 thing, to include one public OpenSSL head file each. This tests that
2150 the public header files can be usefully included in a C++ application.
2152 This test isn't enabled by default. It can be enabled with the option
2153 'enable-buildtest-c++'.
2157 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2161 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2165 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2169 * Added KMAC to EVP_MAC.
2173 * Added property based algorithm implementation selection framework to
2178 * Added SCA hardening for modular field inversion in EC_GROUP through
2179 a new dedicated field_inv() pointer in EC_METHOD.
2180 This also addresses a leakage affecting conversions from projective
2181 to affine coordinates.
2183 *Billy Bob Brumley, Nicola Tuveri*
2185 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2186 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2187 those algorithms that were already supported through the EVP_PKEY API
2188 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2189 and scrypt are now wrappers that call EVP_KDF.
2193 * Build devcrypto engine as a dynamic engine.
2195 *Eneas U de Queiroz*
2197 * Add keyed BLAKE2 to EVP_MAC.
2201 * Fix a bug in the computation of the endpoint-pair shared secret used
2202 by DTLS over SCTP. This breaks interoperability with older versions
2203 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2204 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2205 interoperability with such broken implementations. However, enabling
2206 this switch breaks interoperability with correct implementations.
2208 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2209 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2213 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2217 * Changed the license to the Apache License v2.0.
2221 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2223 - Major releases (indicated by incrementing the MAJOR release number)
2224 may introduce incompatible API/ABI changes.
2225 - Minor releases (indicated by incrementing the MINOR release number)
2226 may introduce new features but retain API/ABI compatibility.
2227 - Patch releases (indicated by incrementing the PATCH number)
2228 are intended for bug fixes and other improvements of existing
2229 features only (like improving performance or adding documentation)
2230 and retain API/ABI compatibility.
2234 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2238 * Remove the 'dist' target and add a tarball building script. The
2239 'dist' target has fallen out of use, and it shouldn't be
2240 necessary to configure just to create a source distribution.
2244 * Recreate the OS390-Unix config target. It no longer relies on a
2245 special script like it did for OpenSSL pre-1.1.0.
2249 * Instead of having the source directories listed in Configure, add
2250 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2255 * Add GMAC to EVP_MAC.
2259 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2263 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2264 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2265 to facilitate the continued use of MACs through raw private keys in
2266 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2270 * Deprecate ECDH_KDF_X9_62().
2274 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2275 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2276 are retained for backwards compatibility.
2280 * AES-XTS mode now enforces that its two keys are different to mitigate
2281 the attacked described in "Efficient Instantiations of Tweakable
2282 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2283 Details of this attack can be obtained from:
2284 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2288 * Rename the object files, i.e. give them other names than in previous
2289 versions. Their names now include the name of the final product, as
2290 well as its type mnemonic (bin, lib, shlib).
2294 * Added new option for 'openssl list', '-objects', which will display the
2295 list of built in objects, i.e. OIDs with names.
2299 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2300 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2305 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2306 improves application performance by removing data copies and providing
2307 applications with zero-copy system calls such as sendfile and splice.
2311 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2315 * `PKCS12_parse` now maintains the order of the parsed certificates
2316 when outputting them via `*ca` (rather than reversing it).
2320 * Deprecated pthread fork support methods.
2324 * Added support for FFDHE key exchange in TLS 1.3.
2328 * Added a new concept for OpenSSL plugability: providers. This
2329 functionality is designed to replace the ENGINE API and ENGINE
2330 implementations, and to be much more dynamic, allowing provider
2331 authors to introduce new algorithms among other things, as long as
2332 there's an API that supports the algorithm type.
2334 With this concept comes a new core API for interaction between
2335 libcrypto and provider implementations. Public libcrypto functions
2336 that want to use providers do so through this core API.
2338 The main documentation for this core API is found in
2339 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2340 refer to other manuals describing the API specific for supported
2341 algorithm types (also called operations).
2348 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2350 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2352 * Avoid loading of a dynamic engine twice.
2356 * Prioritise DANE TLSA issuer certs over peer certs
2360 * Fixed random API for MacOS prior to 10.12
2362 These MacOS versions don't support the CommonCrypto APIs
2366 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2368 * Fixed an SM2 Decryption Buffer Overflow.
2370 In order to decrypt SM2 encrypted data an application is expected to
2371 call the API function EVP_PKEY_decrypt(). Typically an application will
2372 call this function twice. The first time, on entry, the "out" parameter
2373 can be NULL and, on exit, the "outlen" parameter is populated with the
2374 buffer size required to hold the decrypted plaintext. The application
2375 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2376 again, but this time passing a non-NULL value for the "out" parameter.
2378 A bug in the implementation of the SM2 decryption code means that the
2379 calculation of the buffer size required to hold the plaintext returned
2380 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2381 size required by the second call. This can lead to a buffer overflow
2382 when EVP_PKEY_decrypt() is called by the application a second time with
2383 a buffer that is too small.
2385 A malicious attacker who is able present SM2 content for decryption to
2386 an application could cause attacker chosen data to overflow the buffer
2387 by up to a maximum of 62 bytes altering the contents of other data held
2388 after the buffer, possibly changing application behaviour or causing
2389 the application to crash. The location of the buffer is application
2390 dependent but is typically heap allocated.
2395 * Fixed various read buffer overruns processing ASN.1 strings
2397 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2398 structure which contains a buffer holding the string data and a field
2399 holding the buffer length. This contrasts with normal C strings which
2400 are represented as a buffer for the string data which is terminated
2401 with a NUL (0) byte.
2403 Although not a strict requirement, ASN.1 strings that are parsed using
2404 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2405 well as any string whose value has been set with the ASN1_STRING_set()
2406 function will additionally NUL terminate the byte array in the
2407 ASN1_STRING structure.
2409 However, it is possible for applications to directly construct valid
2410 ASN1_STRING structures which do not NUL terminate the byte array by
2411 directly setting the "data" and "length" fields in the ASN1_STRING
2412 array. This can also happen by using the ASN1_STRING_set0() function.
2414 Numerous OpenSSL functions that print ASN.1 data have been found to
2415 assume that the ASN1_STRING byte array will be NUL terminated, even
2416 though this is not guaranteed for strings that have been directly
2417 constructed. Where an application requests an ASN.1 structure to be
2418 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2419 been directly constructed by the application without NUL terminating
2420 the "data" field, then a read buffer overrun can occur.
2422 The same thing can also occur during name constraints processing
2423 of certificates (for example if a certificate has been directly
2424 constructed by the application instead of loading it via the OpenSSL
2425 parsing functions, and the certificate contains non NUL terminated
2426 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2427 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2429 If a malicious actor can cause an application to directly construct an
2430 ASN1_STRING and then process it through one of the affected OpenSSL
2431 functions then this issue could be hit. This might result in a crash
2432 (causing a Denial of Service attack). It could also result in the
2433 disclosure of private memory contents (such as private keys, or
2434 sensitive plaintext).
2439 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2441 * Fixed a problem with verifying a certificate chain when using the
2442 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2443 the certificates present in a certificate chain. It is not set by default.
2445 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2446 the chain that have explicitly encoded elliptic curve parameters was added
2447 as an additional strict check.
2449 An error in the implementation of this check meant that the result of a
2450 previous check to confirm that certificates in the chain are valid CA
2451 certificates was overwritten. This effectively bypasses the check
2452 that non-CA certificates must not be able to issue other certificates.
2454 If a "purpose" has been configured then there is a subsequent opportunity
2455 for checks that the certificate is a valid CA. All of the named "purpose"
2456 values implemented in libcrypto perform this check. Therefore, where
2457 a purpose is set the certificate chain will still be rejected even when the
2458 strict flag has been used. A purpose is set by default in libssl client and
2459 server certificate verification routines, but it can be overridden or
2460 removed by an application.
2462 In order to be affected, an application must explicitly set the
2463 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2464 for the certificate verification or, in the case of TLS client or server
2465 applications, override the default purpose.
2470 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2471 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2472 renegotiation ClientHello omits the signature_algorithms extension (where it
2473 was present in the initial ClientHello), but includes a
2474 signature_algorithms_cert extension then a NULL pointer dereference will
2475 result, leading to a crash and a denial of service attack.
2477 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2478 (which is the default configuration). OpenSSL TLS clients are not impacted by
2482 *Peter Kästle and Samuel Sapalski*
2484 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2486 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2487 create a unique hash value based on the issuer and serial number data
2488 contained within an X509 certificate. However, it was failing to correctly
2489 handle any errors that may occur while parsing the issuer field (which might
2490 occur if the issuer field is maliciously constructed). This may subsequently
2491 result in a NULL pointer deref and a crash leading to a potential denial of
2497 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2498 padding mode to correctly check for rollback attacks. This is considered a
2499 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2504 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2505 functions. Previously they could overflow the output length argument in some
2506 cases where the input length is close to the maximum permissible length for
2507 an integer on the platform. In such cases the return value from the function
2508 call would be 1 (indicating success), but the output length value would be
2509 negative. This could cause applications to behave incorrectly or crash.
2514 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2515 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2516 could be exploited in a side channel attack to recover the password. Since
2517 the attack is local host only this is outside of the current OpenSSL
2518 threat model and therefore no CVE is assigned.
2520 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2525 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2527 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2528 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2529 If an attacker can control both items being compared then this could lead
2530 to a possible denial of service attack. OpenSSL itself uses the
2531 GENERAL_NAME_cmp function for two purposes:
2532 1) Comparing CRL distribution point names between an available CRL and a
2533 CRL distribution point embedded in an X509 certificate
2534 2) When verifying that a timestamp response token signer matches the
2535 timestamp authority name (exposed via the API functions
2536 TS_RESP_verify_response and TS_RESP_verify_token)
2541 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2543 * Certificates with explicit curve parameters are now disallowed in
2544 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2548 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2549 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2550 conversely, silently ignore DTLS protocol version bounds when configuring
2551 TLS-based contexts. The commands can be repeated to set bounds of both
2552 types. The same applies with the corresponding "min_protocol" and
2553 "max_protocol" command-line switches, in case some application uses both TLS
2556 SSL_CTX instances that are created for a fixed protocol version (e.g.
2557 TLSv1_server_method()) also silently ignore version bounds. Previously
2558 attempts to apply bounds to these protocol versions would result in an
2559 error. Now only the "version-flexible" SSL_CTX instances are subject to
2560 limits in configuration files in command-line options.
2564 * Handshake now fails if Extended Master Secret extension is dropped
2569 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2571 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2573 * Fixed segmentation fault in SSL_check_chain()
2574 Server or client applications that call the SSL_check_chain() function
2575 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2576 dereference as a result of incorrect handling of the
2577 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2578 or unrecognised signature algorithm is received from the peer. This could
2579 be exploited by a malicious peer in a Denial of Service attack.
2584 * Added AES consttime code for no-asm configurations
2585 an optional constant time support for AES was added
2586 when building openssl for no-asm.
2587 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2588 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2589 At this time this feature is by default disabled.
2590 It will be enabled by default in 3.0.
2594 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2596 * Revert the change of EOF detection while reading in libssl to avoid
2597 regressions in applications depending on the current way of reporting
2598 the EOF. As the existing method is not fully accurate the change to
2599 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2600 branch and will be present in the 3.0 release.
2604 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2605 when primes for RSA keys are computed.
2606 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2607 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2608 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2609 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2610 This avoids possible fingerprinting of newly generated RSA modules.
2614 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2616 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2617 while reading in libssl then we would report an error back to the
2618 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2619 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2620 therefore give a hint as to what went wrong.
2624 * Check that ed25519 and ed448 are allowed by the security level. Previously
2625 signature algorithms not using an MD were not being checked that they were
2626 allowed by the security level.
2630 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2631 was not quite right. The behaviour was not consistent between resumption
2632 and normal handshakes, and also not quite consistent with historical
2633 behaviour. The behaviour in various scenarios has been clarified and
2634 it has been updated to make it match historical behaviour as closely as
2639 * *[VMS only]* The header files that the VMS compilers include automatically,
2640 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2641 that the C++ compiler doesn't understand. This is a shortcoming in the
2642 compiler, but can be worked around with `__cplusplus` guards.
2644 C++ applications that use OpenSSL libraries must be compiled using the
2645 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2646 functions. Otherwise, only functions with symbols of less than 31
2647 characters can be used, as the linker will not be able to successfully
2648 resolve symbols with longer names.
2652 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2653 The presence of this system service is determined at run-time.
2657 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2662 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2664 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2665 number generator (RNG). This was intended to include protection in the
2666 event of a fork() system call in order to ensure that the parent and child
2667 processes did not share the same RNG state. However, this protection was not
2668 being used in the default case.
2670 A partial mitigation for this issue is that the output from a high
2671 precision timer is mixed into the RNG state so the likelihood of a parent
2672 and child process sharing state is significantly reduced.
2674 If an application already calls OPENSSL_init_crypto() explicitly using
2675 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2678 *Matthias St. Pierre*
2680 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2681 used even when parsing explicit parameters, when loading a encoded key
2682 or calling `EC_GROUP_new_from_ecpkparameters()`/
2683 `EC_GROUP_new_from_ecparameters()`.
2684 This prevents bypass of security hardening and performance gains,
2685 especially for curves with specialized EC_METHODs.
2686 By default, if a key encoded with explicit parameters is loaded and later
2687 encoded, the output is still encoded with explicit parameters, even if
2688 internally a "named" EC_GROUP is used for computation.
2692 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2693 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2694 NULL. After this change, only the cofactor parameter can be NULL. It also
2695 does some minimal sanity checks on the passed order.
2700 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2701 An attack is simple, if the first CMS_recipientInfo is valid but the
2702 second CMS_recipientInfo is chosen ciphertext. If the second
2703 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2704 encryption key will be replaced by garbage, and the message cannot be
2705 decoded, but if the RSA decryption fails, the correct encryption key is
2706 used and the recipient will not notice the attack.
2707 As a work around for this potential attack the length of the decrypted
2708 key must be equal to the cipher default key length, in case the
2709 certificate is not given and all recipientInfo are tried out.
2710 The old behaviour can be re-enabled in the CMS code by setting the
2711 CMS_DEBUG_DECRYPT flag.
2716 * Early start up entropy quality from the DEVRANDOM seed source has been
2717 improved for older Linux systems. The RAND subsystem will wait for
2718 /dev/random to be producing output before seeding from /dev/urandom.
2719 The seeded state is stored for future library initialisations using
2720 a system global shared memory segment. The shared memory identifier
2721 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2722 the desired value. The default identifier is 114.
2726 * Correct the extended master secret constant on EBCDIC systems. Without this
2727 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2728 negotiate EMS will fail. Unfortunately this also means that TLS connections
2729 between EBCDIC systems with this fix, and EBCDIC systems without this
2730 fix will fail if they negotiate EMS.
2734 * Use Windows installation paths in the mingw builds
2736 Mingw isn't a POSIX environment per se, which means that Windows
2737 paths should be used for installation.
2742 * Changed DH_check to accept parameters with order q and 2q subgroups.
2743 With order 2q subgroups the bit 0 of the private key is not secret
2744 but DH_generate_key works around that by clearing bit 0 of the
2745 private key for those. This avoids leaking bit 0 of the private key.
2749 * Significantly reduce secure memory usage by the randomness pools.
2753 * Revert the DEVRANDOM_WAIT feature for Linux systems
2755 The DEVRANDOM_WAIT feature added a select() call to wait for the
2756 /dev/random device to become readable before reading from the
2757 /dev/urandom device.
2759 It turned out that this change had negative side effects on
2760 performance which were not acceptable. After some discussion it
2761 was decided to revert this feature and leave it up to the OS
2762 resp. the platform maintainer to ensure a proper initialization
2763 during early boot time.
2765 *Matthias St. Pierre*
2767 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2769 * Add build tests for C++. These are generated files that only do one
2770 thing, to include one public OpenSSL head file each. This tests that
2771 the public header files can be usefully included in a C++ application.
2773 This test isn't enabled by default. It can be enabled with the option
2774 'enable-buildtest-c++'.
2778 * Enable SHA3 pre-hashing for ECDSA and DSA.
2782 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2783 This changes the size when using the `genpkey` command when no size is given.
2784 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2785 generation commands to use 2048 bits by default.
2789 * Reorganize the manual pages to consistently have RETURN VALUES,
2790 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2791 util/fix-doc-nits accordingly.
2793 *Paul Yang, Joshua Lock*
2795 * Add the missing accessor EVP_PKEY_get0_engine()
2799 * Have commands like `s_client` and `s_server` output the signature scheme
2800 along with other cipher suite parameters when debugging.
2804 * Make OPENSSL_config() error agnostic again.
2808 * Do the error handling in RSA decryption constant time.
2812 * Prevent over long nonces in ChaCha20-Poly1305.
2814 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2815 for every encryption operation. RFC 7539 specifies that the nonce value
2816 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2817 and front pads the nonce with 0 bytes if it is less than 12
2818 bytes. However it also incorrectly allows a nonce to be set of up to 16
2819 bytes. In this case only the last 12 bytes are significant and any
2820 additional leading bytes are ignored.
2822 It is a requirement of using this cipher that nonce values are
2823 unique. Messages encrypted using a reused nonce value are susceptible to
2824 serious confidentiality and integrity attacks. If an application changes
2825 the default nonce length to be longer than 12 bytes and then makes a
2826 change to the leading bytes of the nonce expecting the new value to be a
2827 new unique nonce then such an application could inadvertently encrypt
2828 messages with a reused nonce.
2830 Additionally the ignored bytes in a long nonce are not covered by the
2831 integrity guarantee of this cipher. Any application that relies on the
2832 integrity of these ignored leading bytes of a long nonce may be further
2833 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2834 is safe because no such use sets such a long nonce value. However user
2835 applications that use this cipher directly and set a non-default nonce
2836 length to be longer than 12 bytes may be vulnerable.
2838 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2844 * Add DEVRANDOM_WAIT feature for Linux systems
2846 On older Linux systems where the getrandom() system call is not available,
2847 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2848 Contrary to getrandom(), the /dev/urandom device will not block during
2849 early boot when the kernel CSPRNG has not been seeded yet.
2851 To mitigate this known weakness, use select() to wait for /dev/random to
2852 become readable before reading from /dev/urandom.
2854 * Ensure that SM2 only uses SM3 as digest algorithm
2858 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2860 * Change the info callback signals for the start and end of a post-handshake
2861 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2862 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2863 confused by this and assume that a TLSv1.2 renegotiation has started. This
2864 can break KeyUpdate handling. Instead we no longer signal the start and end
2865 of a post handshake message exchange (although the messages themselves are
2866 still signalled). This could break some applications that were expecting
2867 the old signals. However without this KeyUpdate is not usable for many
2872 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2874 * Timing vulnerability in DSA signature generation
2876 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2877 timing side channel attack. An attacker could use variations in the signing
2878 algorithm to recover the private key.
2880 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2885 * Timing vulnerability in ECDSA signature generation
2887 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2888 timing side channel attack. An attacker could use variations in the signing
2889 algorithm to recover the private key.
2891 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2896 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2897 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2898 of two gigabytes and the error handling improved.
2900 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2901 categorized as a normal bug, not a security issue, because the DRBG reseeds
2902 automatically and is fully functional even without additional randomness
2903 provided by the application.
2905 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2907 * Add a new ClientHello callback. Provides a callback interface that gives
2908 the application the ability to adjust the nascent SSL object at the
2909 earliest stage of ClientHello processing, immediately after extensions have
2910 been collected but before they have been processed. In particular, this
2911 callback can adjust the supported TLS versions in response to the contents
2916 * Add SM2 base algorithm support.
2920 * s390x assembly pack: add (improved) hardware-support for the following
2921 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2922 aes-cfb/cfb8, aes-ecb.
2926 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2927 parameter is no longer accepted, as it leads to a corrupt table. NULL
2928 pem_str is reserved for alias entries only.
2932 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2933 step for prime curves. The new implementation is based on formulae from
2934 differential addition-and-doubling in homogeneous projective coordinates
2935 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2936 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2937 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2938 to work in projective coordinates.
2940 *Billy Bob Brumley, Nicola Tuveri*
2942 * Change generating and checking of primes so that the error rate of not
2943 being prime depends on the intended use based on the size of the input.
2944 For larger primes this will result in more rounds of Miller-Rabin.
2945 The maximal error rate for primes with more than 1080 bits is lowered
2948 *Kurt Roeckx, Annie Yousar*
2950 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2954 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2955 moving between systems, and to avoid confusion when a Windows build is
2956 done with mingw vs with MSVC. For POSIX installs, there's still a
2957 symlink or copy named 'tsget' to avoid that confusion as well.
2961 * Revert blinding in ECDSA sign and instead make problematic addition
2962 length-invariant. Switch even to fixed-length Montgomery multiplication.
2966 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2967 step for binary curves. The new implementation is based on formulae from
2968 differential addition-and-doubling in mixed Lopez-Dahab projective
2969 coordinates, modified to independently blind the operands.
2971 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2973 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2974 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2975 EC_METHODs to implement their own specialized "ladder step", to take
2976 advantage of more favorable coordinate systems or more efficient
2977 differential addition-and-doubling algorithms.
2979 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2981 * Modified the random device based seed sources to keep the relevant
2982 file descriptors open rather than reopening them on each access.
2983 This allows such sources to operate in a chroot() jail without
2984 the associated device nodes being available. This behaviour can be
2985 controlled using RAND_keep_random_devices_open().
2989 * Numerous side-channel attack mitigations have been applied. This may have
2990 performance impacts for some algorithms for the benefit of improved
2991 security. Specific changes are noted in this change log by their respective
2996 * AIX shared library support overhaul. Switch to AIX "natural" way of
2997 handling shared libraries, which means collecting shared objects of
2998 different versions and bitnesses in one common archive. This allows to
2999 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3000 doesn't affect the way 3rd party applications are linked, only how
3001 multi-version installation is managed.
3005 * Make ec_group_do_inverse_ord() more robust and available to other
3006 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3007 mitigations are applied to the fallback BN_mod_inverse().
3008 When using this function rather than BN_mod_inverse() directly, new
3009 EC cryptosystem implementations are then safer-by-default.
3013 * Add coordinate blinding for EC_POINT and implement projective
3014 coordinate blinding for generic prime curves as a countermeasure to
3015 chosen point SCA attacks.
3017 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3019 * Add blinding to ECDSA and DSA signatures to protect against side channel
3020 attacks discovered by Keegan Ryan (NCC Group).
3024 * Enforce checking in the `pkeyutl` command to ensure that the input
3025 length does not exceed the maximum supported digest length when performing
3026 a sign, verify or verifyrecover operation.
3030 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3031 I/O in combination with something like select() or poll() will hang. This
3032 can be turned off again using SSL_CTX_clear_mode().
3033 Many applications do not properly handle non-application data records, and
3034 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3035 around the problems in those applications, but can also break some.
3036 It's recommended to read the manpages about SSL_read(), SSL_write(),
3037 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3038 SSL_CTX_set_read_ahead() again.
3042 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3043 now allow empty (zero character) pass phrases.
3047 * Apply blinding to binary field modular inversion and remove patent
3048 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3052 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3053 binary and prime elliptic curves.
3057 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3058 constant time fixed point multiplication.
3062 * Revise elliptic curve scalar multiplication with timing attack
3063 defenses: ec_wNAF_mul redirects to a constant time implementation
3064 when computing fixed point and variable point multiplication (which
3065 in OpenSSL are mostly used with secret scalars in keygen, sign,
3066 ECDH derive operations).
3067 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3070 * Updated CONTRIBUTING
3074 * Updated DRBG / RAND to request nonce and additional low entropy
3075 randomness from the system.
3077 *Matthias St. Pierre*
3079 * Updated 'openssl rehash' to use OpenSSL consistent default.
3083 * Moved the load of the ssl_conf module to libcrypto, which helps
3084 loading engines that libssl uses before libssl is initialised.
3088 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3092 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3094 *Ingo Schwarze, Rich Salz*
3096 * Added output of accepting IP address and port for 'openssl s_server'
3100 * Added a new API for TLSv1.3 ciphersuites:
3101 SSL_CTX_set_ciphersuites()
3102 SSL_set_ciphersuites()
3106 * Memory allocation failures consistently add an error to the error
3111 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3112 in libcrypto when run as setuid/setgid.
3116 * Load any config file by default when libssl is used.
3120 * Added new public header file <openssl/rand_drbg.h> and documentation
3121 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3123 *Matthias St. Pierre*
3125 * QNX support removed (cannot find contributors to get their approval
3126 for the license change).
3130 * TLSv1.3 replay protection for early data has been implemented. See the
3131 SSL_read_early_data() man page for further details.
3135 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3136 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3137 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3138 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3139 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3140 configuration has been separated out. See the ciphers man page or the
3141 SSL_CTX_set_ciphersuites() man page for more information.
3145 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3146 in responder mode now supports the new "-multi" option, which
3147 spawns the specified number of child processes to handle OCSP
3148 requests. The "-timeout" option now also limits the OCSP
3149 responder's patience to wait to receive the full client request
3150 on a newly accepted connection. Child processes are respawned
3151 as needed, and the CA index file is automatically reloaded
3152 when changed. This makes it possible to run the "ocsp" responder
3153 as a long-running service, making the OpenSSL CA somewhat more
3154 feature-complete. In this mode, most diagnostic messages logged
3155 after entering the event loop are logged via syslog(3) rather than
3160 * Added support for X448 and Ed448. Heavily based on original work by
3165 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3166 objects loaded. This adds the functions OSSL_STORE_expect() and
3167 OSSL_STORE_find() as well as needed tools to construct searches and
3168 get the search data out of them.
3172 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3173 version of OpenSSL should review their configuration settings to ensure
3174 that they are still appropriate for TLSv1.3. For further information see:
3175 <https://wiki.openssl.org/index.php/TLS1.3>
3179 * Grand redesign of the OpenSSL random generator
3181 The default RAND method now utilizes an AES-CTR DRBG according to
3182 NIST standard SP 800-90Ar1. The new random generator is essentially
3183 a port of the default random generator from the OpenSSL FIPS 2.0
3184 object module. It is a hybrid deterministic random bit generator
3185 using an AES-CTR bit stream and which seeds and reseeds itself
3186 automatically using trusted system entropy sources.
3188 Some of its new features are:
3189 - Support for multiple DRBG instances with seed chaining.
3190 - The default RAND method makes use of a DRBG.
3191 - There is a public and private DRBG instance.
3192 - The DRBG instances are fork-safe.
3193 - Keep all global DRBG instances on the secure heap if it is enabled.
3194 - The public and private DRBG instance are per thread for lock free
3197 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3199 * Changed Configure so it only says what it does and doesn't dump
3200 so much data. Instead, ./configdata.pm should be used as a script
3201 to display all sorts of configuration data.
3205 * Added processing of "make variables" to Configure.
3209 * Added SHA512/224 and SHA512/256 algorithm support.
3213 * The last traces of Netware support, first removed in 1.1.0, have
3218 * Get rid of Makefile.shared, and in the process, make the processing
3219 of certain files (rc.obj, or the .def/.map/.opt files produced from
3220 the ordinal files) more visible and hopefully easier to trace and
3221 debug (or make silent).
3225 * Make it possible to have environment variable assignments as
3226 arguments to config / Configure.
3230 * Add multi-prime RSA (RFC 8017) support.
3234 * Add SM3 implemented according to GB/T 32905-2016
3235 *Jack Lloyd <jack.lloyd@ribose.com>,*
3236 *Ronald Tse <ronald.tse@ribose.com>,*
3237 *Erick Borsboom <erick.borsboom@ribose.com>*
3239 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3240 as documented in RFC6066.
3241 Based on a patch from Tomasz Moń
3243 *Filipe Raimundo da Silva*
3245 * Add SM4 implemented according to GB/T 32907-2016.
3246 *Jack Lloyd <jack.lloyd@ribose.com>,*
3247 *Ronald Tse <ronald.tse@ribose.com>,*
3248 *Erick Borsboom <erick.borsboom@ribose.com>*
3250 * Reimplement -newreq-nodes and ERR_error_string_n; the
3251 original author does not agree with the license change.
3255 * Add ARIA AEAD TLS support.
3259 * Some macro definitions to support VS6 have been removed. Visual
3260 Studio 6 has not worked since 1.1.0
3264 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3265 without clearing the errors.
3269 * Add "atfork" functions. If building on a system that without
3270 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3271 requirements. The RAND facility now uses/requires this.
3279 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3280 not possible to disable entirely. However, it's still possible to
3281 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3284 To disable, configure with 'no-ui-console'. 'no-ui' is still
3285 possible to use as an alias. Check at compile time with the
3286 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3287 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3291 * Add a STORE module, which implements a uniform and URI based reader of
3292 stores that can contain keys, certificates, CRLs and numerous other
3293 objects. The main API is loosely based on a few stdio functions,
3294 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3295 OSSL_STORE_error and OSSL_STORE_close.
3296 The implementation uses backends called "loaders" to implement arbitrary
3297 URI schemes. There is one built in "loader" for the 'file' scheme.
3301 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3302 then adjusted to work on FreeBSD 8.4 as well.
3303 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3304 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3308 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3309 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3310 error code calls like this:
3312 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3314 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3315 that can be encoded in C. For the foreseeable future, this will only
3318 *Richard Levitte and Tim Hudson*
3320 * Removed BSD cryptodev engine.
3324 * Add a build target 'build_all_generated', to build all generated files
3325 and only that. This can be used to prepare everything that requires
3326 things like perl for a system that lacks perl and then move everything
3327 to that system and do the rest of the build there.
3331 * In the UI interface, make it possible to duplicate the user data. This
3332 can be used by engines that need to retain the data for a longer time
3333 than just the call where this user data is passed.
3337 * Ignore the '-named_curve auto' value for compatibility of applications
3340 *Tomáš Mráz <tmraz@fedoraproject.org>*
3342 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3343 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3344 alerts across multiple records (some of which could be empty). In practice
3345 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3346 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3347 support this at all. Supporting it adds significant complexity to the
3348 record layer, and its removal is unlikely to cause interoperability
3353 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3354 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3355 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3360 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3361 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3363 *Richard Levitte, Andy Polyakov*
3365 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3370 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3371 platform rather than 'mingw'.
3375 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3376 success if they are asked to add an object which already exists
3377 in the store. This change cascades to other functions which load
3378 certificates and CRLs.
3382 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3383 facilitate stack unwinding even from assembly subroutines.
3387 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3388 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3392 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3393 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3394 which is the minimum version we support.
3398 * Certificate time validation (X509_cmp_time) enforces stricter
3399 compliance with RFC 5280. Fractional seconds and timezone offsets
3400 are no longer allowed.
3404 * Add support for ARIA
3408 * s_client will now send the Server Name Indication (SNI) extension by
3409 default unless the new "-noservername" option is used. The server name is
3410 based on the host provided to the "-connect" option unless overridden by
3411 using "-servername".
3415 * Add support for SipHash
3419 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3420 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3421 prevent issues where no progress is being made and the peer continually
3422 sends unrecognised record types, using up resources processing them.
3426 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3427 using the algorithm defined in
3428 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3432 * Heartbeat support has been removed; the ABI is changed for now.
3434 *Richard Levitte, Rich Salz*
3436 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3440 * The RSA "null" method, which was partially supported to avoid patent
3441 issues, has been replaced to always returns NULL.
3448 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3450 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3451 used even when parsing explicit parameters, when loading a encoded key
3452 or calling `EC_GROUP_new_from_ecpkparameters()`/
3453 `EC_GROUP_new_from_ecparameters()`.
3454 This prevents bypass of security hardening and performance gains,
3455 especially for curves with specialized EC_METHODs.
3456 By default, if a key encoded with explicit parameters is loaded and later
3457 encoded, the output is still encoded with explicit parameters, even if
3458 internally a "named" EC_GROUP is used for computation.
3462 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3463 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3464 NULL. After this change, only the cofactor parameter can be NULL. It also
3465 does some minimal sanity checks on the passed order.
3470 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3471 An attack is simple, if the first CMS_recipientInfo is valid but the
3472 second CMS_recipientInfo is chosen ciphertext. If the second
3473 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3474 encryption key will be replaced by garbage, and the message cannot be
3475 decoded, but if the RSA decryption fails, the correct encryption key is
3476 used and the recipient will not notice the attack.
3477 As a work around for this potential attack the length of the decrypted
3478 key must be equal to the cipher default key length, in case the
3479 certificate is not given and all recipientInfo are tried out.
3480 The old behaviour can be re-enabled in the CMS code by setting the
3481 CMS_DEBUG_DECRYPT flag.
3486 * Use Windows installation paths in the mingw builds
3488 Mingw isn't a POSIX environment per se, which means that Windows
3489 paths should be used for installation.
3494 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3496 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3497 This changes the size when using the `genpkey` command when no size is given.
3498 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3499 generation commands to use 2048 bits by default.
3503 * Prevent over long nonces in ChaCha20-Poly1305.
3505 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3506 for every encryption operation. RFC 7539 specifies that the nonce value
3507 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3508 and front pads the nonce with 0 bytes if it is less than 12
3509 bytes. However it also incorrectly allows a nonce to be set of up to 16
3510 bytes. In this case only the last 12 bytes are significant and any
3511 additional leading bytes are ignored.
3513 It is a requirement of using this cipher that nonce values are
3514 unique. Messages encrypted using a reused nonce value are susceptible to
3515 serious confidentiality and integrity attacks. If an application changes
3516 the default nonce length to be longer than 12 bytes and then makes a
3517 change to the leading bytes of the nonce expecting the new value to be a
3518 new unique nonce then such an application could inadvertently encrypt
3519 messages with a reused nonce.
3521 Additionally the ignored bytes in a long nonce are not covered by the
3522 integrity guarantee of this cipher. Any application that relies on the
3523 integrity of these ignored leading bytes of a long nonce may be further
3524 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3525 is safe because no such use sets such a long nonce value. However user
3526 applications that use this cipher directly and set a non-default nonce
3527 length to be longer than 12 bytes may be vulnerable.
3529 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3535 * Added SCA hardening for modular field inversion in EC_GROUP through
3536 a new dedicated field_inv() pointer in EC_METHOD.
3537 This also addresses a leakage affecting conversions from projective
3538 to affine coordinates.
3540 *Billy Bob Brumley, Nicola Tuveri*
3542 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3543 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3547 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3551 * Remove the 'dist' target and add a tarball building script. The
3552 'dist' target has fallen out of use, and it shouldn't be
3553 necessary to configure just to create a source distribution.
3557 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3559 * Timing vulnerability in DSA signature generation
3561 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3562 timing side channel attack. An attacker could use variations in the signing
3563 algorithm to recover the private key.
3565 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3570 * Timing vulnerability in ECDSA signature generation
3572 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3573 timing side channel attack. An attacker could use variations in the signing
3574 algorithm to recover the private key.
3576 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3581 * Add coordinate blinding for EC_POINT and implement projective
3582 coordinate blinding for generic prime curves as a countermeasure to
3583 chosen point SCA attacks.
3585 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3587 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3589 * Client DoS due to large DH parameter
3591 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3592 malicious server can send a very large prime value to the client. This will
3593 cause the client to spend an unreasonably long period of time generating a
3594 key for this prime resulting in a hang until the client has finished. This
3595 could be exploited in a Denial Of Service attack.
3597 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3602 * Cache timing vulnerability in RSA Key Generation
3604 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3605 a cache timing side channel attack. An attacker with sufficient access to
3606 mount cache timing attacks during the RSA key generation process could
3607 recover the private key.
3609 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3610 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3615 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3616 parameter is no longer accepted, as it leads to a corrupt table. NULL
3617 pem_str is reserved for alias entries only.
3621 * Revert blinding in ECDSA sign and instead make problematic addition
3622 length-invariant. Switch even to fixed-length Montgomery multiplication.
3626 * Change generating and checking of primes so that the error rate of not
3627 being prime depends on the intended use based on the size of the input.
3628 For larger primes this will result in more rounds of Miller-Rabin.
3629 The maximal error rate for primes with more than 1080 bits is lowered
3632 *Kurt Roeckx, Annie Yousar*
3634 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3638 * Add blinding to ECDSA and DSA signatures to protect against side channel
3639 attacks discovered by Keegan Ryan (NCC Group).
3643 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3644 now allow empty (zero character) pass phrases.
3648 * Certificate time validation (X509_cmp_time) enforces stricter
3649 compliance with RFC 5280. Fractional seconds and timezone offsets
3650 are no longer allowed.
3654 * Fixed a text canonicalisation bug in CMS
3656 Where a CMS detached signature is used with text content the text goes
3657 through a canonicalisation process first prior to signing or verifying a
3658 signature. This process strips trailing space at the end of lines, converts
3659 line terminators to CRLF and removes additional trailing line terminators
3660 at the end of a file. A bug in the canonicalisation process meant that
3661 some characters, such as form-feed, were incorrectly treated as whitespace
3662 and removed. This is contrary to the specification (RFC5485). This fix
3663 could mean that detached text data signed with an earlier version of
3664 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3665 signed with a fixed OpenSSL may fail to verify with an earlier version of
3666 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3667 and use the "-binary" flag (for the "cms" command line application) or set
3668 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3672 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3674 * Constructed ASN.1 types with a recursive definition could exceed the stack
3676 Constructed ASN.1 types with a recursive definition (such as can be found
3677 in PKCS7) could eventually exceed the stack given malicious input with
3678 excessive recursion. This could result in a Denial Of Service attack. There
3679 are no such structures used within SSL/TLS that come from untrusted sources
3680 so this is considered safe.
3682 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3688 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3690 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3691 effectively reduced to only comparing the least significant bit of each
3692 byte. This allows an attacker to forge messages that would be considered as
3693 authenticated in an amount of tries lower than that guaranteed by the
3694 security claims of the scheme. The module can only be compiled by the
3695 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3697 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3703 * Add a build target 'build_all_generated', to build all generated files
3704 and only that. This can be used to prepare everything that requires
3705 things like perl for a system that lacks perl and then move everything
3706 to that system and do the rest of the build there.
3710 * Backport SSL_OP_NO_RENGOTIATION
3712 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3713 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3714 changes this is no longer possible in 1.1.0. Therefore, the new
3715 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3716 1.1.0 to provide equivalent functionality.
3718 Note that if an application built against 1.1.0h headers (or above) is run
3719 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3720 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3724 * Removed the OS390-Unix config target. It relied on a script that doesn't
3729 * rsaz_1024_mul_avx2 overflow bug on x86_64
3731 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3732 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3733 Analysis suggests that attacks against RSA and DSA as a result of this
3734 defect would be very difficult to perform and are not believed likely.
3735 Attacks against DH1024 are considered just feasible, because most of the
3736 work necessary to deduce information about a private key may be performed
3737 offline. The amount of resources required for such an attack would be
3738 significant. However, for an attack on TLS to be meaningful, the server
3739 would have to share the DH1024 private key among multiple clients, which is
3740 no longer an option since CVE-2016-0701.
3742 This only affects processors that support the AVX2 but not ADX extensions
3743 like Intel Haswell (4th generation).
3745 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3746 was originally found via the OSS-Fuzz project.
3751 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3753 * bn_sqrx8x_internal carry bug on x86_64
3755 There is a carry propagating bug in the x86_64 Montgomery squaring
3756 procedure. No EC algorithms are affected. Analysis suggests that attacks
3757 against RSA and DSA as a result of this defect would be very difficult to
3758 perform and are not believed likely. Attacks against DH are considered just
3759 feasible (although very difficult) because most of the work necessary to
3760 deduce information about a private key may be performed offline. The amount
3761 of resources required for such an attack would be very significant and
3762 likely only accessible to a limited number of attackers. An attacker would
3763 additionally need online access to an unpatched system using the target
3764 private key in a scenario with persistent DH parameters and a private
3765 key that is shared between multiple clients.
3767 This only affects processors that support the BMI1, BMI2 and ADX extensions
3768 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3770 This issue was reported to OpenSSL by the OSS-Fuzz project.
3775 * Malformed X.509 IPAddressFamily could cause OOB read
3777 If an X.509 certificate has a malformed IPAddressFamily extension,
3778 OpenSSL could do a one-byte buffer overread. The most likely result
3779 would be an erroneous display of the certificate in text format.
3781 This issue was reported to OpenSSL by the OSS-Fuzz project.
3786 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3788 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3789 platform rather than 'mingw'.
3793 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3794 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3795 which is the minimum version we support.
3799 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3801 * Encrypt-Then-Mac renegotiation crash
3803 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3804 negotiated where it was not in the original handshake (or vice-versa) then
3805 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3806 and servers are affected.
3808 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3813 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3815 * Truncated packet could crash via OOB read
3817 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3818 cipher is being used, then a truncated packet can cause that host to
3819 perform an out-of-bounds read, usually resulting in a crash.
3821 This issue was reported to OpenSSL by Robert Święcki of Google.
3826 * Bad (EC)DHE parameters cause a client crash
3828 If a malicious server supplies bad parameters for a DHE or ECDHE key
3829 exchange then this can result in the client attempting to dereference a
3830 NULL pointer leading to a client crash. This could be exploited in a Denial
3833 This issue was reported to OpenSSL by Guido Vranken.
3838 * BN_mod_exp may produce incorrect results on x86_64
3840 There is a carry propagating bug in the x86_64 Montgomery squaring
3841 procedure. No EC algorithms are affected. Analysis suggests that attacks
3842 against RSA and DSA as a result of this defect would be very difficult to
3843 perform and are not believed likely. Attacks against DH are considered just
3844 feasible (although very difficult) because most of the work necessary to
3845 deduce information about a private key may be performed offline. The amount
3846 of resources required for such an attack would be very significant and
3847 likely only accessible to a limited number of attackers. An attacker would
3848 additionally need online access to an unpatched system using the target
3849 private key in a scenario with persistent DH parameters and a private
3850 key that is shared between multiple clients. For example this can occur by
3851 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3852 similar to CVE-2015-3193 but must be treated as a separate problem.
3854 This issue was reported to OpenSSL by the OSS-Fuzz project.
3859 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3861 * ChaCha20/Poly1305 heap-buffer-overflow
3863 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3864 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3865 crash. This issue is not considered to be exploitable beyond a DoS.
3867 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3872 * CMS Null dereference
3874 Applications parsing invalid CMS structures can crash with a NULL pointer
3875 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3876 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3877 structure callback if an attempt is made to free certain invalid encodings.
3878 Only CHOICE structures using a callback which do not handle NULL value are
3881 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3886 * Montgomery multiplication may produce incorrect results
3888 There is a carry propagating bug in the Broadwell-specific Montgomery
3889 multiplication procedure that handles input lengths divisible by, but
3890 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3891 and DH private keys are impossible. This is because the subroutine in
3892 question is not used in operations with the private key itself and an input
3893 of the attacker's direct choice. Otherwise the bug can manifest itself as
3894 transient authentication and key negotiation failures or reproducible
3895 erroneous outcome of public-key operations with specially crafted input.
3896 Among EC algorithms only Brainpool P-512 curves are affected and one
3897 presumably can attack ECDH key negotiation. Impact was not analyzed in
3898 detail, because pre-requisites for attack are considered unlikely. Namely
3899 multiple clients have to choose the curve in question and the server has to
3900 share the private key among them, neither of which is default behaviour.
3901 Even then only clients that chose the curve will be affected.
3903 This issue was publicly reported as transient failures and was not
3904 initially recognized as a security issue. Thanks to Richard Morgan for
3905 providing reproducible case.
3910 * Removed automatic addition of RPATH in shared libraries and executables,
3911 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3915 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3917 * Fix Use After Free for large message sizes
3919 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3920 message larger than approx 16k is received then the underlying buffer to
3921 store the incoming message is reallocated and moved. Unfortunately a
3922 dangling pointer to the old location is left which results in an attempt to
3923 write to the previously freed location. This is likely to result in a
3924 crash, however it could potentially lead to execution of arbitrary code.
3926 This issue only affects OpenSSL 1.1.0a.
3928 This issue was reported to OpenSSL by Robert Święcki.
3933 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3935 * OCSP Status Request extension unbounded memory growth
3937 A malicious client can send an excessively large OCSP Status Request
3938 extension. If that client continually requests renegotiation, sending a
3939 large OCSP Status Request extension each time, then there will be unbounded
3940 memory growth on the server. This will eventually lead to a Denial Of
3941 Service attack through memory exhaustion. Servers with a default
3942 configuration are vulnerable even if they do not support OCSP. Builds using
3943 the "no-ocsp" build time option are not affected.
3945 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3950 * SSL_peek() hang on empty record
3952 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3953 sends an empty record. This could be exploited by a malicious peer in a
3954 Denial Of Service attack.
3956 This issue was reported to OpenSSL by Alex Gaynor.
3961 * Excessive allocation of memory in tls_get_message_header() and
3962 dtls1_preprocess_fragment()
3964 A (D)TLS message includes 3 bytes for its length in the header for the
3965 message. This would allow for messages up to 16Mb in length. Messages of
3966 this length are excessive and OpenSSL includes a check to ensure that a
3967 peer is sending reasonably sized messages in order to avoid too much memory
3968 being consumed to service a connection. A flaw in the logic of version
3969 1.1.0 means that memory for the message is allocated too early, prior to
3970 the excessive message length check. Due to way memory is allocated in
3971 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3972 to service a connection. This could lead to a Denial of Service through
3973 memory exhaustion. However, the excessive message length check still takes
3974 place, and this would cause the connection to immediately fail. Assuming
3975 that the application calls SSL_free() on the failed connection in a timely
3976 manner then the 21Mb of allocated memory will then be immediately freed
3977 again. Therefore, the excessive memory allocation will be transitory in
3978 nature. This then means that there is only a security impact if:
3980 1) The application does not call SSL_free() in a timely manner in the event
3981 that the connection fails
3983 2) The application is working in a constrained environment where there is
3984 very little free memory
3986 3) The attacker initiates multiple connection attempts such that there are
3987 multiple connections in a state where memory has been allocated for the
3988 connection; SSL_free() has not yet been called; and there is insufficient
3989 memory to service the multiple requests.
3991 Except in the instance of (1) above any Denial Of Service is likely to be
3992 transitory because as soon as the connection fails the memory is
3993 subsequently freed again in the SSL_free() call. However there is an
3994 increased risk during this period of application crashes due to the lack of
3995 memory - which would then mean a more serious Denial of Service.
3997 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3998 (CVE-2016-6307 and CVE-2016-6308)
4002 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4003 had to be removed. Primary reason is that vendor assembler can't
4004 assemble our modules with -KPIC flag. As result it, assembly
4005 support, was not even available as option. But its lack means
4006 lack of side-channel resistant code, which is incompatible with
4007 security by todays standards. Fortunately gcc is readily available
4008 prepackaged option, which we firmly point at...
4012 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4014 * Windows command-line tool supports UTF-8 opt-in option for arguments
4015 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4016 (to any value) allows Windows user to access PKCS#12 file generated
4017 with Windows CryptoAPI and protected with non-ASCII password, as well
4018 as files generated under UTF-8 locale on Linux also protected with
4023 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4024 have been disabled by default and removed from DEFAULT, just like RC4.
4025 See the RC4 item below to re-enable both.
4029 * The method for finding the storage location for the Windows RAND seed file
4030 has changed. First we check %RANDFILE%. If that is not set then we check
4031 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4032 all else fails we fall back to C:\.
4036 * The EVP_EncryptUpdate() function has had its return type changed from void
4037 to int. A return of 0 indicates and error while a return of 1 indicates
4042 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4043 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4044 off the constant time implementation for RSA, DSA and DH have been made
4045 no-ops and deprecated.
4049 * Windows RAND implementation was simplified to only get entropy by
4050 calling CryptGenRandom(). Various other RAND-related tickets
4053 *Joseph Wylie Yandle, Rich Salz*
4055 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4056 and `OPENSSL_LH_`, respectively. The old names are available
4057 with API compatibility. They new names are now completely documented.
4061 * Unify TYPE_up_ref(obj) methods signature.
4062 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4063 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4064 int (instead of void) like all others TYPE_up_ref() methods.
4065 So now these methods also check the return value of CRYPTO_atomic_add(),
4066 and the validity of object reference counter.
4068 *fdasilvayy@gmail.com*
4070 * With Windows Visual Studio builds, the .pdb files are installed
4071 alongside the installed libraries and executables. For a static
4072 library installation, ossl_static.pdb is the associate compiler
4073 generated .pdb file to be used when linking programs.
4077 * Remove openssl.spec. Packaging files belong with the packagers.
4081 * Automatic Darwin/OSX configuration has had a refresh, it will now
4082 recognise x86_64 architectures automatically. You can still decide
4083 to build for a different bitness with the environment variable
4084 KERNEL_BITS (can be 32 or 64), for example:
4086 KERNEL_BITS=32 ./config
4090 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4091 256 bit AES and HMAC with SHA256.
4095 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4099 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4103 * To enable users to have their own config files and build file templates,
4104 Configure looks in the directory indicated by the environment variable
4105 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4106 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4107 name and is used as is.
4111 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4112 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4113 X509_CERT_FILE_CTX was removed.
4117 * "shared" builds are now the default. To create only static libraries use
4118 the "no-shared" Configure option.
4122 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4123 All of these option have not worked for some while and are fundamental
4128 * Make various cleanup routines no-ops and mark them as deprecated. Most
4129 global cleanup functions are no longer required because they are handled
4130 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4131 Explicitly de-initing can cause problems (e.g. where a library that uses
4132 OpenSSL de-inits, but an application is still using it). The affected
4133 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4134 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4135 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4136 COMP_zlib_cleanup().
4140 * --strict-warnings no longer enables runtime debugging options
4141 such as REF_DEBUG. Instead, debug options are automatically
4142 enabled with '--debug' builds.
4144 *Andy Polyakov, Emilia Käsper*
4146 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4147 have been moved out of the public header files. New functions for managing
4148 these have been added.
4152 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4153 objects have been moved out of the public header files. New
4154 functions for managing these have been added.
4158 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4159 have been moved out of the public header files. New functions for managing
4160 these have been added.
4164 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4165 moved out of the public header files. New functions for managing these
4170 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4174 * Removed the mk1mf build scripts.
4178 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4179 it is always safe to #include a header now.
4183 * Removed the aged BC-32 config and all its supporting scripts
4187 * Removed support for Ultrix, Netware, and OS/2.
4191 * Add support for HKDF.
4193 *Alessandro Ghedini*
4195 * Add support for blake2b and blake2s
4199 * Added support for "pipelining". Ciphers that have the
4200 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4201 encryptions/decryptions simultaneously. There are currently no built-in
4202 ciphers with this property but the expectation is that engines will be able
4203 to offer it to significantly improve throughput. Support has been extended
4204 into libssl so that multiple records for a single connection can be
4205 processed in one go (for >=TLS 1.1).
4209 * Added the AFALG engine. This is an async capable engine which is able to
4210 offload work to the Linux kernel. In this initial version it only supports
4211 AES128-CBC. The kernel must be version 4.1.0 or greater.
4215 * OpenSSL now uses a new threading API. It is no longer necessary to
4216 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4217 are two supported threading models: pthreads and windows threads. It is
4218 also possible to configure OpenSSL at compile time for "no-threads". The
4219 old threading API should no longer be used. The functions have been
4220 replaced with "no-op" compatibility macros.
4222 *Alessandro Ghedini, Matt Caswell*
4224 * Modify behavior of ALPN to invoke callback after SNI/servername
4225 callback, such that updates to the SSL_CTX affect ALPN.
4229 * Add SSL_CIPHER queries for authentication and key-exchange.
4233 * Changes to the DEFAULT cipherlist:
4234 - Prefer (EC)DHE handshakes over plain RSA.
4235 - Prefer AEAD ciphers over legacy ciphers.
4236 - Prefer ECDSA over RSA when both certificates are available.
4237 - Prefer TLSv1.2 ciphers/PRF.
4238 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4243 * Change the ECC default curve list to be this, in order: x25519,
4244 secp256r1, secp521r1, secp384r1.
4248 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4249 disabled by default. They can be re-enabled using the
4250 enable-weak-ssl-ciphers option to Configure.
4254 * If the server has ALPN configured, but supports no protocols that the
4255 client advertises, send a fatal "no_application_protocol" alert.
4256 This behaviour is SHALL in RFC 7301, though it isn't universally
4257 implemented by other servers.
4261 * Add X25519 support.
4262 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4263 for public and private key encoding using the format documented in
4264 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4265 key generation and key derivation.
4267 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4272 * Deprecate SRP_VBASE_get_by_user.
4273 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4274 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4275 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4276 seed, even if the seed is configured.
4278 Users should use SRP_VBASE_get1_by_user instead. Note that in
4279 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4280 also that even though configuring the SRP seed attempts to hide
4281 invalid usernames by continuing the handshake with fake
4282 credentials, this behaviour is not constant time and no strong
4283 guarantees are made that the handshake is indistinguishable from
4284 that of a valid user.
4288 * Configuration change; it's now possible to build dynamic engines
4289 without having to build shared libraries and vice versa. This
4290 only applies to the engines in `engines/`, those in `crypto/engine/`
4291 will always be built into libcrypto (i.e. "static").
4293 Building dynamic engines is enabled by default; to disable, use
4294 the configuration option "disable-dynamic-engine".
4296 The only requirements for building dynamic engines are the
4297 presence of the DSO module and building with position independent
4298 code, so they will also automatically be disabled if configuring
4299 with "disable-dso" or "disable-pic".
4301 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4302 are also taken away from openssl/opensslconf.h, as they are
4307 * Configuration change; if there is a known flag to compile
4308 position independent code, it will always be applied on the
4309 libcrypto and libssl object files, and never on the application
4310 object files. This means other libraries that use routines from
4311 libcrypto / libssl can be made into shared libraries regardless
4312 of how OpenSSL was configured.
4314 If this isn't desirable, the configuration options "disable-pic"
4315 or "no-pic" can be used to disable the use of PIC. This will
4316 also disable building shared libraries and dynamic engines.
4320 * Removed JPAKE code. It was experimental and has no wide use.
4324 * The INSTALL_PREFIX Makefile variable has been renamed to
4325 DESTDIR. That makes for less confusion on what this variable
4326 is for. Also, the configuration option --install_prefix is
4331 * Heartbeat for TLS has been removed and is disabled by default
4332 for DTLS; configure with enable-heartbeats. Code that uses the
4333 old #define's might need to be updated.
4335 *Emilia Käsper, Rich Salz*
4337 * Rename REF_CHECK to REF_DEBUG.
4341 * New "unified" build system
4343 The "unified" build system is aimed to be a common system for all
4344 platforms we support. With it comes new support for VMS.
4346 This system builds supports building in a different directory tree
4347 than the source tree. It produces one Makefile (for unix family
4348 or lookalikes), or one descrip.mms (for VMS).
4350 The source of information to make the Makefile / descrip.mms is
4351 small files called 'build.info', holding the necessary
4352 information for each directory with source to compile, and a
4353 template in Configurations, like unix-Makefile.tmpl or
4356 With this change, the library names were also renamed on Windows
4357 and on VMS. They now have names that are closer to the standard
4358 on Unix, and include the major version number, and in certain
4359 cases, the architecture they are built for. See "Notes on shared
4360 libraries" in INSTALL.
4362 We rely heavily on the perl module Text::Template.
4366 * Added support for auto-initialisation and de-initialisation of the library.
4367 OpenSSL no longer requires explicit init or deinit routines to be called,
4368 except in certain circumstances. See the OPENSSL_init_crypto() and
4369 OPENSSL_init_ssl() man pages for further information.
4373 * The arguments to the DTLSv1_listen function have changed. Specifically the
4374 "peer" argument is now expected to be a BIO_ADDR object.
4376 * Rewrite of BIO networking library. The BIO library lacked consistent
4377 support of IPv6, and adding it required some more extensive
4378 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4379 which hold all types of addresses and chains of address information.
4380 It also introduces a new API, with functions like BIO_socket,
4381 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4382 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4383 have been adapted accordingly.
4387 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4392 * CRIME protection: disable compression by default, even if OpenSSL is
4393 compiled with zlib enabled. Applications can still enable compression
4394 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4395 using the SSL_CONF library to configure compression.
4399 * The signature of the session callback configured with
4400 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4401 was explicitly marked as `const unsigned char*` instead of
4406 * Always DPURIFY. Remove the use of uninitialized memory in the
4407 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4411 * Removed many obsolete configuration items, including
4412 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4413 MD2_CHAR, MD2_INT, MD2_LONG
4415 IDEA_SHORT, IDEA_LONG
4416 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4418 *Rich Salz, with advice from Andy Polyakov*
4420 * Many BN internals have been moved to an internal header file.
4422 *Rich Salz with help from Andy Polyakov*
4424 * Configuration and writing out the results from it has changed.
4425 Files such as Makefile include/openssl/opensslconf.h and are now
4426 produced through general templates, such as Makefile.in and
4427 crypto/opensslconf.h.in and some help from the perl module
4430 Also, the center of configuration information is no longer
4431 Makefile. Instead, Configure produces a perl module in
4432 configdata.pm which holds most of the config data (in the hash
4433 table %config), the target data that comes from the target
4434 configuration in one of the `Configurations/*.conf` files (in
4439 * To clarify their intended purposes, the Configure options
4440 --prefix and --openssldir change their semantics, and become more
4441 straightforward and less interdependent.
4443 --prefix shall be used exclusively to give the location INSTALLTOP
4444 where programs, scripts, libraries, include files and manuals are
4445 going to be installed. The default is now /usr/local.
4447 --openssldir shall be used exclusively to give the default
4448 location OPENSSLDIR where certificates, private keys, CRLs are
4449 managed. This is also where the default openssl.cnf gets
4451 If the directory given with this option is a relative path, the
4452 values of both the --prefix value and the --openssldir value will
4453 be combined to become OPENSSLDIR.
4454 The default for --openssldir is INSTALLTOP/ssl.
4456 Anyone who uses --openssldir to specify where OpenSSL is to be
4457 installed MUST change to use --prefix instead.
4461 * The GOST engine was out of date and therefore it has been removed. An up
4462 to date GOST engine is now being maintained in an external repository.
4463 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4464 support for GOST ciphersuites (these are only activated if a GOST engine
4469 * EGD is no longer supported by default; use enable-egd when
4472 *Ben Kaduk and Rich Salz*
4474 * The distribution now has Makefile.in files, which are used to
4475 create Makefile's when Configure is run. *Configure must be run
4476 before trying to build now.*
4480 * The return value for SSL_CIPHER_description() for error conditions
4485 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4487 Obtaining and performing DNSSEC validation of TLSA records is
4488 the application's responsibility. The application provides
4489 the TLSA records of its choice to OpenSSL, and these are then
4490 used to authenticate the peer.
4492 The TLSA records need not even come from DNS. They can, for
4493 example, be used to implement local end-entity certificate or
4494 trust-anchor "pinning", where the "pin" data takes the form
4495 of TLSA records, which can augment or replace verification
4496 based on the usual WebPKI public certification authorities.
4500 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4501 continues to support deprecated interfaces in default builds.
4502 However, applications are strongly advised to compile their
4503 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4504 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4505 or the 1.1.0 releases.
4507 In environments in which all applications have been ported to
4508 not use any deprecated interfaces OpenSSL's Configure script
4509 should be used with the --api=1.1.0 option to entirely remove
4510 support for the deprecated features from the library and
4511 unconditionally disable them in the installed headers.
4512 Essentially the same effect can be achieved with the "no-deprecated"
4513 argument to Configure, except that this will always restrict
4514 the build to just the latest API, rather than a fixed API
4517 As applications are ported to future revisions of the API,
4518 they should update their compile-time OPENSSL_API_COMPAT define
4519 accordingly, but in most cases should be able to continue to
4520 compile with later releases.
4522 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4523 0x10000000L and 0x00908000L, respectively. However those
4524 versions did not support the OPENSSL_API_COMPAT feature, and
4525 so applications are not typically tested for explicit support
4526 of just the undeprecated features of either release.
4530 * Add support for setting the minimum and maximum supported protocol.
4531 It can bet set via the SSL_set_min_proto_version() and
4532 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4533 MaxProtocol. It's recommended to use the new APIs to disable
4534 protocols instead of disabling individual protocols using
4535 SSL_set_options() or SSL_CONF's Protocol. This change also
4536 removes support for disabling TLS 1.2 in the OpenSSL TLS
4537 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4541 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4545 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4546 and integrates ECDSA and ECDH functionality into EC. Implementations can
4547 now redirect key generation and no longer need to convert to or from
4550 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4551 include the ec.h header file instead.
4555 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4556 ciphers who are no longer supported and drops support the ephemeral RSA key
4557 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4561 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4562 opaque. For HMAC_CTX, the following constructors and destructors
4565 HMAC_CTX *HMAC_CTX_new(void);
4566 void HMAC_CTX_free(HMAC_CTX *ctx);
4568 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4569 destroy such methods has been added. See EVP_MD_meth_new(3) and
4570 EVP_CIPHER_meth_new(3) for documentation.
4573 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4574 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4575 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4576 an already created structure.
4577 2) For consistency with the majority of our object creators and
4578 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4579 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4580 for deprecated builds.
4584 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4585 cryptographic operations to be performed asynchronously as long as an
4586 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4587 further details. Libssl has also had this capability integrated with the
4588 introduction of the new mode SSL_MODE_ASYNC and associated error
4589 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4590 pages. This work was developed in partnership with Intel Corp.
4594 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4595 always enabled now. If you want to disable the support you should
4596 exclude it using the list of supported ciphers. This also means that the
4597 "-no_ecdhe" option has been removed from s_server.
4601 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4602 SSL_{CTX_}set1_curves() which can set a list.
4606 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4607 curve you want to support using SSL_{CTX_}set1_curves().
4611 * State machine rewrite. The state machine code has been significantly
4612 refactored in order to remove much duplication of code and solve issues
4613 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4614 further details). This change does have some associated API changes.
4615 Notably the SSL_state() function has been removed and replaced by
4616 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4617 SSL_set_state() has been removed altogether. The previous handshake states
4618 defined in ssl.h and ssl3.h have also been removed.
4622 * All instances of the string "ssleay" in the public API were replaced
4623 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4624 Some error codes related to internal RSA_eay API's were renamed.
4628 * The demo files in crypto/threads were moved to demo/threads.
4632 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4635 *Matt Caswell, Rich Salz*
4637 * New ASN.1 embed macro.
4639 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4640 structure is not allocated: it is part of the parent. That is instead of
4648 This reduces memory fragmentation and make it impossible to accidentally
4649 set a mandatory field to NULL.
4651 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4652 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4653 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4658 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4662 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4663 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4664 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4665 DES and RC4 ciphersuites.
4669 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4670 This changes the decoding behaviour for some invalid messages,
4671 though the change is mostly in the more lenient direction, and
4672 legacy behaviour is preserved as much as possible.
4676 * Fix no-stdio build.
4677 *David Woodhouse <David.Woodhouse@intel.com> and also*
4678 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4680 * New testing framework
4681 The testing framework has been largely rewritten and is now using
4682 perl and the perl modules Test::Harness and an extended variant of
4683 Test::More called OpenSSL::Test to do its work. All test scripts in
4684 test/ have been rewritten into test recipes, and all direct calls to
4685 executables in test/Makefile have become individual recipes using the
4686 simplified testing OpenSSL::Test::Simple.
4688 For documentation on our testing modules, do:
4690 perldoc test/testlib/OpenSSL/Test/Simple.pm
4691 perldoc test/testlib/OpenSSL/Test.pm
4695 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4696 are used; the latter aborts on memory leaks (usually checked on exit).
4697 Some undocumented "set malloc, etc., hooks" functions were removed
4698 and others were changed. All are now documented.
4702 * In DSA_generate_parameters_ex, if the provided seed is too short,
4705 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4707 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4708 from RFC4279, RFC4785, RFC5487, RFC5489.
4710 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4711 original RSA_PSK patch.
4715 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4716 era flag was never set throughout the codebase (only read). Also removed
4717 SSL3_FLAGS_POP_BUFFER which was only used if
4718 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4722 * Changed the default name options in the "ca", "crl", "req" and "x509"
4723 to be "oneline" instead of "compat".
4727 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4728 not aware of clients that still exhibit this bug, and the workaround
4729 hasn't been working properly for a while.
4733 * The return type of BIO_number_read() and BIO_number_written() as well as
4734 the corresponding num_read and num_write members in the BIO structure has
4735 changed from unsigned long to uint64_t. On platforms where an unsigned
4736 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4741 * Given the pervasive nature of TLS extensions it is inadvisable to run
4742 OpenSSL without support for them. It also means that maintaining
4743 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4744 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4748 * Removed support for the two export grade static DH ciphersuites
4749 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4750 were newly added (along with a number of other static DH ciphersuites) to
4751 1.0.2. However the two export ones have *never* worked since they were
4752 introduced. It seems strange in any case to be adding new export
4753 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4757 * Version negotiation has been rewritten. In particular SSLv23_method(),
4758 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4759 and turned into macros which simply call the new preferred function names
4760 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4761 should use the new names instead. Also as part of this change the ssl23.h
4762 header file has been removed.
4766 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4767 code and the associated standard is no longer considered fit-for-purpose.
4771 * RT2547 was closed. When generating a private key, try to make the
4772 output file readable only by the owner. This behavior change might
4773 be noticeable when interacting with other software.
4775 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4780 * Added HTTP GET support to the ocsp command.
4784 * Changed default digest for the dgst and enc commands from MD5 to
4789 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4793 * Added support for TLS extended master secret from
4794 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4795 initial patch which was a great help during development.
4799 * All libssl internal structures have been removed from the public header
4800 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4801 now redundant). Users should not attempt to access internal structures
4802 directly. Instead they should use the provided API functions.
4806 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4807 Access to deprecated functions can be re-enabled by running config with
4808 "enable-deprecated". In addition applications wishing to use deprecated
4809 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4810 will, by default, disable some transitive includes that previously existed
4811 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4815 * Added support for OCB mode. OpenSSL has been granted a patent license
4816 compatible with the OpenSSL license for use of OCB. Details are available
4817 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4818 for OCB can be removed by calling config with no-ocb.
4822 * SSLv2 support has been removed. It still supports receiving an SSLv2
4823 compatible client hello.
4827 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4828 done while fixing the error code for the key-too-small case.
4830 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4832 * CA.sh has been removed; use CA.pl instead.
4836 * Removed old DES API.
4840 * Remove various unsupported platforms:
4846 Sinix/ReliantUNIX RM400
4851 16-bit platforms such as WIN16
4855 * Clean up OPENSSL_NO_xxx #define's
4856 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4857 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4858 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4859 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4860 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4861 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4862 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4863 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4864 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4865 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4869 * Cleaned up dead code
4870 Remove all but one '#ifdef undef' which is to be looked at.
4874 * Clean up calling of xxx_free routines.
4875 Just like free(), fix most of the xxx_free routines to accept
4876 NULL. Remove the non-null checks from callers. Save much code.
4880 * Add secure heap for storage of private keys (when possible).
4881 Add BIO_s_secmem(), CBIGNUM, etc.
4882 Contributed by Akamai Technologies under our Corporate CLA.
4886 * Experimental support for a new, fast, unbiased prime candidate generator,
4887 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4889 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4891 * New output format NSS in the sess_id command line tool. This allows
4892 exporting the session id and the master key in NSS keylog format.
4894 *Martin Kaiser <martin@kaiser.cx>*
4896 * Harmonize version and its documentation. -f flag is used to display
4899 *mancha <mancha1@zoho.com>*
4901 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4902 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4904 *mancha <mancha1@zoho.com>*
4906 * Fix some double frees. These are not thought to be exploitable.
4908 *mancha <mancha1@zoho.com>*
4910 * A missing bounds check in the handling of the TLS heartbeat extension
4911 can be used to reveal up to 64k of memory to a connected client or
4914 Thanks for Neel Mehta of Google Security for discovering this bug and to
4915 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4916 preparing the fix ([CVE-2014-0160])
4918 *Adam Langley, Bodo Moeller*
4920 * Fix for the attack described in the paper "Recovering OpenSSL
4921 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4922 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4923 <http://eprint.iacr.org/2014/140>
4925 Thanks to Yuval Yarom and Naomi Benger for discovering this
4926 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4928 *Yuval Yarom and Naomi Benger*
4930 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4931 this fixes a limitation in previous versions of OpenSSL.
4935 * Experimental encrypt-then-mac support.
4937 Experimental support for encrypt then mac from
4938 draft-gutmann-tls-encrypt-then-mac-02.txt
4940 To enable it set the appropriate extension number (0x42 for the test
4941 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4943 For non-compliant peers (i.e. just about everything) this should have no
4946 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4950 * Add EVP support for key wrapping algorithms, to avoid problems with
4951 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4952 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4953 algorithms and include tests cases.
4957 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4962 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4963 MGF1 digest and OAEP label.
4967 * Make openssl verify return errors.
4969 *Chris Palmer <palmer@google.com> and Ben Laurie*
4971 * New function ASN1_TIME_diff to calculate the difference between two
4972 ASN1_TIME structures or one structure and the current time.
4976 * Update fips_test_suite to support multiple command line options. New
4977 test to induce all self test errors in sequence and check expected
4982 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4983 sign or verify all in one operation.
4987 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4988 test programs and fips_test_suite. Includes functionality to parse
4989 the minimal script output of fipsalgest.pl directly.
4993 * Add authorisation parameter to FIPS_module_mode_set().
4997 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5001 * Use separate DRBG fields for internal and external flags. New function
5002 FIPS_drbg_health_check() to perform on demand health checking. Add
5003 generation tests to fips_test_suite with reduced health check interval to
5004 demonstrate periodic health checking. Add "nodh" option to
5005 fips_test_suite to skip very slow DH test.
5009 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5014 * More extensive health check for DRBG checking many more failure modes.
5015 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5016 combination: call this in fips_test_suite.
5020 * Add support for canonical generation of DSA parameter 'g'. See
5023 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5024 POST to handle HMAC cases.
5028 * Add functions FIPS_module_version() and FIPS_module_version_text()
5029 to return numerical and string versions of the FIPS module number.
5033 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5034 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5035 outside the validated module in the FIPS capable OpenSSL.
5039 * Minor change to DRBG entropy callback semantics. In some cases
5040 there is no multiple of the block length between min_len and
5041 max_len. Allow the callback to return more than max_len bytes
5042 of entropy but discard any extra: it is the callback's responsibility
5043 to ensure that the extra data discarded does not impact the
5044 requested amount of entropy.
5048 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5049 information in FIPS186-3, SP800-57 and SP800-131A.
5053 * CCM support via EVP. Interface is very similar to GCM case except we
5054 must supply all data in one chunk (i.e. no update, final) and the
5055 message length must be supplied if AAD is used. Add algorithm test
5060 * Initial version of POST overhaul. Add POST callback to allow the status
5061 of POST to be monitored and/or failures induced. Modify fips_test_suite
5062 to use callback. Always run all selftests even if one fails.
5066 * XTS support including algorithm test driver in the fips_gcmtest program.
5067 Note: this does increase the maximum key length from 32 to 64 bytes but
5068 there should be no binary compatibility issues as existing applications
5069 will never use XTS mode.
5073 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5074 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5075 performs algorithm blocking for unapproved PRNG types. Also do not
5076 set PRNG type in FIPS_mode_set(): leave this to the application.
5077 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5078 the standard OpenSSL PRNG: set additional data to a date time vector.
5082 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5083 This shouldn't present any incompatibility problems because applications
5084 shouldn't be using these directly and any that are will need to rethink
5085 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5089 * Extensive self tests and health checking required by SP800-90 DRBG.
5090 Remove strength parameter from FIPS_drbg_instantiate and always
5091 instantiate at maximum supported strength.
5095 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5099 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5103 * New function DH_compute_key_padded() to compute a DH key and pad with
5104 leading zeroes if needed: this complies with SP800-56A et al.
5108 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5109 anything, incomplete, subject to change and largely untested at present.
5113 * Modify fipscanisteronly build option to only build the necessary object
5114 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5118 * Add experimental option FIPSSYMS to give all symbols in
5119 fipscanister.o and FIPS or fips prefix. This will avoid
5120 conflicts with future versions of OpenSSL. Add perl script
5121 util/fipsas.pl to preprocess assembly language source files
5122 and rename any affected symbols.
5126 * Add selftest checks and algorithm block of non-fips algorithms in
5127 FIPS mode. Remove DES2 from selftests.
5131 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5132 return internal method without any ENGINE dependencies. Add new
5133 tiny fips sign and verify functions.
5137 * New build option no-ec2m to disable characteristic 2 code.
5141 * New build option "fipscanisteronly". This only builds fipscanister.o
5142 and (currently) associated fips utilities. Uses the file Makefile.fips
5143 instead of Makefile.org as the prototype.
5147 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5148 Update fips_gcmtest to use IV generator.
5152 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5153 setting output buffer to NULL. The `*Final` function must be
5154 called although it will not retrieve any additional data. The tag
5155 can be set or retrieved with a ctrl. The IV length is by default 12
5156 bytes (96 bits) but can be set to an alternative value. If the IV
5157 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5162 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5163 underlying do_cipher function handles all cipher semantics itself
5164 including padding and finalisation. This is useful if (for example)
5165 an ENGINE cipher handles block padding itself. The behaviour of
5166 do_cipher is subtly changed if this flag is set: the return value
5167 is the number of characters written to the output buffer (zero is
5168 no longer an error code) or a negative error code. Also if the
5169 input buffer is NULL and length 0 finalisation should be performed.
5173 * If a candidate issuer certificate is already part of the constructed
5174 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5178 * Improve forward-security support: add functions
5180 void SSL_CTX_set_not_resumable_session_callback(
5181 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5182 void SSL_set_not_resumable_session_callback(
5183 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5185 for use by SSL/TLS servers; the callback function will be called whenever a
5186 new session is created, and gets to decide whether the session may be
5187 cached to make it resumable (return 0) or not (return 1). (As by the
5188 SSL/TLS protocol specifications, the session_id sent by the server will be
5189 empty to indicate that the session is not resumable; also, the server will
5190 not generate RFC 4507 (RFC 5077) session tickets.)
5192 A simple reasonable callback implementation is to return is_forward_secure.
5193 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5194 by the SSL/TLS server library, indicating whether it can provide forward
5197 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5199 * New -verify_name option in command line utilities to set verification
5204 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5205 Add CMAC pkey methods.
5209 * Experimental renegotiation in s_server -www mode. If the client
5210 browses /reneg connection is renegotiated. If /renegcert it is
5211 renegotiated requesting a certificate.
5215 * Add an "external" session cache for debugging purposes to s_server. This
5216 should help trace issues which normally are only apparent in deployed
5217 multi-process servers.
5221 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5222 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5223 BIO_set_cipher() and some obscure PEM functions were changed so they
5224 can now return an error. The RAND changes required a change to the
5225 RAND_METHOD structure.
5229 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5230 a gcc attribute to warn if the result of a function is ignored. This
5231 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5232 whose return value is often ignored.
5236 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5237 These allow SCTs (signed certificate timestamps) to be requested and
5238 validated when establishing a connection.
5240 *Rob Percival <robpercival@google.com>*
5245 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5247 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5248 used even when parsing explicit parameters, when loading a encoded key
5249 or calling `EC_GROUP_new_from_ecpkparameters()`/
5250 `EC_GROUP_new_from_ecparameters()`.
5251 This prevents bypass of security hardening and performance gains,
5252 especially for curves with specialized EC_METHODs.
5253 By default, if a key encoded with explicit parameters is loaded and later
5254 encoded, the output is still encoded with explicit parameters, even if
5255 internally a "named" EC_GROUP is used for computation.
5259 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5260 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5261 NULL. After this change, only the cofactor parameter can be NULL. It also
5262 does some minimal sanity checks on the passed order.
5267 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5268 An attack is simple, if the first CMS_recipientInfo is valid but the
5269 second CMS_recipientInfo is chosen ciphertext. If the second
5270 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5271 encryption key will be replaced by garbage, and the message cannot be
5272 decoded, but if the RSA decryption fails, the correct encryption key is
5273 used and the recipient will not notice the attack.
5274 As a work around for this potential attack the length of the decrypted
5275 key must be equal to the cipher default key length, in case the
5276 certificate is not given and all recipientInfo are tried out.
5277 The old behaviour can be re-enabled in the CMS code by setting the
5278 CMS_DEBUG_DECRYPT flag.
5283 * Document issue with installation paths in diverse Windows builds
5285 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5286 binaries and run-time config file.
5291 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5293 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5294 This changes the size when using the `genpkey` command when no size is given.
5295 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5296 generation commands to use 2048 bits by default.
5300 * Add FIPS support for Android Arm 64-bit
5302 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5303 Module in Version 2.0.10. For some reason, the corresponding target
5304 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5305 built with FIPS support on Android Arm 64-bit. This omission has been
5308 *Matthias St. Pierre*
5310 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5312 * 0-byte record padding oracle
5314 If an application encounters a fatal protocol error and then calls
5315 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5316 then OpenSSL can respond differently to the calling application if a 0 byte
5317 record is received with invalid padding compared to if a 0 byte record is
5318 received with an invalid MAC. If the application then behaves differently
5319 based on that in a way that is detectable to the remote peer, then this
5320 amounts to a padding oracle that could be used to decrypt data.
5322 In order for this to be exploitable "non-stitched" ciphersuites must be in
5323 use. Stitched ciphersuites are optimised implementations of certain
5324 commonly used ciphersuites. Also the application must call SSL_shutdown()
5325 twice even if a protocol error has occurred (applications should not do
5326 this but some do anyway).
5328 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5329 Aviram, with additional investigation by Steven Collison and Andrew
5330 Hourselt. It was reported to OpenSSL on 10th December 2018.
5335 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5339 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5341 * Microarchitecture timing vulnerability in ECC scalar multiplication
5343 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5344 shown to be vulnerable to a microarchitecture timing side channel attack.
5345 An attacker with sufficient access to mount local timing attacks during
5346 ECDSA signature generation could recover the private key.
5348 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5349 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5355 * Timing vulnerability in DSA signature generation
5357 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5358 timing side channel attack. An attacker could use variations in the signing
5359 algorithm to recover the private key.
5361 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5366 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5367 Module, accidentally introduced while backporting security fixes from the
5368 development branch and hindering the use of ECC in FIPS mode.
5372 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5374 * Client DoS due to large DH parameter
5376 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5377 malicious server can send a very large prime value to the client. This will
5378 cause the client to spend an unreasonably long period of time generating a
5379 key for this prime resulting in a hang until the client has finished. This
5380 could be exploited in a Denial Of Service attack.
5382 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5387 * Cache timing vulnerability in RSA Key Generation
5389 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5390 a cache timing side channel attack. An attacker with sufficient access to
5391 mount cache timing attacks during the RSA key generation process could
5392 recover the private key.
5394 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5395 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5400 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5401 parameter is no longer accepted, as it leads to a corrupt table. NULL
5402 pem_str is reserved for alias entries only.
5406 * Revert blinding in ECDSA sign and instead make problematic addition
5407 length-invariant. Switch even to fixed-length Montgomery multiplication.
5411 * Change generating and checking of primes so that the error rate of not
5412 being prime depends on the intended use based on the size of the input.
5413 For larger primes this will result in more rounds of Miller-Rabin.
5414 The maximal error rate for primes with more than 1080 bits is lowered
5417 *Kurt Roeckx, Annie Yousar*
5419 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5423 * Add blinding to ECDSA and DSA signatures to protect against side channel
5424 attacks discovered by Keegan Ryan (NCC Group).
5428 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5429 now allow empty (zero character) pass phrases.
5433 * Certificate time validation (X509_cmp_time) enforces stricter
5434 compliance with RFC 5280. Fractional seconds and timezone offsets
5435 are no longer allowed.
5439 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5441 * Constructed ASN.1 types with a recursive definition could exceed the stack
5443 Constructed ASN.1 types with a recursive definition (such as can be found
5444 in PKCS7) could eventually exceed the stack given malicious input with
5445 excessive recursion. This could result in a Denial Of Service attack. There
5446 are no such structures used within SSL/TLS that come from untrusted sources
5447 so this is considered safe.
5449 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5455 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5457 * Read/write after SSL object in error state
5459 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5460 mechanism. The intent was that if a fatal error occurred during a handshake
5461 then OpenSSL would move into the error state and would immediately fail if
5462 you attempted to continue the handshake. This works as designed for the
5463 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5464 SSL_connect()), however due to a bug it does not work correctly if
5465 SSL_read() or SSL_write() is called directly. In that scenario, if the
5466 handshake fails then a fatal error will be returned in the initial function
5467 call. If SSL_read()/SSL_write() is subsequently called by the application
5468 for the same SSL object then it will succeed and the data is passed without
5469 being decrypted/encrypted directly from the SSL/TLS record layer.
5471 In order to exploit this issue an application bug would have to be present
5472 that resulted in a call to SSL_read()/SSL_write() being issued after having
5473 already received a fatal error.
5475 This issue was reported to OpenSSL by David Benjamin (Google).
5480 * rsaz_1024_mul_avx2 overflow bug on x86_64
5482 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5483 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5484 Analysis suggests that attacks against RSA and DSA as a result of this
5485 defect would be very difficult to perform and are not believed likely.
5486 Attacks against DH1024 are considered just feasible, because most of the
5487 work necessary to deduce information about a private key may be performed
5488 offline. The amount of resources required for such an attack would be
5489 significant. However, for an attack on TLS to be meaningful, the server
5490 would have to share the DH1024 private key among multiple clients, which is
5491 no longer an option since CVE-2016-0701.
5493 This only affects processors that support the AVX2 but not ADX extensions
5494 like Intel Haswell (4th generation).
5496 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5497 was originally found via the OSS-Fuzz project.
5502 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5504 * bn_sqrx8x_internal carry bug on x86_64
5506 There is a carry propagating bug in the x86_64 Montgomery squaring
5507 procedure. No EC algorithms are affected. Analysis suggests that attacks
5508 against RSA and DSA as a result of this defect would be very difficult to
5509 perform and are not believed likely. Attacks against DH are considered just
5510 feasible (although very difficult) because most of the work necessary to
5511 deduce information about a private key may be performed offline. The amount
5512 of resources required for such an attack would be very significant and
5513 likely only accessible to a limited number of attackers. An attacker would
5514 additionally need online access to an unpatched system using the target
5515 private key in a scenario with persistent DH parameters and a private
5516 key that is shared between multiple clients.
5518 This only affects processors that support the BMI1, BMI2 and ADX extensions
5519 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5521 This issue was reported to OpenSSL by the OSS-Fuzz project.
5526 * Malformed X.509 IPAddressFamily could cause OOB read
5528 If an X.509 certificate has a malformed IPAddressFamily extension,
5529 OpenSSL could do a one-byte buffer overread. The most likely result
5530 would be an erroneous display of the certificate in text format.
5532 This issue was reported to OpenSSL by the OSS-Fuzz project.
5536 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5538 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5539 platform rather than 'mingw'.
5543 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5545 * Truncated packet could crash via OOB read
5547 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5548 cipher is being used, then a truncated packet can cause that host to
5549 perform an out-of-bounds read, usually resulting in a crash.
5551 This issue was reported to OpenSSL by Robert Święcki of Google.
5556 * BN_mod_exp may produce incorrect results on x86_64
5558 There is a carry propagating bug in the x86_64 Montgomery squaring
5559 procedure. No EC algorithms are affected. Analysis suggests that attacks
5560 against RSA and DSA as a result of this defect would be very difficult to
5561 perform and are not believed likely. Attacks against DH are considered just
5562 feasible (although very difficult) because most of the work necessary to
5563 deduce information about a private key may be performed offline. The amount
5564 of resources required for such an attack would be very significant and
5565 likely only accessible to a limited number of attackers. An attacker would
5566 additionally need online access to an unpatched system using the target
5567 private key in a scenario with persistent DH parameters and a private
5568 key that is shared between multiple clients. For example this can occur by
5569 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5570 similar to CVE-2015-3193 but must be treated as a separate problem.
5572 This issue was reported to OpenSSL by the OSS-Fuzz project.
5577 * Montgomery multiplication may produce incorrect results
5579 There is a carry propagating bug in the Broadwell-specific Montgomery
5580 multiplication procedure that handles input lengths divisible by, but
5581 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5582 and DH private keys are impossible. This is because the subroutine in
5583 question is not used in operations with the private key itself and an input
5584 of the attacker's direct choice. Otherwise the bug can manifest itself as
5585 transient authentication and key negotiation failures or reproducible
5586 erroneous outcome of public-key operations with specially crafted input.
5587 Among EC algorithms only Brainpool P-512 curves are affected and one
5588 presumably can attack ECDH key negotiation. Impact was not analyzed in
5589 detail, because pre-requisites for attack are considered unlikely. Namely
5590 multiple clients have to choose the curve in question and the server has to
5591 share the private key among them, neither of which is default behaviour.
5592 Even then only clients that chose the curve will be affected.
5594 This issue was publicly reported as transient failures and was not
5595 initially recognized as a security issue. Thanks to Richard Morgan for
5596 providing reproducible case.
5601 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5602 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5603 prevent issues where no progress is being made and the peer continually
5604 sends unrecognised record types, using up resources processing them.
5608 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5610 * Missing CRL sanity check
5612 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5613 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5614 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5616 This issue only affects the OpenSSL 1.0.2i
5621 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5623 * OCSP Status Request extension unbounded memory growth
5625 A malicious client can send an excessively large OCSP Status Request
5626 extension. If that client continually requests renegotiation, sending a
5627 large OCSP Status Request extension each time, then there will be unbounded
5628 memory growth on the server. This will eventually lead to a Denial Of
5629 Service attack through memory exhaustion. Servers with a default
5630 configuration are vulnerable even if they do not support OCSP. Builds using
5631 the "no-ocsp" build time option are not affected.
5633 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5638 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5641 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5647 * OOB write in MDC2_Update()
5649 An overflow can occur in MDC2_Update() either if called directly or
5650 through the EVP_DigestUpdate() function using MDC2. If an attacker
5651 is able to supply very large amounts of input data after a previous
5652 call to EVP_EncryptUpdate() with a partial block then a length check
5653 can overflow resulting in a heap corruption.
5655 The amount of data needed is comparable to SIZE_MAX which is impractical
5658 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5663 * Malformed SHA512 ticket DoS
5665 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5666 DoS attack where a malformed ticket will result in an OOB read which will
5669 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5670 a custom server callback and ticket lookup mechanism.
5672 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5677 * OOB write in BN_bn2dec()
5679 The function BN_bn2dec() does not check the return value of BN_div_word().
5680 This can cause an OOB write if an application uses this function with an
5681 overly large BIGNUM. This could be a problem if an overly large certificate
5682 or CRL is printed out from an untrusted source. TLS is not affected because
5683 record limits will reject an oversized certificate before it is parsed.
5685 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5690 * OOB read in TS_OBJ_print_bio()
5692 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5693 the total length the OID text representation would use and not the amount
5694 of data written. This will result in OOB reads when large OIDs are
5697 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5702 * Pointer arithmetic undefined behaviour
5704 Avoid some undefined pointer arithmetic
5706 A common idiom in the codebase is to check limits in the following manner:
5709 Where "p" points to some malloc'd data of SIZE bytes and
5712 "len" here could be from some externally supplied data (e.g. from a TLS
5715 The rules of C pointer arithmetic are such that "p + len" is only well
5716 defined where len <= SIZE. Therefore the above idiom is actually
5717 undefined behaviour.
5719 For example this could cause problems if some malloc implementation
5720 provides an address for "p" such that "p + len" actually overflows for
5721 values of len that are too big and therefore p + len < limit.
5723 This issue was reported to OpenSSL by Guido Vranken
5728 * Constant time flag not preserved in DSA signing
5730 Operations in the DSA signing algorithm should run in constant time in
5731 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5732 implementation means that a non-constant time codepath is followed for
5733 certain operations. This has been demonstrated through a cache-timing
5734 attack to be sufficient for an attacker to recover the private DSA key.
5736 This issue was reported by César Pereida (Aalto University), Billy Brumley
5737 (Tampere University of Technology), and Yuval Yarom (The University of
5738 Adelaide and NICTA).
5743 * DTLS buffered message DoS
5745 In a DTLS connection where handshake messages are delivered out-of-order
5746 those messages that OpenSSL is not yet ready to process will be buffered
5747 for later use. Under certain circumstances, a flaw in the logic means that
5748 those messages do not get removed from the buffer even though the handshake
5749 has been completed. An attacker could force up to approx. 15 messages to
5750 remain in the buffer when they are no longer required. These messages will
5751 be cleared when the DTLS connection is closed. The default maximum size for
5752 a message is 100k. Therefore, the attacker could force an additional 1500k
5753 to be consumed per connection. By opening many simultaneous connections an
5754 attacker could cause a DoS attack through memory exhaustion.
5756 This issue was reported to OpenSSL by Quan Luo.
5761 * DTLS replay protection DoS
5763 A flaw in the DTLS replay attack protection mechanism means that records
5764 that arrive for future epochs update the replay protection "window" before
5765 the MAC for the record has been validated. This could be exploited by an
5766 attacker by sending a record for the next epoch (which does not have to
5767 decrypt or have a valid MAC), with a very large sequence number. This means
5768 that all subsequent legitimate packets are dropped causing a denial of
5769 service for a specific DTLS connection.
5771 This issue was reported to OpenSSL by the OCAP audit team.
5776 * Certificate message OOB reads
5778 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5779 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5780 theoretical DoS risk but this has not been observed in practice on common
5783 The messages affected are client certificate, client certificate request
5784 and server certificate. As a result the attack can only be performed
5785 against a client or a server which enables client authentication.
5787 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5792 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5794 * Prevent padding oracle in AES-NI CBC MAC check
5796 A MITM attacker can use a padding oracle attack to decrypt traffic
5797 when the connection uses an AES CBC cipher and the server support
5800 This issue was introduced as part of the fix for Lucky 13 padding
5801 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5802 constant time by making sure that always the same bytes are read and
5803 compared against either the MAC or padding bytes. But it no longer
5804 checked that there was enough data to have both the MAC and padding
5807 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5811 * Fix EVP_EncodeUpdate overflow
5813 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5814 Base64 encoding of binary data. If an attacker is able to supply very large
5815 amounts of input data then a length check can overflow resulting in a heap
5818 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5819 the `PEM_write_bio*` family of functions. These are mainly used within the
5820 OpenSSL command line applications, so any application which processes data
5821 from an untrusted source and outputs it as a PEM file should be considered
5822 vulnerable to this issue. User applications that call these APIs directly
5823 with large amounts of untrusted data may also be vulnerable.
5825 This issue was reported by Guido Vranken.
5830 * Fix EVP_EncryptUpdate overflow
5832 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5833 is able to supply very large amounts of input data after a previous call to
5834 EVP_EncryptUpdate() with a partial block then a length check can overflow
5835 resulting in a heap corruption. Following an analysis of all OpenSSL
5836 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5837 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5838 the first called function after an EVP_EncryptInit(), and therefore that
5839 specific call must be safe. The second form is where the length passed to
5840 EVP_EncryptUpdate() can be seen from the code to be some small value and
5841 therefore there is no possibility of an overflow. Since all instances are
5842 one of these two forms, it is believed that there can be no overflows in
5843 internal code due to this problem. It should be noted that
5844 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5845 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5846 of these calls have also been analysed too and it is believed there are no
5847 instances in internal usage where an overflow could occur.
5849 This issue was reported by Guido Vranken.
5854 * Prevent ASN.1 BIO excessive memory allocation
5856 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5857 a short invalid encoding can cause allocation of large amounts of memory
5858 potentially consuming excessive resources or exhausting memory.
5860 Any application parsing untrusted data through d2i BIO functions is
5861 affected. The memory based functions such as d2i_X509() are *not* affected.
5862 Since the memory based functions are used by the TLS library, TLS
5863 applications are not affected.
5865 This issue was reported by Brian Carpenter.
5872 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5873 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5874 in arbitrary stack data being returned in the buffer.
5876 This issue was reported by Guido Vranken.
5881 * Modify behavior of ALPN to invoke callback after SNI/servername
5882 callback, such that updates to the SSL_CTX affect ALPN.
5886 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5891 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5892 methods are enabled and ssl2 is disabled the methods return NULL.
5896 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5898 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5899 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5900 provide any "EXPORT" or "LOW" strength ciphers.
5904 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5905 is by default disabled at build-time. Builds that are not configured with
5906 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5907 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5908 will need to explicitly call either of:
5910 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5912 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5914 as appropriate. Even if either of those is used, or the application
5915 explicitly uses the version-specific SSLv2_method() or its client and
5916 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5917 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5918 ciphers, and SSLv2 56-bit DES are no longer available.
5923 * Fix a double-free in DSA code
5925 A double free bug was discovered when OpenSSL parses malformed DSA private
5926 keys and could lead to a DoS attack or memory corruption for applications
5927 that receive DSA private keys from untrusted sources. This scenario is
5930 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5936 * Disable SRP fake user seed to address a server memory leak.
5938 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5940 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5941 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5942 was changed to ignore the "fake user" SRP seed, even if the seed
5945 Users should use SRP_VBASE_get1_by_user instead. Note that in
5946 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5947 also that even though configuring the SRP seed attempts to hide
5948 invalid usernames by continuing the handshake with fake
5949 credentials, this behaviour is not constant time and no strong
5950 guarantees are made that the handshake is indistinguishable from
5951 that of a valid user.
5956 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5958 In the BN_hex2bn function the number of hex digits is calculated using an
5959 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5960 large values of `i` this can result in `bn_expand` not allocating any
5961 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5962 field as NULL leading to a subsequent NULL ptr deref. For very large values
5963 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5964 In this case memory is allocated to the internal BIGNUM data field, but it
5965 is insufficiently sized leading to heap corruption. A similar issue exists
5966 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5967 is ever called by user applications with very large untrusted hex/dec data.
5968 This is anticipated to be a rare occurrence.
5970 All OpenSSL internal usage of these functions use data that is not expected
5971 to be untrusted, e.g. config file data or application command line
5972 arguments. If user developed applications generate config file data based
5973 on untrusted data then it is possible that this could also lead to security
5974 consequences. This is also anticipated to be rare.
5976 This issue was reported to OpenSSL by Guido Vranken.
5981 * Fix memory issues in `BIO_*printf` functions
5983 The internal `fmtstr` function used in processing a "%s" format string in
5984 the `BIO_*printf` functions could overflow while calculating the length of a
5985 string and cause an OOB read when printing very long strings.
5987 Additionally the internal `doapr_outch` function can attempt to write to an
5988 OOB memory location (at an offset from the NULL pointer) in the event of a
5989 memory allocation failure. In 1.0.2 and below this could be caused where
5990 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5991 could be in processing a very long "%s" format string. Memory leaks can
5994 The first issue may mask the second issue dependent on compiler behaviour.
5995 These problems could enable attacks where large amounts of untrusted data
5996 is passed to the `BIO_*printf` functions. If applications use these functions
5997 in this way then they could be vulnerable. OpenSSL itself uses these
5998 functions when printing out human-readable dumps of ASN.1 data. Therefore
5999 applications that print this data could be vulnerable if the data is from
6000 untrusted sources. OpenSSL command line applications could also be
6001 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6002 as command line arguments.
6004 Libssl is not considered directly vulnerable. Additionally certificates etc
6005 received via remote connections via libssl are also unlikely to be able to
6006 trigger these issues because of message size limits enforced within libssl.
6008 This issue was reported to OpenSSL Guido Vranken.
6013 * Side channel attack on modular exponentiation
6015 A side-channel attack was found which makes use of cache-bank conflicts on
6016 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6017 of RSA keys. The ability to exploit this issue is limited as it relies on
6018 an attacker who has control of code in a thread running on the same
6019 hyper-threaded core as the victim thread which is performing decryptions.
6021 This issue was reported to OpenSSL by Yuval Yarom, The University of
6022 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6023 Nadia Heninger, University of Pennsylvania with more information at
6024 <http://cachebleed.info>.
6029 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6030 if no keysize is specified with default_bits. This fixes an
6031 omission in an earlier change that changed all RSA/DSA key generation
6032 commands to use 2048 bits by default.
6036 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6038 * DH small subgroups
6040 Historically OpenSSL only ever generated DH parameters based on "safe"
6041 primes. More recently (in version 1.0.2) support was provided for
6042 generating X9.42 style parameter files such as those required for RFC 5114
6043 support. The primes used in such files may not be "safe". Where an
6044 application is using DH configured with parameters based on primes that are
6045 not "safe" then an attacker could use this fact to find a peer's private
6046 DH exponent. This attack requires that the attacker complete multiple
6047 handshakes in which the peer uses the same private DH exponent. For example
6048 this could be used to discover a TLS server's private DH exponent if it's
6049 reusing the private DH exponent or it's using a static DH ciphersuite.
6051 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6052 TLS. It is not on by default. If the option is not set then the server
6053 reuses the same private DH exponent for the life of the server process and
6054 would be vulnerable to this attack. It is believed that many popular
6055 applications do set this option and would therefore not be at risk.
6057 The fix for this issue adds an additional check where a "q" parameter is
6058 available (as is the case in X9.42 based parameters). This detects the
6059 only known attack, and is the only possible defense for static DH
6060 ciphersuites. This could have some performance impact.
6062 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6063 default and cannot be disabled. This could have some performance impact.
6065 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6070 * SSLv2 doesn't block disabled ciphers
6072 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6073 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6074 been disabled, provided that the SSLv2 protocol was not also disabled via
6077 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6078 and Sebastian Schinzel.
6083 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6085 * BN_mod_exp may produce incorrect results on x86_64
6087 There is a carry propagating bug in the x86_64 Montgomery squaring
6088 procedure. No EC algorithms are affected. Analysis suggests that attacks
6089 against RSA and DSA as a result of this defect would be very difficult to
6090 perform and are not believed likely. Attacks against DH are considered just
6091 feasible (although very difficult) because most of the work necessary to
6092 deduce information about a private key may be performed offline. The amount
6093 of resources required for such an attack would be very significant and
6094 likely only accessible to a limited number of attackers. An attacker would
6095 additionally need online access to an unpatched system using the target
6096 private key in a scenario with persistent DH parameters and a private
6097 key that is shared between multiple clients. For example this can occur by
6098 default in OpenSSL DHE based SSL/TLS ciphersuites.
6100 This issue was reported to OpenSSL by Hanno Böck.
6105 * Certificate verify crash with missing PSS parameter
6107 The signature verification routines will crash with a NULL pointer
6108 dereference if presented with an ASN.1 signature using the RSA PSS
6109 algorithm and absent mask generation function parameter. Since these
6110 routines are used to verify certificate signature algorithms this can be
6111 used to crash any certificate verification operation and exploited in a
6112 DoS attack. Any application which performs certificate verification is
6113 vulnerable including OpenSSL clients and servers which enable client
6116 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6121 * X509_ATTRIBUTE memory leak
6123 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6124 memory. This structure is used by the PKCS#7 and CMS routines so any
6125 application which reads PKCS#7 or CMS data from untrusted sources is
6126 affected. SSL/TLS is not affected.
6128 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6134 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6135 This changes the decoding behaviour for some invalid messages,
6136 though the change is mostly in the more lenient direction, and
6137 legacy behaviour is preserved as much as possible.
6141 * In DSA_generate_parameters_ex, if the provided seed is too short,
6144 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6146 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6148 * Alternate chains certificate forgery
6150 During certificate verification, OpenSSL will attempt to find an
6151 alternative certificate chain if the first attempt to build such a chain
6152 fails. An error in the implementation of this logic can mean that an
6153 attacker could cause certain checks on untrusted certificates to be
6154 bypassed, such as the CA flag, enabling them to use a valid leaf
6155 certificate to act as a CA and "issue" an invalid certificate.
6157 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6162 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6164 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6165 incompatibility in the handling of HMAC. The previous ABI has now been
6170 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6172 * Malformed ECParameters causes infinite loop
6174 When processing an ECParameters structure OpenSSL enters an infinite loop
6175 if the curve specified is over a specially malformed binary polynomial
6178 This can be used to perform denial of service against any
6179 system which processes public keys, certificate requests or
6180 certificates. This includes TLS clients and TLS servers with
6181 client authentication enabled.
6183 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6188 * Exploitable out-of-bounds read in X509_cmp_time
6190 X509_cmp_time does not properly check the length of the ASN1_TIME
6191 string and can read a few bytes out of bounds. In addition,
6192 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6195 An attacker can use this to craft malformed certificates and CRLs of
6196 various sizes and potentially cause a segmentation fault, resulting in
6197 a DoS on applications that verify certificates or CRLs. TLS clients
6198 that verify CRLs are affected. TLS clients and servers with client
6199 authentication enabled may be affected if they use custom verification
6202 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6203 independently by Hanno Böck.
6208 * PKCS7 crash with missing EnvelopedContent
6210 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6211 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6212 with missing content and trigger a NULL pointer dereference on parsing.
6214 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6215 structures from untrusted sources are affected. OpenSSL clients and
6216 servers are not affected.
6218 This issue was reported to OpenSSL by Michal Zalewski (Google).
6223 * CMS verify infinite loop with unknown hash function
6225 When verifying a signedData message the CMS code can enter an infinite loop
6226 if presented with an unknown hash function OID. This can be used to perform
6227 denial of service against any system which verifies signedData messages using
6229 This issue was reported to OpenSSL by Johannes Bauer.
6234 * Race condition handling NewSessionTicket
6236 If a NewSessionTicket is received by a multi-threaded client when attempting to
6237 reuse a previous ticket then a race condition can occur potentially leading to
6238 a double free of the ticket data.
6243 * Only support 256-bit or stronger elliptic curves with the
6244 'ecdh_auto' setting (server) or by default (client). Of supported
6245 curves, prefer P-256 (both).
6249 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6251 * ClientHello sigalgs DoS fix
6253 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6254 invalid signature algorithms extension a NULL pointer dereference will
6255 occur. This can be exploited in a DoS attack against the server.
6257 This issue was was reported to OpenSSL by David Ramos of Stanford
6261 *Stephen Henson and Matt Caswell*
6263 * Multiblock corrupted pointer fix
6265 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6266 feature only applies on 64 bit x86 architecture platforms that support AES
6267 NI instructions. A defect in the implementation of "multiblock" can cause
6268 OpenSSL's internal write buffer to become incorrectly set to NULL when
6269 using non-blocking IO. Typically, when the user application is using a
6270 socket BIO for writing, this will only result in a failed connection.
6271 However if some other BIO is used then it is likely that a segmentation
6272 fault will be triggered, thus enabling a potential DoS attack.
6274 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6279 * Segmentation fault in DTLSv1_listen fix
6281 The DTLSv1_listen function is intended to be stateless and processes the
6282 initial ClientHello from many peers. It is common for user code to loop
6283 over the call to DTLSv1_listen until a valid ClientHello is received with
6284 an associated cookie. A defect in the implementation of DTLSv1_listen means
6285 that state is preserved in the SSL object from one invocation to the next
6286 that can lead to a segmentation fault. Errors processing the initial
6287 ClientHello can trigger this scenario. An example of such an error could be
6288 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6291 This issue was reported to OpenSSL by Per Allansson.
6296 * Segmentation fault in ASN1_TYPE_cmp fix
6298 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6299 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6300 certificate signature algorithm consistency this can be used to crash any
6301 certificate verification operation and exploited in a DoS attack. Any
6302 application which performs certificate verification is vulnerable including
6303 OpenSSL clients and servers which enable client authentication.
6308 * Segmentation fault for invalid PSS parameters fix
6310 The signature verification routines will crash with a NULL pointer
6311 dereference if presented with an ASN.1 signature using the RSA PSS
6312 algorithm and invalid parameters. Since these routines are used to verify
6313 certificate signature algorithms this can be used to crash any
6314 certificate verification operation and exploited in a DoS attack. Any
6315 application which performs certificate verification is vulnerable including
6316 OpenSSL clients and servers which enable client authentication.
6318 This issue was was reported to OpenSSL by Brian Carpenter.
6323 * ASN.1 structure reuse memory corruption fix
6325 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6326 memory corruption via an invalid write. Such reuse is and has been
6327 strongly discouraged and is believed to be rare.
6329 Applications that parse structures containing CHOICE or ANY DEFINED BY
6330 components may be affected. Certificate parsing (d2i_X509 and related
6331 functions) are however not affected. OpenSSL clients and servers are
6337 * PKCS7 NULL pointer dereferences fix
6339 The PKCS#7 parsing code does not handle missing outer ContentInfo
6340 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6341 missing content and trigger a NULL pointer dereference on parsing.
6343 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6344 otherwise parse PKCS#7 structures from untrusted sources are
6345 affected. OpenSSL clients and servers are not affected.
6347 This issue was reported to OpenSSL by Michal Zalewski (Google).
6352 * DoS via reachable assert in SSLv2 servers fix
6354 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6355 servers that both support SSLv2 and enable export cipher suites by sending
6356 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6358 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6359 (OpenSSL development team).
6364 * Empty CKE with client auth and DHE fix
6366 If client auth is used then a server can seg fault in the event of a DHE
6367 ciphersuite being selected and a zero length ClientKeyExchange message
6368 being sent by the client. This could be exploited in a DoS attack.
6373 * Handshake with unseeded PRNG fix
6375 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6376 with an unseeded PRNG. The conditions are:
6377 - The client is on a platform where the PRNG has not been seeded
6378 automatically, and the user has not seeded manually
6379 - A protocol specific client method version has been used (i.e. not
6380 SSL_client_methodv23)
6381 - A ciphersuite is used that does not require additional random data from
6382 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6384 If the handshake succeeds then the client random that has been used will
6385 have been generated from a PRNG with insufficient entropy and therefore the
6386 output may be predictable.
6388 For example using the following command with an unseeded openssl will
6389 succeed on an unpatched platform:
6391 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6396 * Use After Free following d2i_ECPrivatekey error fix
6398 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6399 could cause a use after free condition. This, in turn, could cause a double
6400 free in several private key parsing functions (such as d2i_PrivateKey
6401 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6402 for applications that receive EC private keys from untrusted
6403 sources. This scenario is considered rare.
6405 This issue was discovered by the BoringSSL project and fixed in their
6411 * X509_to_X509_REQ NULL pointer deref fix
6413 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6414 the certificate key is invalid. This function is rarely used in practice.
6416 This issue was discovered by Brian Carpenter.
6421 * Removed the export ciphers from the DEFAULT ciphers
6425 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6427 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6428 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6429 So far those who have to target multiple platforms would compromise
6430 and argue that binary targeting say ARMv5 would still execute on
6431 ARMv8. "Universal" build resolves this compromise by providing
6432 near-optimal performance even on newer platforms.
6436 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6437 (other platforms pending).
6439 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6441 * Add support for the SignedCertificateTimestampList certificate and
6442 OCSP response extensions from RFC6962.
6446 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6447 for corner cases. (Certain input points at infinity could lead to
6448 bogus results, with non-infinity inputs mapped to infinity too.)
6452 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6453 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6454 common cases are optimized and there still is room for further
6455 improvements. Vector Permutation AES for Altivec is also added.
6459 * Add support for little-endian ppc64 Linux target.
6461 *Marcelo Cerri (IBM)*
6463 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6464 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6465 are optimized and there still is room for further improvements.
6466 Both 32- and 64-bit modes are supported.
6468 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6470 * Improved ARMv7 NEON support.
6474 * Support for SPARC Architecture 2011 crypto extensions, first
6475 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6476 SHA256/512, MD5, GHASH and modular exponentiation.
6478 *Andy Polyakov, David Miller*
6480 * Accelerated modular exponentiation for Intel processors, a.k.a.
6483 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6485 * Support for new and upcoming Intel processors, including AVX2,
6486 BMI and SHA ISA extensions. This includes additional "stitched"
6487 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6490 This work was sponsored by Intel Corp.
6494 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6495 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6496 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6500 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6501 this fixes a limitation in previous versions of OpenSSL.
6505 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6506 MGF1 digest and OAEP label.
6510 * Add EVP support for key wrapping algorithms, to avoid problems with
6511 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6512 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6513 algorithms and include tests cases.
6517 * Add functions to allocate and set the fields of an ECDSA_METHOD
6520 *Douglas E. Engert, Steve Henson*
6522 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6523 difference in days and seconds between two tm or ASN1_TIME structures.
6527 * Add -rev test option to s_server to just reverse order of characters
6528 received by client and send back to server. Also prints an abbreviated
6529 summary of the connection parameters.
6533 * New option -brief for s_client and s_server to print out a brief summary
6534 of connection parameters.
6538 * Add callbacks for arbitrary TLS extensions.
6540 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6542 * New option -crl_download in several openssl utilities to download CRLs
6543 from CRLDP extension in certificates.
6547 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6551 * New function X509_CRL_diff to generate a delta CRL from the difference
6552 of two full CRLs. Add support to "crl" utility.
6556 * New functions to set lookup_crls function and to retrieve
6557 X509_STORE from X509_STORE_CTX.
6561 * Print out deprecated issuer and subject unique ID fields in
6566 * Extend OCSP I/O functions so they can be used for simple general purpose
6567 HTTP as well as OCSP. New wrapper function which can be used to download
6568 CRLs using the OCSP API.
6572 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6576 * `SSL_CONF*` functions. These provide a common framework for application
6577 configuration using configuration files or command lines.
6581 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6582 message callback and prints the results. Needs compile time option
6583 "enable-ssl-trace". New options to s_client and s_server to enable
6588 * New ctrl and macro to retrieve supported points extensions.
6589 Print out extension in s_server and s_client.
6593 * New functions to retrieve certificate signature and signature
6598 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6603 * New Suite B modes for TLS code. These use and enforce the requirements
6604 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6605 only use Suite B curves. The Suite B modes can be set by using the
6606 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6610 * New chain verification flags for Suite B levels of security. Check
6611 algorithms are acceptable when flags are set in X509_verify_cert.
6615 * Make tls1_check_chain return a set of flags indicating checks passed
6616 by a certificate chain. Add additional tests to handle client
6617 certificates: checks for matching certificate type and issuer name
6622 * If an attempt is made to use a signature algorithm not in the peer
6623 preference list abort the handshake. If client has no suitable
6624 signature algorithms in response to a certificate request do not
6625 use the certificate.
6629 * If server EC tmp key is not in client preference list abort handshake.
6633 * Add support for certificate stores in CERT structure. This makes it
6634 possible to have different stores per SSL structure or one store in
6635 the parent SSL_CTX. Include distinct stores for certificate chain
6636 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6637 to build and store a certificate chain in CERT structure: returning
6638 an error if the chain cannot be built: this will allow applications
6639 to test if a chain is correctly configured.
6641 Note: if the CERT based stores are not set then the parent SSL_CTX
6642 store is used to retain compatibility with existing behaviour.
6646 * New function ssl_set_client_disabled to set a ciphersuite disabled
6647 mask based on the current session, check mask when sending client
6648 hello and checking the requested ciphersuite.
6652 * New ctrls to retrieve and set certificate types in a certificate
6653 request message. Print out received values in s_client. If certificate
6654 types is not set with custom values set sensible values based on
6655 supported signature algorithms.
6659 * Support for distinct client and server supported signature algorithms.
6663 * Add certificate callback. If set this is called whenever a certificate
6664 is required by client or server. An application can decide which
6665 certificate chain to present based on arbitrary criteria: for example
6666 supported signature algorithms. Add very simple example to s_server.
6667 This fixes many of the problems and restrictions of the existing client
6668 certificate callback: for example you can now clear an existing
6669 certificate and specify the whole chain.
6673 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6674 the certificate can be used for (if anything). Set valid_flags field
6675 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6676 to have similar checks in it.
6678 Add new "cert_flags" field to CERT structure and include a "strict mode".
6679 This enforces some TLS certificate requirements (such as only permitting
6680 certificate signature algorithms contained in the supported algorithms
6681 extension) which some implementations ignore: this option should be used
6682 with caution as it could cause interoperability issues.
6686 * Update and tidy signature algorithm extension processing. Work out
6687 shared signature algorithms based on preferences and peer algorithms
6688 and print them out in s_client and s_server. Abort handshake if no
6689 shared signature algorithms.
6693 * Add new functions to allow customised supported signature algorithms
6694 for SSL and SSL_CTX structures. Add options to s_client and s_server
6699 * New function SSL_certs_clear() to delete all references to certificates
6700 from an SSL structure. Before this once a certificate had been added
6701 it couldn't be removed.
6705 * Integrate hostname, email address and IP address checking with certificate
6706 verification. New verify options supporting checking in openssl utility.
6710 * Fixes and wildcard matching support to hostname and email checking
6711 functions. Add manual page.
6713 *Florian Weimer (Red Hat Product Security Team)*
6715 * New functions to check a hostname email or IP address against a
6716 certificate. Add options x509 utility to print results of checks against
6721 * Fix OCSP checking.
6723 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6725 * Initial experimental support for explicitly trusted non-root CAs.
6726 OpenSSL still tries to build a complete chain to a root but if an
6727 intermediate CA has a trust setting included that is used. The first
6728 setting is used: whether to trust (e.g., -addtrust option to the x509
6733 * Add -trusted_first option which attempts to find certificates in the
6734 trusted store even if an untrusted chain is also supplied.
6738 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6739 platform support for Linux and Android.
6743 * Support for linux-x32, ILP32 environment in x86_64 framework.
6747 * Experimental multi-implementation support for FIPS capable OpenSSL.
6748 When in FIPS mode the approved implementations are used as normal,
6749 when not in FIPS mode the internal unapproved versions are used instead.
6750 This means that the FIPS capable OpenSSL isn't forced to use the
6751 (often lower performance) FIPS implementations outside FIPS mode.
6755 * Transparently support X9.42 DH parameters when calling
6756 PEM_read_bio_DHparameters. This means existing applications can handle
6757 the new parameter format automatically.
6761 * Initial experimental support for X9.42 DH parameter format: mainly
6762 to support use of 'q' parameter for RFC5114 parameters.
6766 * Add DH parameters from RFC5114 including test data to dhtest.
6770 * Support for automatic EC temporary key parameter selection. If enabled
6771 the most preferred EC parameters are automatically used instead of
6772 hardcoded fixed parameters. Now a server just has to call:
6773 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6774 support ECDH and use the most appropriate parameters.
6778 * Enhance and tidy EC curve and point format TLS extension code. Use
6779 static structures instead of allocation if default values are used.
6780 New ctrls to set curves we wish to support and to retrieve shared curves.
6781 Print out shared curves in s_server. New options to s_server and s_client
6782 to set list of supported curves.
6786 * New ctrls to retrieve supported signature algorithms and
6787 supported curve values as an array of NIDs. Extend openssl utility
6788 to print out received values.
6792 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6793 between NIDs and the more common NIST names such as "P-256". Enhance
6794 ecparam utility and ECC method to recognise the NIST names for curves.
6798 * Enhance SSL/TLS certificate chain handling to support different
6799 chains for each certificate instead of one chain in the parent SSL_CTX.
6803 * Support for fixed DH ciphersuite client authentication: where both
6804 server and client use DH certificates with common parameters.
6808 * Support for fixed DH ciphersuites: those requiring DH server
6813 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6815 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6816 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6817 X509_CINF_get_signature were reverted post internal team review.
6822 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6824 * OCSP Status Request extension unbounded memory growth
6826 A malicious client can send an excessively large OCSP Status Request
6827 extension. If that client continually requests renegotiation, sending a
6828 large OCSP Status Request extension each time, then there will be unbounded
6829 memory growth on the server. This will eventually lead to a Denial Of
6830 Service attack through memory exhaustion. Servers with a default
6831 configuration are vulnerable even if they do not support OCSP. Builds using
6832 the "no-ocsp" build time option are not affected.
6834 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6839 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6842 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6848 * OOB write in MDC2_Update()
6850 An overflow can occur in MDC2_Update() either if called directly or
6851 through the EVP_DigestUpdate() function using MDC2. If an attacker
6852 is able to supply very large amounts of input data after a previous
6853 call to EVP_EncryptUpdate() with a partial block then a length check
6854 can overflow resulting in a heap corruption.
6856 The amount of data needed is comparable to SIZE_MAX which is impractical
6859 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6864 * Malformed SHA512 ticket DoS
6866 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6867 DoS attack where a malformed ticket will result in an OOB read which will
6870 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6871 a custom server callback and ticket lookup mechanism.
6873 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6878 * OOB write in BN_bn2dec()
6880 The function BN_bn2dec() does not check the return value of BN_div_word().
6881 This can cause an OOB write if an application uses this function with an
6882 overly large BIGNUM. This could be a problem if an overly large certificate
6883 or CRL is printed out from an untrusted source. TLS is not affected because
6884 record limits will reject an oversized certificate before it is parsed.
6886 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6891 * OOB read in TS_OBJ_print_bio()
6893 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6894 the total length the OID text representation would use and not the amount
6895 of data written. This will result in OOB reads when large OIDs are
6898 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6903 * Pointer arithmetic undefined behaviour
6905 Avoid some undefined pointer arithmetic
6907 A common idiom in the codebase is to check limits in the following manner:
6910 Where "p" points to some malloc'd data of SIZE bytes and
6913 "len" here could be from some externally supplied data (e.g. from a TLS
6916 The rules of C pointer arithmetic are such that "p + len" is only well
6917 defined where len <= SIZE. Therefore, the above idiom is actually
6918 undefined behaviour.
6920 For example this could cause problems if some malloc implementation
6921 provides an address for "p" such that "p + len" actually overflows for
6922 values of len that are too big and therefore p + len < limit.
6924 This issue was reported to OpenSSL by Guido Vranken
6929 * Constant time flag not preserved in DSA signing
6931 Operations in the DSA signing algorithm should run in constant time in
6932 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6933 implementation means that a non-constant time codepath is followed for
6934 certain operations. This has been demonstrated through a cache-timing
6935 attack to be sufficient for an attacker to recover the private DSA key.
6937 This issue was reported by César Pereida (Aalto University), Billy Brumley
6938 (Tampere University of Technology), and Yuval Yarom (The University of
6939 Adelaide and NICTA).
6944 * DTLS buffered message DoS
6946 In a DTLS connection where handshake messages are delivered out-of-order
6947 those messages that OpenSSL is not yet ready to process will be buffered
6948 for later use. Under certain circumstances, a flaw in the logic means that
6949 those messages do not get removed from the buffer even though the handshake
6950 has been completed. An attacker could force up to approx. 15 messages to
6951 remain in the buffer when they are no longer required. These messages will
6952 be cleared when the DTLS connection is closed. The default maximum size for
6953 a message is 100k. Therefore, the attacker could force an additional 1500k
6954 to be consumed per connection. By opening many simultaneous connections an
6955 attacker could cause a DoS attack through memory exhaustion.
6957 This issue was reported to OpenSSL by Quan Luo.
6962 * DTLS replay protection DoS
6964 A flaw in the DTLS replay attack protection mechanism means that records
6965 that arrive for future epochs update the replay protection "window" before
6966 the MAC for the record has been validated. This could be exploited by an
6967 attacker by sending a record for the next epoch (which does not have to
6968 decrypt or have a valid MAC), with a very large sequence number. This means
6969 that all subsequent legitimate packets are dropped causing a denial of
6970 service for a specific DTLS connection.
6972 This issue was reported to OpenSSL by the OCAP audit team.
6977 * Certificate message OOB reads
6979 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6980 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6981 theoretical DoS risk but this has not been observed in practice on common
6984 The messages affected are client certificate, client certificate request
6985 and server certificate. As a result the attack can only be performed
6986 against a client or a server which enables client authentication.
6988 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6993 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6995 * Prevent padding oracle in AES-NI CBC MAC check
6997 A MITM attacker can use a padding oracle attack to decrypt traffic
6998 when the connection uses an AES CBC cipher and the server support
7001 This issue was introduced as part of the fix for Lucky 13 padding
7002 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7003 constant time by making sure that always the same bytes are read and
7004 compared against either the MAC or padding bytes. But it no longer
7005 checked that there was enough data to have both the MAC and padding
7008 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7013 * Fix EVP_EncodeUpdate overflow
7015 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7016 Base64 encoding of binary data. If an attacker is able to supply very large
7017 amounts of input data then a length check can overflow resulting in a heap
7020 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7021 the `PEM_write_bio*` family of functions. These are mainly used within the
7022 OpenSSL command line applications, so any application which processes data
7023 from an untrusted source and outputs it as a PEM file should be considered
7024 vulnerable to this issue. User applications that call these APIs directly
7025 with large amounts of untrusted data may also be vulnerable.
7027 This issue was reported by Guido Vranken.
7032 * Fix EVP_EncryptUpdate overflow
7034 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7035 is able to supply very large amounts of input data after a previous call to
7036 EVP_EncryptUpdate() with a partial block then a length check can overflow
7037 resulting in a heap corruption. Following an analysis of all OpenSSL
7038 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7039 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7040 the first called function after an EVP_EncryptInit(), and therefore that
7041 specific call must be safe. The second form is where the length passed to
7042 EVP_EncryptUpdate() can be seen from the code to be some small value and
7043 therefore there is no possibility of an overflow. Since all instances are
7044 one of these two forms, it is believed that there can be no overflows in
7045 internal code due to this problem. It should be noted that
7046 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7047 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7048 of these calls have also been analysed too and it is believed there are no
7049 instances in internal usage where an overflow could occur.
7051 This issue was reported by Guido Vranken.
7056 * Prevent ASN.1 BIO excessive memory allocation
7058 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7059 a short invalid encoding can casuse allocation of large amounts of memory
7060 potentially consuming excessive resources or exhausting memory.
7062 Any application parsing untrusted data through d2i BIO functions is
7063 affected. The memory based functions such as d2i_X509() are *not* affected.
7064 Since the memory based functions are used by the TLS library, TLS
7065 applications are not affected.
7067 This issue was reported by Brian Carpenter.
7074 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7075 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7076 in arbitrary stack data being returned in the buffer.
7078 This issue was reported by Guido Vranken.
7083 * Modify behavior of ALPN to invoke callback after SNI/servername
7084 callback, such that updates to the SSL_CTX affect ALPN.
7088 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7093 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7094 methods are enabled and ssl2 is disabled the methods return NULL.
7098 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7100 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7101 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7102 provide any "EXPORT" or "LOW" strength ciphers.
7106 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7107 is by default disabled at build-time. Builds that are not configured with
7108 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7109 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7110 will need to explicitly call either of:
7112 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7114 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7116 as appropriate. Even if either of those is used, or the application
7117 explicitly uses the version-specific SSLv2_method() or its client and
7118 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7119 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7120 ciphers, and SSLv2 56-bit DES are no longer available.
7125 * Fix a double-free in DSA code
7127 A double free bug was discovered when OpenSSL parses malformed DSA private
7128 keys and could lead to a DoS attack or memory corruption for applications
7129 that receive DSA private keys from untrusted sources. This scenario is
7132 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7138 * Disable SRP fake user seed to address a server memory leak.
7140 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7142 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7143 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7144 was changed to ignore the "fake user" SRP seed, even if the seed
7147 Users should use SRP_VBASE_get1_by_user instead. Note that in
7148 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7149 also that even though configuring the SRP seed attempts to hide
7150 invalid usernames by continuing the handshake with fake
7151 credentials, this behaviour is not constant time and no strong
7152 guarantees are made that the handshake is indistinguishable from
7153 that of a valid user.
7158 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7160 In the BN_hex2bn function the number of hex digits is calculated using an
7161 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7162 large values of `i` this can result in `bn_expand` not allocating any
7163 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7164 field as NULL leading to a subsequent NULL ptr deref. For very large values
7165 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7166 In this case memory is allocated to the internal BIGNUM data field, but it
7167 is insufficiently sized leading to heap corruption. A similar issue exists
7168 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7169 is ever called by user applications with very large untrusted hex/dec data.
7170 This is anticipated to be a rare occurrence.
7172 All OpenSSL internal usage of these functions use data that is not expected
7173 to be untrusted, e.g. config file data or application command line
7174 arguments. If user developed applications generate config file data based
7175 on untrusted data then it is possible that this could also lead to security
7176 consequences. This is also anticipated to be rare.
7178 This issue was reported to OpenSSL by Guido Vranken.
7183 * Fix memory issues in `BIO_*printf` functions
7185 The internal `fmtstr` function used in processing a "%s" format string in
7186 the `BIO_*printf` functions could overflow while calculating the length of a
7187 string and cause an OOB read when printing very long strings.
7189 Additionally the internal `doapr_outch` function can attempt to write to an
7190 OOB memory location (at an offset from the NULL pointer) in the event of a
7191 memory allocation failure. In 1.0.2 and below this could be caused where
7192 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7193 could be in processing a very long "%s" format string. Memory leaks can
7196 The first issue may mask the second issue dependent on compiler behaviour.
7197 These problems could enable attacks where large amounts of untrusted data
7198 is passed to the `BIO_*printf` functions. If applications use these functions
7199 in this way then they could be vulnerable. OpenSSL itself uses these
7200 functions when printing out human-readable dumps of ASN.1 data. Therefore
7201 applications that print this data could be vulnerable if the data is from
7202 untrusted sources. OpenSSL command line applications could also be
7203 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7204 as command line arguments.
7206 Libssl is not considered directly vulnerable. Additionally certificates etc
7207 received via remote connections via libssl are also unlikely to be able to
7208 trigger these issues because of message size limits enforced within libssl.
7210 This issue was reported to OpenSSL Guido Vranken.
7215 * Side channel attack on modular exponentiation
7217 A side-channel attack was found which makes use of cache-bank conflicts on
7218 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7219 of RSA keys. The ability to exploit this issue is limited as it relies on
7220 an attacker who has control of code in a thread running on the same
7221 hyper-threaded core as the victim thread which is performing decryptions.
7223 This issue was reported to OpenSSL by Yuval Yarom, The University of
7224 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7225 Nadia Heninger, University of Pennsylvania with more information at
7226 <http://cachebleed.info>.
7231 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7232 if no keysize is specified with default_bits. This fixes an
7233 omission in an earlier change that changed all RSA/DSA key generation
7234 commands to use 2048 bits by default.
7238 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7240 * Protection for DH small subgroup attacks
7242 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7243 switched on by default and cannot be disabled. This could have some
7248 * SSLv2 doesn't block disabled ciphers
7250 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7251 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7252 been disabled, provided that the SSLv2 protocol was not also disabled via
7255 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7256 and Sebastian Schinzel.
7261 * Reject DH handshakes with parameters shorter than 1024 bits.
7265 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7267 * Certificate verify crash with missing PSS parameter
7269 The signature verification routines will crash with a NULL pointer
7270 dereference if presented with an ASN.1 signature using the RSA PSS
7271 algorithm and absent mask generation function parameter. Since these
7272 routines are used to verify certificate signature algorithms this can be
7273 used to crash any certificate verification operation and exploited in a
7274 DoS attack. Any application which performs certificate verification is
7275 vulnerable including OpenSSL clients and servers which enable client
7278 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7283 * X509_ATTRIBUTE memory leak
7285 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7286 memory. This structure is used by the PKCS#7 and CMS routines so any
7287 application which reads PKCS#7 or CMS data from untrusted sources is
7288 affected. SSL/TLS is not affected.
7290 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7296 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7297 This changes the decoding behaviour for some invalid messages,
7298 though the change is mostly in the more lenient direction, and
7299 legacy behaviour is preserved as much as possible.
7303 * In DSA_generate_parameters_ex, if the provided seed is too short,
7304 use a random seed, as already documented.
7306 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7308 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7310 * Alternate chains certificate forgery
7312 During certificate verfification, OpenSSL will attempt to find an
7313 alternative certificate chain if the first attempt to build such a chain
7314 fails. An error in the implementation of this logic can mean that an
7315 attacker could cause certain checks on untrusted certificates to be
7316 bypassed, such as the CA flag, enabling them to use a valid leaf
7317 certificate to act as a CA and "issue" an invalid certificate.
7319 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7325 * Race condition handling PSK identify hint
7327 If PSK identity hints are received by a multi-threaded client then
7328 the values are wrongly updated in the parent SSL_CTX structure. This can
7329 result in a race condition potentially leading to a double free of the
7335 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7337 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7338 incompatibility in the handling of HMAC. The previous ABI has now been
7341 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7343 * Malformed ECParameters causes infinite loop
7345 When processing an ECParameters structure OpenSSL enters an infinite loop
7346 if the curve specified is over a specially malformed binary polynomial
7349 This can be used to perform denial of service against any
7350 system which processes public keys, certificate requests or
7351 certificates. This includes TLS clients and TLS servers with
7352 client authentication enabled.
7354 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7359 * Exploitable out-of-bounds read in X509_cmp_time
7361 X509_cmp_time does not properly check the length of the ASN1_TIME
7362 string and can read a few bytes out of bounds. In addition,
7363 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7366 An attacker can use this to craft malformed certificates and CRLs of
7367 various sizes and potentially cause a segmentation fault, resulting in
7368 a DoS on applications that verify certificates or CRLs. TLS clients
7369 that verify CRLs are affected. TLS clients and servers with client
7370 authentication enabled may be affected if they use custom verification
7373 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7374 independently by Hanno Böck.
7379 * PKCS7 crash with missing EnvelopedContent
7381 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7382 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7383 with missing content and trigger a NULL pointer dereference on parsing.
7385 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7386 structures from untrusted sources are affected. OpenSSL clients and
7387 servers are not affected.
7389 This issue was reported to OpenSSL by Michal Zalewski (Google).
7394 * CMS verify infinite loop with unknown hash function
7396 When verifying a signedData message the CMS code can enter an infinite loop
7397 if presented with an unknown hash function OID. This can be used to perform
7398 denial of service against any system which verifies signedData messages using
7400 This issue was reported to OpenSSL by Johannes Bauer.
7405 * Race condition handling NewSessionTicket
7407 If a NewSessionTicket is received by a multi-threaded client when attempting to
7408 reuse a previous ticket then a race condition can occur potentially leading to
7409 a double free of the ticket data.
7414 * Reject DH handshakes with parameters shorter than 768 bits.
7416 *Kurt Roeckx and Emilia Kasper*
7418 * dhparam: generate 2048-bit parameters by default.
7420 *Kurt Roeckx and Emilia Kasper*
7422 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7424 * Segmentation fault in ASN1_TYPE_cmp fix
7426 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7427 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7428 certificate signature algorithm consistency this can be used to crash any
7429 certificate verification operation and exploited in a DoS attack. Any
7430 application which performs certificate verification is vulnerable including
7431 OpenSSL clients and servers which enable client authentication.
7436 * ASN.1 structure reuse memory corruption fix
7438 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7439 memory corruption via an invalid write. Such reuse is and has been
7440 strongly discouraged and is believed to be rare.
7442 Applications that parse structures containing CHOICE or ANY DEFINED BY
7443 components may be affected. Certificate parsing (d2i_X509 and related
7444 functions) are however not affected. OpenSSL clients and servers are
7450 * PKCS7 NULL pointer dereferences fix
7452 The PKCS#7 parsing code does not handle missing outer ContentInfo
7453 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7454 missing content and trigger a NULL pointer dereference on parsing.
7456 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7457 otherwise parse PKCS#7 structures from untrusted sources are
7458 affected. OpenSSL clients and servers are not affected.
7460 This issue was reported to OpenSSL by Michal Zalewski (Google).
7465 * DoS via reachable assert in SSLv2 servers fix
7467 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7468 servers that both support SSLv2 and enable export cipher suites by sending
7469 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7471 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7472 (OpenSSL development team).
7477 * Use After Free following d2i_ECPrivatekey error fix
7479 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7480 could cause a use after free condition. This, in turn, could cause a double
7481 free in several private key parsing functions (such as d2i_PrivateKey
7482 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7483 for applications that receive EC private keys from untrusted
7484 sources. This scenario is considered rare.
7486 This issue was discovered by the BoringSSL project and fixed in their
7492 * X509_to_X509_REQ NULL pointer deref fix
7494 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7495 the certificate key is invalid. This function is rarely used in practice.
7497 This issue was discovered by Brian Carpenter.
7502 * Removed the export ciphers from the DEFAULT ciphers
7506 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7508 * Build fixes for the Windows and OpenVMS platforms
7510 *Matt Caswell and Richard Levitte*
7512 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7514 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7515 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7516 dereference. This could lead to a Denial Of Service attack. Thanks to
7517 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7522 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7523 dtls1_buffer_record function under certain conditions. In particular this
7524 could occur if an attacker sent repeated DTLS records with the same
7525 sequence number but for the next epoch. The memory leak could be exploited
7526 by an attacker in a Denial of Service attack through memory exhaustion.
7527 Thanks to Chris Mueller for reporting this issue.
7532 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7533 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7534 method would be set to NULL which could later result in a NULL pointer
7535 dereference. Thanks to Frank Schmirler for reporting this issue.
7540 * Abort handshake if server key exchange message is omitted for ephemeral
7543 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7544 reporting this issue.
7549 * Remove non-export ephemeral RSA code on client and server. This code
7550 violated the TLS standard by allowing the use of temporary RSA keys in
7551 non-export ciphersuites and could be used by a server to effectively
7552 downgrade the RSA key length used to a value smaller than the server
7553 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7554 INRIA or reporting this issue.
7559 * Fixed issue where DH client certificates are accepted without verification.
7560 An OpenSSL server will accept a DH certificate for client authentication
7561 without the certificate verify message. This effectively allows a client to
7562 authenticate without the use of a private key. This only affects servers
7563 which trust a client certificate authority which issues certificates
7564 containing DH keys: these are extremely rare and hardly ever encountered.
7565 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7571 * Ensure that the session ID context of an SSL is updated when its
7572 SSL_CTX is updated via SSL_set_SSL_CTX.
7574 The session ID context is typically set from the parent SSL_CTX,
7575 and can vary with the CTX.
7579 * Fix various certificate fingerprint issues.
7581 By using non-DER or invalid encodings outside the signed portion of a
7582 certificate the fingerprint can be changed without breaking the signature.
7583 Although no details of the signed portion of the certificate can be changed
7584 this can cause problems with some applications: e.g. those using the
7585 certificate fingerprint for blacklists.
7587 1. Reject signatures with non zero unused bits.
7589 If the BIT STRING containing the signature has non zero unused bits reject
7590 the signature. All current signature algorithms require zero unused bits.
7592 2. Check certificate algorithm consistency.
7594 Check the AlgorithmIdentifier inside TBS matches the one in the
7595 certificate signature. NB: this will result in signature failure
7596 errors for some broken certificates.
7598 Thanks to Konrad Kraszewski from Google for reporting this issue.
7600 3. Check DSA/ECDSA signatures use DER.
7602 Re-encode DSA/ECDSA signatures and compare with the original received
7603 signature. Return an error if there is a mismatch.
7605 This will reject various cases including garbage after signature
7606 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7607 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7608 (negative or with leading zeroes).
7610 Further analysis was conducted and fixes were developed by Stephen Henson
7611 of the OpenSSL core team.
7617 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7618 results on some platforms, including x86_64. This bug occurs at random
7619 with a very low probability, and is not known to be exploitable in any
7620 way, though its exact impact is difficult to determine. Thanks to Pieter
7621 Wuille (Blockstream) who reported this issue and also suggested an initial
7622 fix. Further analysis was conducted by the OpenSSL development team and
7623 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7624 the OpenSSL core team.
7629 * Do not resume sessions on the server if the negotiated protocol
7630 version does not match the session's version. Resuming with a different
7631 version, while not strictly forbidden by the RFC, is of questionable
7632 sanity and breaks all known clients.
7634 *David Benjamin, Emilia Käsper*
7636 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7637 early CCS messages during renegotiation. (Note that because
7638 renegotiation is encrypted, this early CCS was not exploitable.)
7642 * Tighten client-side session ticket handling during renegotiation:
7643 ensure that the client only accepts a session ticket if the server sends
7644 the extension anew in the ServerHello. Previously, a TLS client would
7645 reuse the old extension state and thus accept a session ticket if one was
7646 announced in the initial ServerHello.
7648 Similarly, ensure that the client requires a session ticket if one
7649 was advertised in the ServerHello. Previously, a TLS client would
7650 ignore a missing NewSessionTicket message.
7654 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7658 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7659 sends a carefully crafted handshake message, to cause OpenSSL to fail
7660 to free up to 64k of memory causing a memory leak. This could be
7661 exploited in a Denial Of Service attack. This issue affects OpenSSL
7662 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7663 whether SRTP is used or configured. Implementations of OpenSSL that
7664 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7666 The fix was developed by the OpenSSL team.
7671 * Session Ticket Memory Leak.
7673 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7674 integrity of that ticket is first verified. In the event of a session
7675 ticket integrity check failing, OpenSSL will fail to free memory
7676 causing a memory leak. By sending a large number of invalid session
7677 tickets an attacker could exploit this issue in a Denial Of Service
7683 * Build option no-ssl3 is incomplete.
7685 When OpenSSL is configured with "no-ssl3" as a build option, servers
7686 could accept and complete an SSL 3.0 handshake, and clients could be
7687 configured to send them.
7690 *Akamai and the OpenSSL team*
7692 * Add support for TLS_FALLBACK_SCSV.
7693 Client applications doing fallback retries should call
7694 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7697 *Adam Langley, Bodo Moeller*
7699 * Add additional DigestInfo checks.
7701 Re-encode DigestInto in DER and check against the original when
7702 verifying RSA signature: this will reject any improperly encoded
7703 DigestInfo structures.
7705 Note: this is a precautionary measure and no attacks are currently known.
7709 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7711 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7712 SRP code can be overrun an internal buffer. Add sanity check that
7713 g, A, B < N to SRP code.
7715 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7716 Group for discovering this issue.
7721 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7722 TLS 1.0 instead of higher protocol versions when the ClientHello message
7723 is badly fragmented. This allows a man-in-the-middle attacker to force a
7724 downgrade to TLS 1.0 even if both the server and the client support a
7725 higher protocol version, by modifying the client's TLS records.
7727 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7728 researching this issue.
7733 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7734 to a denial of service attack. A malicious server can crash the client
7735 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7736 ciphersuite and sending carefully crafted handshake messages.
7738 Thanks to Felix Gröbert (Google) for discovering and researching this
7744 * By sending carefully crafted DTLS packets an attacker could cause openssl
7745 to leak memory. This can be exploited through a Denial of Service attack.
7746 Thanks to Adam Langley for discovering and researching this issue.
7751 * An attacker can force openssl to consume large amounts of memory whilst
7752 processing DTLS handshake messages. This can be exploited through a
7753 Denial of Service attack.
7754 Thanks to Adam Langley for discovering and researching this issue.
7759 * An attacker can force an error condition which causes openssl to crash
7760 whilst processing DTLS packets due to memory being freed twice. This
7761 can be exploited through a Denial of Service attack.
7762 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7768 * If a multithreaded client connects to a malicious server using a resumed
7769 session and the server sends an ec point format extension it could write
7770 up to 255 bytes to freed memory.
7772 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7778 * A malicious server can crash an OpenSSL client with a null pointer
7779 dereference (read) by specifying an SRP ciphersuite even though it was not
7780 properly negotiated with the client. This can be exploited through a
7781 Denial of Service attack.
7783 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7784 discovering and researching this issue.
7789 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7790 X509_name_oneline, X509_name_print_ex et al. to leak some information
7791 from the stack. Applications may be affected if they echo pretty printing
7792 output to the attacker.
7794 Thanks to Ivan Fratric (Google) for discovering this issue.
7797 *Emilia Käsper, and Steve Henson*
7799 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7800 for corner cases. (Certain input points at infinity could lead to
7801 bogus results, with non-infinity inputs mapped to infinity too.)
7805 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7807 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7808 handshake can force the use of weak keying material in OpenSSL
7809 SSL/TLS clients and servers.
7811 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7812 researching this issue. ([CVE-2014-0224])
7814 *KIKUCHI Masashi, Steve Henson*
7816 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7817 OpenSSL DTLS client the code can be made to recurse eventually crashing
7820 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7823 *Imre Rad, Steve Henson*
7825 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7826 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7827 client or server. This is potentially exploitable to run arbitrary
7828 code on a vulnerable client or server.
7830 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7832 *Jüri Aedla, Steve Henson*
7834 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7835 are subject to a denial of service attack.
7837 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7838 this issue. ([CVE-2014-3470])
7840 *Felix Gröbert, Ivan Fratric, Steve Henson*
7842 * Harmonize version and its documentation. -f flag is used to display
7845 *mancha <mancha1@zoho.com>*
7847 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7848 in i2d_ECPrivateKey.
7850 *mancha <mancha1@zoho.com>*
7852 * Fix some double frees. These are not thought to be exploitable.
7854 *mancha <mancha1@zoho.com>*
7856 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7858 * A missing bounds check in the handling of the TLS heartbeat extension
7859 can be used to reveal up to 64k of memory to a connected client or
7862 Thanks for Neel Mehta of Google Security for discovering this bug and to
7863 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7864 preparing the fix ([CVE-2014-0160])
7866 *Adam Langley, Bodo Moeller*
7868 * Fix for the attack described in the paper "Recovering OpenSSL
7869 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7870 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7871 <http://eprint.iacr.org/2014/140>
7873 Thanks to Yuval Yarom and Naomi Benger for discovering this
7874 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7876 *Yuval Yarom and Naomi Benger*
7878 * TLS pad extension: draft-agl-tls-padding-03
7880 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7881 TLS client Hello record length value would otherwise be > 255 and
7882 less that 512 pad with a dummy extension containing zeroes so it
7883 is at least 512 bytes long.
7885 *Adam Langley, Steve Henson*
7887 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7889 * Fix for TLS record tampering bug. A carefully crafted invalid
7890 handshake could crash OpenSSL with a NULL pointer exception.
7891 Thanks to Anton Johansson for reporting this issues.
7894 * Keep original DTLS digest and encryption contexts in retransmission
7895 structures so we can use the previous session parameters if they need
7896 to be resent. ([CVE-2013-6450])
7900 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7901 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7902 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7903 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7904 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7905 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7907 *Rob Stradling, Adam Langley*
7909 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7911 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7912 supporting platforms or when small records were transferred.
7914 *Andy Polyakov, Steve Henson*
7916 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7918 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7920 This addresses the flaw in CBC record processing discovered by
7921 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7922 at: <http://www.isg.rhul.ac.uk/tls/>
7924 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7925 Security Group at Royal Holloway, University of London
7926 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7927 Emilia Käsper for the initial patch.
7930 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7932 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7933 ciphersuites which can be exploited in a denial of service attack.
7934 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7935 and detecting this bug and to Wolfgang Ettlinger
7936 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7941 * Return an error when checking OCSP signatures when key is NULL.
7942 This fixes a DoS attack. ([CVE-2013-0166])
7946 * Make openssl verify return errors.
7948 *Chris Palmer <palmer@google.com> and Ben Laurie*
7950 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7951 the right response is stapled. Also change SSL_get_certificate()
7952 so it returns the certificate actually sent.
7953 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7955 *Rob Stradling <rob.stradling@comodo.com>*
7957 * Fix possible deadlock when decoding public keys.
7961 * Don't use TLS 1.0 record version number in initial client hello
7966 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7968 * Sanity check record length before skipping explicit IV in TLS
7969 1.2, 1.1 and DTLS to fix DoS attack.
7971 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7972 fuzzing as a service testing platform.
7977 * Initialise tkeylen properly when encrypting CMS messages.
7978 Thanks to Solar Designer of Openwall for reporting this issue.
7982 * In FIPS mode don't try to use composite ciphers as they are not
7987 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7989 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7990 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7991 mean any application compiled against OpenSSL 1.0.0 headers setting
7992 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7993 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7994 0x10000000L Any application which was previously compiled against
7995 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7996 will need to be recompiled as a result. Letting be results in
7997 inability to disable specifically TLS 1.1 and in client context,
7998 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8002 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8003 disable just protocol X, but all protocols above X *if* there are
8004 protocols *below* X still enabled. In more practical terms it means
8005 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8006 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8007 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8012 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8014 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8015 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8016 in CRYPTO_realloc_clean.
8018 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8019 issue and to Adam Langley <agl@chromium.org> for fixing it.
8022 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8024 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8028 * Workarounds for some broken servers that "hang" if a client hello
8029 record length exceeds 255 bytes.
8031 1. Do not use record version number > TLS 1.0 in initial client
8032 hello: some (but not all) hanging servers will now work.
8033 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8034 the number of ciphers sent in the client hello. This should be
8035 set to an even number, such as 50, for example by passing:
8036 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8037 Most broken servers should now work.
8038 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8039 TLS 1.2 client support entirely.
8043 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8047 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8049 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8050 STRING form instead of a DigestInfo.
8054 * The format used for MDC2 RSA signatures is inconsistent between EVP
8055 and the RSA_sign/RSA_verify functions. This was made more apparent when
8056 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8057 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8058 the correct format in RSA_verify so both forms transparently work.
8062 * Some servers which support TLS 1.0 can choke if we initially indicate
8063 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8064 encrypted premaster secret. As a workaround use the maximum permitted
8065 client version in client hello, this should keep such servers happy
8066 and still work with previous versions of OpenSSL.
8070 * Add support for TLS/DTLS heartbeats.
8072 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8074 * Add support for SCTP.
8076 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8078 * Improved PRNG seeding for VOS.
8080 *Paul Green <Paul.Green@stratus.com>*
8082 * Extensive assembler packs updates, most notably:
8084 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8085 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8086 - x86_64: bit-sliced AES implementation;
8087 - ARM: NEON support, contemporary platforms optimizations;
8088 - s390x: z196 support;
8089 - `*`: GHASH and GF(2^m) multiplication implementations;
8093 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8094 (removal of unnecessary code)
8096 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8098 * Add TLS key material exporter from RFC 5705.
8102 * Add DTLS-SRTP negotiation from RFC 5764.
8106 * Add Next Protocol Negotiation,
8107 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8108 disabled with a no-npn flag to config or Configure. Code donated
8111 *Adam Langley <agl@google.com> and Ben Laurie*
8113 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8114 NIST-P256, NIST-P521, with constant-time single point multiplication on
8115 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8116 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8117 Code made available under Apache License version 2.0.
8119 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8120 line to include this in your build of OpenSSL, and run "make depend" (or
8121 "make update"). This enables the following EC_METHODs:
8123 EC_GFp_nistp224_method()
8124 EC_GFp_nistp256_method()
8125 EC_GFp_nistp521_method()
8127 EC_GROUP_new_by_curve_name() will automatically use these (while
8128 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8131 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8133 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8134 all platforms. Move ssize_t definition from e_os.h to the public
8135 header file e_os2.h as it now appears in public header file cms.h
8139 * New -sigopt option to the ca, req and x509 utilities. Additional
8140 signature parameters can be passed using this option and in
8145 * Add RSA PSS signing function. This will generate and set the
8146 appropriate AlgorithmIdentifiers for PSS based on those in the
8147 corresponding EVP_MD_CTX structure. No application support yet.
8151 * Support for companion algorithm specific ASN1 signing routines.
8152 New function ASN1_item_sign_ctx() signs a pre-initialised
8153 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8154 the appropriate parameters.
8158 * Add new algorithm specific ASN1 verification initialisation function
8159 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8160 handling will be the same no matter what EVP_PKEY_METHOD is used.
8161 Add a PSS handler to support verification of PSS signatures: checked
8162 against a number of sample certificates.
8166 * Add signature printing for PSS. Add PSS OIDs.
8168 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8170 * Add algorithm specific signature printing. An individual ASN1 method
8171 can now print out signatures instead of the standard hex dump.
8173 More complex signatures (e.g. PSS) can print out more meaningful
8174 information. Include DSA version that prints out the signature
8179 * Password based recipient info support for CMS library: implementing
8184 * Split password based encryption into PBES2 and PBKDF2 functions. This
8185 neatly separates the code into cipher and PBE sections and is required
8186 for some algorithms that split PBES2 into separate pieces (such as
8187 password based CMS).
8191 * Session-handling fixes:
8192 - Fix handling of connections that are resuming with a session ID,
8193 but also support Session Tickets.
8194 - Fix a bug that suppressed issuing of a new ticket if the client
8195 presented a ticket with an expired session.
8196 - Try to set the ticket lifetime hint to something reasonable.
8197 - Make tickets shorter by excluding irrelevant information.
8198 - On the client side, don't ignore renewed tickets.
8200 *Adam Langley, Bodo Moeller (Google)*
8202 * Fix PSK session representation.
8206 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8208 This work was sponsored by Intel.
8212 * Add GCM support to TLS library. Some custom code is needed to split
8213 the IV between the fixed (from PRF) and explicit (from TLS record)
8214 portions. This adds all GCM ciphersuites supported by RFC5288 and
8215 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8216 add a special AESGCM string for GCM only.
8220 * Expand range of ctrls for AES GCM. Permit setting invocation
8221 field on decrypt and retrieval of invocation field only on encrypt.
8225 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8226 As required by RFC5289 these ciphersuites cannot be used if for
8227 versions of TLS earlier than 1.2.
8231 * For FIPS capable OpenSSL interpret a NULL default public key method
8232 as unset and return the appropriate default but do *not* set the default.
8233 This means we can return the appropriate method in applications that
8234 switch between FIPS and non-FIPS modes.
8238 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8239 ENGINE is used then we cannot handle that in the FIPS module so we
8240 keep original code iff non-FIPS operations are allowed.
8244 * Add -attime option to openssl utilities.
8246 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8248 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8252 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8253 FIPS EC methods unconditionally for now.
8257 * New build option no-ec2m to disable characteristic 2 code.
8261 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8262 all cases can be covered as some introduce binary incompatibilities.
8266 * Redirect RSA operations to FIPS module including keygen,
8267 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8271 * Add similar low-level API blocking to ciphers.
8275 * low-level digest APIs are not approved in FIPS mode: any attempt
8276 to use these will cause a fatal error. Applications that *really* want
8277 to use them can use the `private_*` version instead.
8281 * Redirect cipher operations to FIPS module for FIPS builds.
8285 * Redirect digest operations to FIPS module for FIPS builds.
8289 * Update build system to add "fips" flag which will link in fipscanister.o
8290 for static and shared library builds embedding a signature if needed.
8294 * Output TLS supported curves in preference order instead of numerical
8295 order. This is currently hardcoded for the highest order curves first.
8296 This should be configurable so applications can judge speed vs strength.
8300 * Add TLS v1.2 server support for client authentication.
8304 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8309 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8310 FIPS modules versions.
8314 * Add TLS v1.2 client side support for client authentication. Keep cache
8315 of handshake records longer as we don't know the hash algorithm to use
8316 until after the certificate request message is received.
8320 * Initial TLS v1.2 client support. Add a default signature algorithms
8321 extension including all the algorithms we support. Parse new signature
8322 format in client key exchange. Relax some ECC signing restrictions for
8323 TLS v1.2 as indicated in RFC5246.
8327 * Add server support for TLS v1.2 signature algorithms extension. Switch
8328 to new signature format when needed using client digest preference.
8329 All server ciphersuites should now work correctly in TLS v1.2. No client
8330 support yet and no support for client certificates.
8334 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8335 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8336 ciphersuites. At present only RSA key exchange ciphersuites work with
8337 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8338 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8339 and version checking.
8343 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8344 with this defined it will not be affected by any changes to ssl internal
8345 structures. Add several utility functions to allow openssl application
8346 to work with OPENSSL_NO_SSL_INTERN defined.
8350 * A long standing patch to add support for SRP from EdelWeb (Peter
8351 Sylvester and Christophe Renou) was integrated.
8352 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8353 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8356 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8360 * Permit abbreviated handshakes when renegotiating using the function
8361 SSL_renegotiate_abbreviated().
8363 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8365 * Add call to ENGINE_register_all_complete() to
8366 ENGINE_load_builtin_engines(), so some implementations get used
8367 automatically instead of needing explicit application support.
8371 * Add support for TLS key exporter as described in RFC5705.
8373 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8375 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8376 a few changes are required:
8378 Add SSL_OP_NO_TLSv1_1 flag.
8379 Add TLSv1_1 methods.
8380 Update version checking logic to handle version 1.1.
8381 Add explicit IV handling (ported from DTLS code).
8382 Add command line options to s_client/s_server.
8389 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8391 * X509_ATTRIBUTE memory leak
8393 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8394 memory. This structure is used by the PKCS#7 and CMS routines so any
8395 application which reads PKCS#7 or CMS data from untrusted sources is
8396 affected. SSL/TLS is not affected.
8398 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8404 * Race condition handling PSK identify hint
8406 If PSK identity hints are received by a multi-threaded client then
8407 the values are wrongly updated in the parent SSL_CTX structure. This can
8408 result in a race condition potentially leading to a double free of the
8414 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8416 * Malformed ECParameters causes infinite loop
8418 When processing an ECParameters structure OpenSSL enters an infinite loop
8419 if the curve specified is over a specially malformed binary polynomial
8422 This can be used to perform denial of service against any
8423 system which processes public keys, certificate requests or
8424 certificates. This includes TLS clients and TLS servers with
8425 client authentication enabled.
8427 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8432 * Exploitable out-of-bounds read in X509_cmp_time
8434 X509_cmp_time does not properly check the length of the ASN1_TIME
8435 string and can read a few bytes out of bounds. In addition,
8436 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8439 An attacker can use this to craft malformed certificates and CRLs of
8440 various sizes and potentially cause a segmentation fault, resulting in
8441 a DoS on applications that verify certificates or CRLs. TLS clients
8442 that verify CRLs are affected. TLS clients and servers with client
8443 authentication enabled may be affected if they use custom verification
8446 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8447 independently by Hanno Böck.
8452 * PKCS7 crash with missing EnvelopedContent
8454 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8455 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8456 with missing content and trigger a NULL pointer dereference on parsing.
8458 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8459 structures from untrusted sources are affected. OpenSSL clients and
8460 servers are not affected.
8462 This issue was reported to OpenSSL by Michal Zalewski (Google).
8467 * CMS verify infinite loop with unknown hash function
8469 When verifying a signedData message the CMS code can enter an infinite loop
8470 if presented with an unknown hash function OID. This can be used to perform
8471 denial of service against any system which verifies signedData messages using
8473 This issue was reported to OpenSSL by Johannes Bauer.
8478 * Race condition handling NewSessionTicket
8480 If a NewSessionTicket is received by a multi-threaded client when attempting to
8481 reuse a previous ticket then a race condition can occur potentially leading to
8482 a double free of the ticket data.
8487 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8489 * Segmentation fault in ASN1_TYPE_cmp fix
8491 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8492 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8493 certificate signature algorithm consistency this can be used to crash any
8494 certificate verification operation and exploited in a DoS attack. Any
8495 application which performs certificate verification is vulnerable including
8496 OpenSSL clients and servers which enable client authentication.
8501 * ASN.1 structure reuse memory corruption fix
8503 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8504 memory corruption via an invalid write. Such reuse is and has been
8505 strongly discouraged and is believed to be rare.
8507 Applications that parse structures containing CHOICE or ANY DEFINED BY
8508 components may be affected. Certificate parsing (d2i_X509 and related
8509 functions) are however not affected. OpenSSL clients and servers are
8515 * PKCS7 NULL pointer dereferences fix
8517 The PKCS#7 parsing code does not handle missing outer ContentInfo
8518 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8519 missing content and trigger a NULL pointer dereference on parsing.
8521 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8522 otherwise parse PKCS#7 structures from untrusted sources are
8523 affected. OpenSSL clients and servers are not affected.
8525 This issue was reported to OpenSSL by Michal Zalewski (Google).
8530 * DoS via reachable assert in SSLv2 servers fix
8532 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8533 servers that both support SSLv2 and enable export cipher suites by sending
8534 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8536 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8537 (OpenSSL development team).
8542 * Use After Free following d2i_ECPrivatekey error fix
8544 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8545 could cause a use after free condition. This, in turn, could cause a double
8546 free in several private key parsing functions (such as d2i_PrivateKey
8547 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8548 for applications that receive EC private keys from untrusted
8549 sources. This scenario is considered rare.
8551 This issue was discovered by the BoringSSL project and fixed in their
8557 * X509_to_X509_REQ NULL pointer deref fix
8559 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8560 the certificate key is invalid. This function is rarely used in practice.
8562 This issue was discovered by Brian Carpenter.
8567 * Removed the export ciphers from the DEFAULT ciphers
8571 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8573 * Build fixes for the Windows and OpenVMS platforms
8575 *Matt Caswell and Richard Levitte*
8577 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8579 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8580 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8581 dereference. This could lead to a Denial Of Service attack. Thanks to
8582 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8587 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8588 dtls1_buffer_record function under certain conditions. In particular this
8589 could occur if an attacker sent repeated DTLS records with the same
8590 sequence number but for the next epoch. The memory leak could be exploited
8591 by an attacker in a Denial of Service attack through memory exhaustion.
8592 Thanks to Chris Mueller for reporting this issue.
8597 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8598 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8599 method would be set to NULL which could later result in a NULL pointer
8600 dereference. Thanks to Frank Schmirler for reporting this issue.
8605 * Abort handshake if server key exchange message is omitted for ephemeral
8608 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8609 reporting this issue.
8614 * Remove non-export ephemeral RSA code on client and server. This code
8615 violated the TLS standard by allowing the use of temporary RSA keys in
8616 non-export ciphersuites and could be used by a server to effectively
8617 downgrade the RSA key length used to a value smaller than the server
8618 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8619 INRIA or reporting this issue.
8624 * Fixed issue where DH client certificates are accepted without verification.
8625 An OpenSSL server will accept a DH certificate for client authentication
8626 without the certificate verify message. This effectively allows a client to
8627 authenticate without the use of a private key. This only affects servers
8628 which trust a client certificate authority which issues certificates
8629 containing DH keys: these are extremely rare and hardly ever encountered.
8630 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8636 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8637 results on some platforms, including x86_64. This bug occurs at random
8638 with a very low probability, and is not known to be exploitable in any
8639 way, though its exact impact is difficult to determine. Thanks to Pieter
8640 Wuille (Blockstream) who reported this issue and also suggested an initial
8641 fix. Further analysis was conducted by the OpenSSL development team and
8642 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8643 the OpenSSL core team.
8648 * Fix various certificate fingerprint issues.
8650 By using non-DER or invalid encodings outside the signed portion of a
8651 certificate the fingerprint can be changed without breaking the signature.
8652 Although no details of the signed portion of the certificate can be changed
8653 this can cause problems with some applications: e.g. those using the
8654 certificate fingerprint for blacklists.
8656 1. Reject signatures with non zero unused bits.
8658 If the BIT STRING containing the signature has non zero unused bits reject
8659 the signature. All current signature algorithms require zero unused bits.
8661 2. Check certificate algorithm consistency.
8663 Check the AlgorithmIdentifier inside TBS matches the one in the
8664 certificate signature. NB: this will result in signature failure
8665 errors for some broken certificates.
8667 Thanks to Konrad Kraszewski from Google for reporting this issue.
8669 3. Check DSA/ECDSA signatures use DER.
8671 Reencode DSA/ECDSA signatures and compare with the original received
8672 signature. Return an error if there is a mismatch.
8674 This will reject various cases including garbage after signature
8675 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8676 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8677 (negative or with leading zeroes).
8679 Further analysis was conducted and fixes were developed by Stephen Henson
8680 of the OpenSSL core team.
8686 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8688 * Session Ticket Memory Leak.
8690 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8691 integrity of that ticket is first verified. In the event of a session
8692 ticket integrity check failing, OpenSSL will fail to free memory
8693 causing a memory leak. By sending a large number of invalid session
8694 tickets an attacker could exploit this issue in a Denial Of Service
8700 * Build option no-ssl3 is incomplete.
8702 When OpenSSL is configured with "no-ssl3" as a build option, servers
8703 could accept and complete an SSL 3.0 handshake, and clients could be
8704 configured to send them.
8707 *Akamai and the OpenSSL team*
8709 * Add support for TLS_FALLBACK_SCSV.
8710 Client applications doing fallback retries should call
8711 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8714 *Adam Langley, Bodo Moeller*
8716 * Add additional DigestInfo checks.
8718 Reencode DigestInto in DER and check against the original when
8719 verifying RSA signature: this will reject any improperly encoded
8720 DigestInfo structures.
8722 Note: this is a precautionary measure and no attacks are currently known.
8726 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8728 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8729 to a denial of service attack. A malicious server can crash the client
8730 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8731 ciphersuite and sending carefully crafted handshake messages.
8733 Thanks to Felix Gröbert (Google) for discovering and researching this
8739 * By sending carefully crafted DTLS packets an attacker could cause openssl
8740 to leak memory. This can be exploited through a Denial of Service attack.
8741 Thanks to Adam Langley for discovering and researching this issue.
8746 * An attacker can force openssl to consume large amounts of memory whilst
8747 processing DTLS handshake messages. This can be exploited through a
8748 Denial of Service attack.
8749 Thanks to Adam Langley for discovering and researching this issue.
8754 * An attacker can force an error condition which causes openssl to crash
8755 whilst processing DTLS packets due to memory being freed twice. This
8756 can be exploited through a Denial of Service attack.
8757 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8763 * If a multithreaded client connects to a malicious server using a resumed
8764 session and the server sends an ec point format extension it could write
8765 up to 255 bytes to freed memory.
8767 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8773 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8774 X509_name_oneline, X509_name_print_ex et al. to leak some information
8775 from the stack. Applications may be affected if they echo pretty printing
8776 output to the attacker.
8778 Thanks to Ivan Fratric (Google) for discovering this issue.
8781 *Emilia Käsper, and Steve Henson*
8783 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8784 for corner cases. (Certain input points at infinity could lead to
8785 bogus results, with non-infinity inputs mapped to infinity too.)
8789 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8791 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8792 handshake can force the use of weak keying material in OpenSSL
8793 SSL/TLS clients and servers.
8795 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8796 researching this issue. ([CVE-2014-0224])
8798 *KIKUCHI Masashi, Steve Henson*
8800 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8801 OpenSSL DTLS client the code can be made to recurse eventually crashing
8804 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8807 *Imre Rad, Steve Henson*
8809 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8810 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8811 client or server. This is potentially exploitable to run arbitrary
8812 code on a vulnerable client or server.
8814 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8816 *Jüri Aedla, Steve Henson*
8818 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8819 are subject to a denial of service attack.
8821 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8822 this issue. ([CVE-2014-3470])
8824 *Felix Gröbert, Ivan Fratric, Steve Henson*
8826 * Harmonize version and its documentation. -f flag is used to display
8829 *mancha <mancha1@zoho.com>*
8831 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8832 in i2d_ECPrivateKey.
8834 *mancha <mancha1@zoho.com>*
8836 * Fix some double frees. These are not thought to be exploitable.
8838 *mancha <mancha1@zoho.com>*
8840 * Fix for the attack described in the paper "Recovering OpenSSL
8841 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8842 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8843 <http://eprint.iacr.org/2014/140>
8845 Thanks to Yuval Yarom and Naomi Benger for discovering this
8846 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8848 *Yuval Yarom and Naomi Benger*
8850 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8852 * Keep original DTLS digest and encryption contexts in retransmission
8853 structures so we can use the previous session parameters if they need
8854 to be resent. ([CVE-2013-6450])
8858 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8859 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8860 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8861 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8862 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8863 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8865 *Rob Stradling, Adam Langley*
8867 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8869 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8871 This addresses the flaw in CBC record processing discovered by
8872 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8873 at: <http://www.isg.rhul.ac.uk/tls/>
8875 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8876 Security Group at Royal Holloway, University of London
8877 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8878 Emilia Käsper for the initial patch.
8881 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8883 * Return an error when checking OCSP signatures when key is NULL.
8884 This fixes a DoS attack. ([CVE-2013-0166])
8888 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8889 the right response is stapled. Also change SSL_get_certificate()
8890 so it returns the certificate actually sent.
8891 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8892 (This is a backport)
8894 *Rob Stradling <rob.stradling@comodo.com>*
8896 * Fix possible deadlock when decoding public keys.
8900 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8902 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8905 * Sanity check record length before skipping explicit IV in DTLS
8908 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8909 fuzzing as a service testing platform.
8914 * Initialise tkeylen properly when encrypting CMS messages.
8915 Thanks to Solar Designer of Openwall for reporting this issue.
8919 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8921 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8922 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8923 in CRYPTO_realloc_clean.
8925 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8926 issue and to Adam Langley <agl@chromium.org> for fixing it.
8929 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8931 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8933 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8934 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8935 content decryption and always return the same error. Note: this attack
8936 needs on average 2^20 messages so it only affects automated senders. The
8937 old behaviour can be re-enabled in the CMS code by setting the
8938 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8939 an MMA defence is not necessary.
8940 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8941 this issue. ([CVE-2012-0884])
8945 * Fix CVE-2011-4619: make sure we really are receiving a
8946 client hello before rejecting multiple SGC restarts. Thanks to
8947 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8951 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8953 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8954 Thanks to Antonio Martin, Enterprise Secure Access Research and
8955 Development, Cisco Systems, Inc. for discovering this bug and
8956 preparing a fix. ([CVE-2012-0050])
8960 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8962 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8963 of the Vaudenay padding oracle attack on CBC mode encryption
8964 which enables an efficient plaintext recovery attack against
8965 the OpenSSL implementation of DTLS. Their attack exploits timing
8966 differences arising during decryption processing. A research
8967 paper describing this attack can be found at:
8968 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8969 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8970 Security Group at Royal Holloway, University of London
8971 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8972 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8973 for preparing the fix. ([CVE-2011-4108])
8975 *Robin Seggelmann, Michael Tuexen*
8977 * Clear bytes used for block padding of SSL 3.0 records.
8980 *Adam Langley (Google)*
8982 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8983 Kadianakis <desnacked@gmail.com> for discovering this issue and
8984 Adam Langley for preparing the fix. ([CVE-2011-4619])
8986 *Adam Langley (Google)*
8988 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8990 *Andrey Kulikov <amdeich@gmail.com>*
8992 * Prevent malformed RFC3779 data triggering an assertion failure.
8993 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8994 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8996 *Rob Austein <sra@hactrn.net>*
8998 * Improved PRNG seeding for VOS.
9000 *Paul Green <Paul.Green@stratus.com>*
9002 * Fix ssl_ciph.c set-up race.
9004 *Adam Langley (Google)*
9006 * Fix spurious failures in ecdsatest.c.
9008 *Emilia Käsper (Google)*
9010 * Fix the BIO_f_buffer() implementation (which was mixing different
9011 interpretations of the `..._len` fields).
9013 *Adam Langley (Google)*
9015 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9016 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9017 threads won't reuse the same blinding coefficients.
9019 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9020 lock to call BN_BLINDING_invert_ex, and avoids one use of
9021 BN_BLINDING_update for each BN_BLINDING structure (previously,
9022 the last update always remained unused).
9024 *Emilia Käsper (Google)*
9026 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9028 *Bob Buckholz (Google)*
9030 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9032 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9033 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9035 *Kaspar Brand <ossl@velox.ch>*
9037 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9038 for multi-threaded use of ECDH. ([CVE-2011-3210])
9040 *Adam Langley (Google)*
9042 * Fix x509_name_ex_d2i memory leak on bad inputs.
9046 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9047 signature public key algorithm by using OID xref utilities instead.
9048 Before this you could only use some ECC ciphersuites with SHA1 only.
9052 * Add protection against ECDSA timing attacks as mentioned in the paper
9053 by Billy Bob Brumley and Nicola Tuveri, see:
9054 <http://eprint.iacr.org/2011/232.pdf>
9056 *Billy Bob Brumley and Nicola Tuveri*
9058 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9060 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9062 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9064 * Fix bug in string printing code: if *any* escaping is enabled we must
9065 escape the escape character (backslash) or the resulting string is
9070 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9072 * Disable code workaround for ancient and obsolete Netscape browsers
9073 and servers: an attacker can use it in a ciphersuite downgrade attack.
9074 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9078 * Fixed J-PAKE implementation error, originally discovered by
9079 Sebastien Martini, further info and confirmation from Stefan
9080 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9084 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9086 * Fix extension code to avoid race conditions which can result in a buffer
9087 overrun vulnerability: resumed sessions must not be modified as they can
9088 be shared by multiple threads. CVE-2010-3864
9092 * Fix WIN32 build system to correctly link an ENGINE directory into
9097 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9099 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9102 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9104 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9106 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9107 context. The operation can be customised via the ctrl mechanism in
9108 case ENGINEs want to include additional functionality.
9112 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9116 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9117 output hashes compatible with older versions of OpenSSL.
9119 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9121 * Fix compression algorithm handling: if resuming a session use the
9122 compression algorithm of the resumed session instead of determining
9123 it from client hello again. Don't allow server to change algorithm.
9127 * Add load_crls() function to commands tidying load_certs() too. Add option
9128 to verify utility to allow additional CRLs to be included.
9132 * Update OCSP request code to permit adding custom headers to the request:
9133 some responders need this.
9137 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9140 *Julia Lawall <julia@diku.dk>*
9142 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9143 needlessly dereferenced structures, used obsolete functions and
9144 didn't handle all updated verify codes correctly.
9148 * Disable MD2 in the default configuration.
9152 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9153 indicate the initial BIO being pushed or popped. This makes it possible
9154 to determine whether the BIO is the one explicitly called or as a result
9155 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9156 it handles reference counts correctly and doesn't zero out the I/O bio
9157 when it is not being explicitly popped. WARNING: applications which
9158 included workarounds for the old buggy behaviour will need to be modified
9159 or they could free up already freed BIOs.
9163 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9164 renaming to all platforms (within the 0.9.8 branch, this was
9165 done conditionally on Netware platforms to avoid a name clash).
9167 *Guenter <lists@gknw.net>*
9169 * Add ECDHE and PSK support to DTLS.
9171 *Michael Tuexen <tuexen@fh-muenster.de>*
9173 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9178 * Add "missing" function EVP_MD_flags() (without this the only way to
9179 retrieve a digest flags is by accessing the structure directly. Update
9180 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9181 or cipher is registered as in the "from" argument. Print out all
9182 registered digests in the dgst usage message instead of manually
9183 attempting to work them out.
9187 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9188 this allows the use of compression and extensions. Change default cipher
9189 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9190 by default unless an application cipher string requests it.
9194 * Alter match criteria in PKCS12_parse(). It used to try to use local
9195 key ids to find matching certificates and keys but some PKCS#12 files
9196 don't follow the (somewhat unwritten) rules and this strategy fails.
9197 Now just gather all certificates together and the first private key
9198 then look for the first certificate that matches the key.
9202 * Support use of registered digest and cipher names for dgst and cipher
9203 commands instead of having to add each one as a special case. So now
9210 openssl dgst -sha256 foo
9212 and this works for ENGINE based algorithms too.
9216 * Update Gost ENGINE to support parameter files.
9218 *Victor B. Wagner <vitus@cryptocom.ru>*
9220 * Support GeneralizedTime in ca utility.
9222 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9224 * Enhance the hash format used for certificate directory links. The new
9225 form uses the canonical encoding (meaning equivalent names will work
9226 even if they aren't identical) and uses SHA1 instead of MD5. This form
9227 is incompatible with the older format and as a result c_rehash should
9228 be used to rebuild symbolic links.
9232 * Make PKCS#8 the default write format for private keys, replacing the
9233 traditional format. This form is standardised, more secure and doesn't
9234 include an implicit MD5 dependency.
9238 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9239 committed to OpenSSL should pass this lot as a minimum.
9243 * Add session ticket override functionality for use by EAP-FAST.
9245 *Jouni Malinen <j@w1.fi>*
9247 * Modify HMAC functions to return a value. Since these can be implemented
9248 in an ENGINE errors can occur.
9252 * Type-checked OBJ_bsearch_ex.
9256 * Type-checked OBJ_bsearch. Also some constification necessitated
9257 by type-checking. Still to come: TXT_DB, bsearch(?),
9258 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9263 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9264 seconds to a tm structure directly, instead of going through OS
9265 specific date routines. This avoids any issues with OS routines such
9266 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9267 and X509_time_adj_ex() to cover the extended range. The existing
9268 X509_time_adj() is still usable and will no longer have any date issues.
9272 * Delta CRL support. New use deltas option which will attempt to locate
9273 and search any appropriate delta CRLs available.
9275 This work was sponsored by Google.
9279 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9280 code and add additional score elements. Validate alternate CRL paths
9281 as part of the CRL checking and indicate a new error "CRL path validation
9282 error" in this case. Applications wanting additional details can use
9283 the verify callback and check the new "parent" field. If this is not
9284 NULL CRL path validation is taking place. Existing applications won't
9285 see this because it requires extended CRL support which is off by
9288 This work was sponsored by Google.
9292 * Support for freshest CRL extension.
9294 This work was sponsored by Google.
9298 * Initial indirect CRL support. Currently only supported in the CRLs
9299 passed directly and not via lookup. Process certificate issuer
9300 CRL entry extension and lookup CRL entries by bother issuer name
9301 and serial number. Check and process CRL issuer entry in IDP extension.
9303 This work was sponsored by Google.
9307 * Add support for distinct certificate and CRL paths. The CRL issuer
9308 certificate is validated separately in this case. Only enabled if
9309 an extended CRL support flag is set: this flag will enable additional
9310 CRL functionality in future.
9312 This work was sponsored by Google.
9316 * Add support for policy mappings extension.
9318 This work was sponsored by Google.
9322 * Fixes to pathlength constraint, self issued certificate handling,
9323 policy processing to align with RFC3280 and PKITS tests.
9325 This work was sponsored by Google.
9329 * Support for name constraints certificate extension. DN, email, DNS
9330 and URI types are currently supported.
9332 This work was sponsored by Google.
9336 * To cater for systems that provide a pointer-based thread ID rather
9337 than numeric, deprecate the current numeric thread ID mechanism and
9338 replace it with a structure and associated callback type. This
9339 mechanism allows a numeric "hash" to be extracted from a thread ID in
9340 either case, and on platforms where pointers are larger than 'long',
9341 mixing is done to help ensure the numeric 'hash' is usable even if it
9342 can't be guaranteed unique. The default mechanism is to use "&errno"
9343 as a pointer-based thread ID to distinguish between threads.
9345 Applications that want to provide their own thread IDs should now use
9346 CRYPTO_THREADID_set_callback() to register a callback that will call
9347 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9349 Note that ERR_remove_state() is now deprecated, because it is tied
9350 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9351 to free the current thread's error state should be replaced by
9352 ERR_remove_thread_state(NULL).
9354 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9355 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9356 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9357 application was previously providing a numeric thread callback that
9358 was inappropriate for distinguishing threads, then uniqueness might
9359 have been obtained with &errno that happened immediately in the
9360 intermediate development versions of OpenSSL; this is no longer the
9361 case, the numeric thread callback will now override the automatic use
9364 *Geoff Thorpe, with help from Bodo Moeller*
9366 * Initial support for different CRL issuing certificates. This covers a
9367 simple case where the self issued certificates in the chain exist and
9368 the real CRL issuer is higher in the existing chain.
9370 This work was sponsored by Google.
9374 * Removed effectively defunct crypto/store from the build.
9378 * Revamp of STACK to provide stronger type-checking. Still to come:
9379 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9380 ASN1_STRING, CONF_VALUE.
9384 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9385 RAM on SSL connections. This option can save about 34k per idle SSL.
9389 * Revamp of LHASH to provide stronger type-checking. Still to come:
9390 STACK, TXT_DB, bsearch, qsort.
9394 * Initial support for Cryptographic Message Syntax (aka CMS) based
9395 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9396 support for data, signedData, compressedData, digestedData and
9397 encryptedData, envelopedData types included. Scripts to check against
9398 RFC4134 examples draft and interop and consistency checks of many
9399 content types and variants.
9403 * Add options to enc utility to support use of zlib compression BIO.
9407 * Extend mk1mf to support importing of options and assembly language
9408 files from Configure script, currently only included in VC-WIN32.
9409 The assembly language rules can now optionally generate the source
9410 files from the associated perl scripts.
9414 * Implement remaining functionality needed to support GOST ciphersuites.
9415 Interop testing has been performed using CryptoPro implementations.
9417 *Victor B. Wagner <vitus@cryptocom.ru>*
9419 * s390x assembler pack.
9423 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9428 * Implement Opaque PRF Input TLS extension as specified in
9429 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9430 official specification yet and no extension type assignment by
9431 IANA exists, this extension (for now) will have to be explicitly
9432 enabled when building OpenSSL by providing the extension number
9433 to use. For example, specify an option
9435 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9437 to the "config" or "Configure" script to enable the extension,
9438 assuming extension number 0x9527 (which is a completely arbitrary
9439 and unofficial assignment based on the MD5 hash of the Internet
9440 Draft). Note that by doing so, you potentially lose
9441 interoperability with other TLS implementations since these might
9442 be using the same extension number for other purposes.
9444 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9445 opaque PRF input value to use in the handshake. This will create
9446 an internal copy of the length-'len' string at 'src', and will
9447 return non-zero for success.
9449 To get more control and flexibility, provide a callback function
9452 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9453 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9457 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9460 Callback function 'cb' will be called in handshakes, and is
9461 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9462 Argument 'arg' is for application purposes (the value as given to
9463 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9464 be provided to the callback function). The callback function
9465 has to return non-zero to report success: usually 1 to use opaque
9466 PRF input just if possible, or 2 to enforce use of the opaque PRF
9467 input. In the latter case, the library will abort the handshake
9468 if opaque PRF input is not successfully negotiated.
9470 Arguments 'peerinput' and 'len' given to the callback function
9471 will always be NULL and 0 in the case of a client. A server will
9472 see the client's opaque PRF input through these variables if
9473 available (NULL and 0 otherwise). Note that if the server
9474 provides an opaque PRF input, the length must be the same as the
9475 length of the client's opaque PRF input.
9477 Note that the callback function will only be called when creating
9478 a new session (session resumption can resume whatever was
9479 previously negotiated), and will not be called in SSL 2.0
9480 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9481 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9482 for applications that need to enforce opaque PRF input.
9486 * Update ssl code to support digests other than SHA1+MD5 for handshake
9489 *Victor B. Wagner <vitus@cryptocom.ru>*
9491 * Add RFC4507 support to OpenSSL. This includes the corrections in
9492 RFC4507bis. The encrypted ticket format is an encrypted encoded
9493 SSL_SESSION structure, that way new session features are automatically
9496 If a client application caches session in an SSL_SESSION structure
9497 support is transparent because tickets are now stored in the encoded
9500 The SSL_CTX structure automatically generates keys for ticket
9501 protection in servers so again support should be possible
9502 with no application modification.
9504 If a client or server wishes to disable RFC4507 support then the option
9505 SSL_OP_NO_TICKET can be set.
9507 Add a TLS extension debugging callback to allow the contents of any client
9508 or server extensions to be examined.
9510 This work was sponsored by Google.
9514 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9515 OpenSSL should now compile cleanly on gcc 4.2
9517 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9519 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9520 support including streaming MAC support: this is required for GOST
9521 ciphersuite support.
9523 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9525 * Add option -stream to use PKCS#7 streaming in smime utility. New
9526 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9527 to output in BER and PEM format.
9531 * Experimental support for use of HMAC via EVP_PKEY interface. This
9532 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9533 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9534 ENGINE support for HMAC keys which are unextractable. New -mac and
9535 -macopt options to dgst utility.
9539 * New option -sigopt to dgst utility. Update dgst to use
9540 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9541 alternative signing parameters such as X9.31 or PSS in the dgst
9546 * Change ssl_cipher_apply_rule(), the internal function that does
9547 the work each time a ciphersuite string requests enabling
9548 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9549 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9550 the order of disabled ciphersuites such that those ciphersuites
9551 that most recently went from enabled to disabled not only stay
9552 in order with respect to each other, but also have higher priority
9553 than other disabled ciphersuites the next time ciphersuites are
9556 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9557 the same ciphersuites as with "HIGH" alone, but in a specific
9558 order where the PSK ciphersuites come first (since they are the
9559 most recently disabled ciphersuites when "HIGH" is parsed).
9561 Also, change ssl_create_cipher_list() (using this new
9562 functionality) such that between otherwise identical
9563 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9568 * Change ssl_create_cipher_list() so that it automatically
9569 arranges the ciphersuites in reasonable order before starting
9570 to process the rule string. Thus, the definition for "DEFAULT"
9571 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9572 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9573 This makes it much easier to arrive at a reasonable default order
9574 in applications for which anonymous ciphers are OK (meaning
9575 that you can't actually use DEFAULT).
9577 *Bodo Moeller; suggested by Victor Duchovni*
9579 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9580 processing) into multiple integers instead of setting
9581 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9582 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9583 (These masks as well as the individual bit definitions are hidden
9584 away into the non-exported interface ssl/ssl_locl.h, so this
9585 change to the definition of the SSL_CIPHER structure shouldn't
9586 affect applications.) This give us more bits for each of these
9587 categories, so there is no longer a need to coagulate AES128 and
9588 AES256 into a single algorithm bit, and to coagulate Camellia128
9589 and Camellia256 into a single algorithm bit, which has led to all
9592 Thus, among other things, the kludge introduced in 0.9.7m and
9593 0.9.8e for masking out AES256 independently of AES128 or masking
9594 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9596 With the change, we also introduce new ciphersuite aliases that
9597 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9602 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9603 Use the leftmost N bytes of the signature input if the input is
9604 larger than the prime q (with N being the size in bytes of q).
9608 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9609 it yet and it is largely untested.
9613 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9617 * Initial incomplete changes to avoid need for function casts in OpenSSL
9618 some compilers (gcc 4.2 and later) reject their use. Safestack is
9619 reimplemented. Update ASN1 to avoid use of legacy functions.
9623 * Win32/64 targets are linked with Winsock2.
9627 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9628 to external functions. This can be used to increase CRL handling
9629 efficiency especially when CRLs are very large by (for example) storing
9630 the CRL revoked certificates in a database.
9634 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9635 new CRLs added to a directory can be used. New command line option
9636 -verify_return_error to s_client and s_server. This causes real errors
9637 to be returned by the verify callback instead of carrying on no matter
9638 what. This reflects the way a "real world" verify callback would behave.
9642 * GOST engine, supporting several GOST algorithms and public key formats.
9643 Kindly donated by Cryptocom.
9647 * Partial support for Issuing Distribution Point CRL extension. CRLs
9648 partitioned by DP are handled but no indirect CRL or reason partitioning
9649 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9650 selected via a scoring technique which handles IDP and AKID in CRLs.
9654 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9655 will ultimately be used for all verify operations: this will remove the
9656 X509_STORE dependency on certificate verification and allow alternative
9657 lookup methods. X509_STORE based implementations of these two callbacks.
9661 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9662 Modify get_crl() to find a valid (unexpired) CRL if possible.
9666 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9667 this would be called X509_CRL_cmp() but that name is already used by
9668 a function that just compares CRL issuer names. Cache several CRL
9669 extensions in X509_CRL structure and cache CRLDP in X509.
9673 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9674 this maps equivalent X509_NAME structures into a consistent structure.
9675 Name comparison can then be performed rapidly using memcmp().
9679 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9684 * Allow digests to supply their own micalg string for S/MIME type using
9685 the ctrl EVP_MD_CTRL_MICALG.
9689 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9690 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9691 ctrl. It can then customise the structure before and/or after signing
9696 * New function OBJ_add_sigid() to allow application defined signature OIDs
9697 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9698 to free up any added signature OIDs.
9702 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9703 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9704 digest and cipher tables. New options added to openssl utility:
9705 list-message-digest-algorithms and list-cipher-algorithms.
9709 * Change the array representation of binary polynomials: the list
9710 of degrees of non-zero coefficients is now terminated with -1.
9711 Previously it was terminated with 0, which was also part of the
9712 value; thus, the array representation was not applicable to
9713 polynomials where t^0 has coefficient zero. This change makes
9714 the array representation useful in a more general context.
9718 * Various modifications and fixes to SSL/TLS cipher string
9719 handling. For ECC, the code now distinguishes between fixed ECDH
9720 with RSA certificates on the one hand and with ECDSA certificates
9721 on the other hand, since these are separate ciphersuites. The
9722 unused code for Fortezza ciphersuites has been removed.
9724 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9725 (not "ECDHE"). For consistency with the code for DH
9726 certificates, use of ECDH certificates is now considered ECDH
9727 authentication, not RSA or ECDSA authentication (the latter is
9728 merely the CA's signing algorithm and not actively used in the
9731 The temporary ciphersuite alias "ECCdraft" is no longer
9732 available, and ECC ciphersuites are no longer excluded from "ALL"
9733 and "DEFAULT". The following aliases now exist for RFC 4492
9734 ciphersuites, most of these by analogy with the DH case:
9736 kECDHr - ECDH cert, signed with RSA
9737 kECDHe - ECDH cert, signed with ECDSA
9738 kECDH - ECDH cert (signed with either RSA or ECDSA)
9739 kEECDH - ephemeral ECDH
9740 ECDH - ECDH cert or ephemeral ECDH
9746 AECDH - anonymous ECDH
9747 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9751 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9752 Use correct micalg parameters depending on digest(s) in signed message.
9756 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9757 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9761 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9762 an engine to register a method. Add ENGINE lookups for methods and
9763 functional reference processing.
9767 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9768 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9773 * New -resign option to smime utility. This adds one or more signers
9774 to an existing PKCS#7 signedData structure. Also -md option to use an
9775 alternative message digest algorithm for signing.
9779 * Tidy up PKCS#7 routines and add new functions to make it easier to
9780 create PKCS7 structures containing multiple signers. Update smime
9781 application to support multiple signers.
9785 * New -macalg option to pkcs12 utility to allow setting of an alternative
9790 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9791 Reorganize PBE internals to lookup from a static table using NIDs,
9792 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9793 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9794 PRF which will be automatically used with PBES2.
9798 * Replace the algorithm specific calls to generate keys in "req" with the
9803 * Update PKCS#7 enveloped data routines to use new API. This is now
9804 supported by any public key method supporting the encrypt operation. A
9805 ctrl is added to allow the public key algorithm to examine or modify
9806 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9811 * Add a ctrl to asn1 method to allow a public key algorithm to express
9812 a default digest type to use. In most cases this will be SHA1 but some
9813 algorithms (such as GOST) need to specify an alternative digest. The
9814 return value indicates how strong the preference is 1 means optional and
9815 2 is mandatory (that is it is the only supported type). Modify
9816 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9817 use the default md. Update openssl utilities to use the default digest
9818 type for signing if it is not explicitly indicated.
9822 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9823 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9824 signing method from the key type. This effectively removes the link
9825 between digests and public key types.
9829 * Add an OID cross reference table and utility functions. Its purpose is to
9830 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9831 rsaEncryption. This will allow some of the algorithm specific hackery
9832 needed to use the correct OID to be removed.
9836 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9837 structures for PKCS7_sign(). They are now set up by the relevant public
9842 * Add provisional EC pkey method with support for ECDSA and ECDH.
9846 * Add support for key derivation (agreement) in the API, DH method and
9851 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9852 public and private key formats. As a side effect these add additional
9853 command line functionality not previously available: DSA signatures can be
9854 generated and verified using pkeyutl and DH key support and generation in
9861 *Oliver Tappe <zooey@hirschkaefer.de>*
9863 * New make target "install_html_docs" installs HTML renditions of the
9866 *Oliver Tappe <zooey@hirschkaefer.de>*
9868 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9869 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9870 support key and parameter generation and add initial key generation
9871 functionality for RSA.
9875 * Add functions for main EVP_PKEY_method operations. The undocumented
9876 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9877 `EVP_PKEY_{encrypt,decrypt}_old`.
9881 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9882 key API, doesn't do much yet.
9886 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9887 public key algorithms. New option to openssl utility:
9888 "list-public-key-algorithms" to print out info.
9892 * Implement the Supported Elliptic Curves Extension for
9893 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9897 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9898 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9902 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9903 utilities such as rsa, dsa, dsaparam etc except they process any key
9908 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9909 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9910 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9915 * Initial support for pluggable public key ASN1.
9916 De-spaghettify the public key ASN1 handling. Move public and private
9917 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9918 algorithm specific handling to a single module within the relevant
9919 algorithm directory. Add functions to allow (near) opaque processing
9920 of public and private key structures.
9924 * Implement the Supported Point Formats Extension for
9925 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9929 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9930 for the psk identity [hint] and the psk callback functions to the
9931 SSL_SESSION, SSL and SSL_CTX structure.
9934 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9938 SSL_CTX_use_psk_identity_hint
9939 SSL_get_psk_identity_hint
9940 SSL_get_psk_identity
9941 SSL_use_psk_identity_hint
9943 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9945 * Add RFC 3161 compliant time stamp request creation, response generation
9946 and response verification functionality.
9948 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9950 * Add initial support for TLS extensions, specifically for the server_name
9951 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9952 have new members for a hostname. The SSL data structure has an
9953 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9954 stored in that context to allow for session resumption, even after the
9955 SSL has been switched to a new SSL_CTX in reaction to a client's
9956 server_name extension.
9958 New functions (subject to change):
9960 SSL_get_servername()
9961 SSL_get_servername_type()
9964 New CTRL codes and macros (subject to change):
9966 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9967 - SSL_CTX_set_tlsext_servername_callback()
9968 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9969 - SSL_CTX_set_tlsext_servername_arg()
9970 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9972 openssl s_client has a new '-servername ...' option.
9974 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9975 '-key2 ...', '-servername_fatal' (subject to change). This allows
9976 testing the HostName extension for a specific single hostname ('-cert'
9977 and '-key' remain fallbacks for handshakes without HostName
9978 negotiation). If the unrecognized_name alert has to be sent, this by
9979 default is a warning; it becomes fatal with the '-servername_fatal'
9982 *Peter Sylvester, Remy Allais, Christophe Renou*
9984 * Whirlpool hash implementation is added.
9988 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9989 bn(64,32). Because of instruction set limitations it doesn't have
9990 any negative impact on performance. This was done mostly in order
9991 to make it possible to share assembler modules, such as bn_mul_mont
9992 implementations, between 32- and 64-bit builds without hassle.
9996 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9997 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10002 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10003 dedicated Montgomery multiplication procedure, is introduced.
10004 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10005 "64-bit" performance on certain 32-bit targets.
10009 * New option SSL_OP_NO_COMP to disable use of compression selectively
10010 in SSL structures. New SSL ctrl to set maximum send fragment size.
10011 Save memory by setting the I/O buffer sizes dynamically instead of
10012 using the maximum available value.
10016 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10017 in addition to the text details.
10021 * Very, very preliminary EXPERIMENTAL support for printing of general
10022 ASN1 structures. This currently produces rather ugly output and doesn't
10023 handle several customised structures at all.
10027 * Integrated support for PVK file format and some related formats such
10028 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10029 these in the 'rsa' and 'dsa' utilities.
10033 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10037 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10038 place for the (very old) "NETSCAPE" format certificates which are now
10039 handled using new ASN1 code equivalents.
10043 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10044 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10045 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10049 * Modify CRL distribution points extension code to print out previously
10050 unsupported fields. Enhance extension setting code to allow setting of
10055 * Add print and set support for Issuing Distribution Point CRL extension.
10059 * Change 'Configure' script to enable Camellia by default.
10066 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10068 * When rejecting SSL/TLS records due to an incorrect version number, never
10069 update s->server with a new major version number. As of
10070 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10071 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10072 the previous behavior could result in a read attempt at NULL when
10073 receiving specific incorrect SSL/TLS records once record payload
10074 protection is active. ([CVE-2010-0740])
10076 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10078 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10079 could be crashed if the relevant tables were not present (e.g. chrooted).
10081 *Tomas Hoger <thoger@redhat.com>*
10083 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10085 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10087 *Martin Olsson, Neel Mehta*
10089 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10090 accommodate for stack sorting, always a write lock!).
10094 * On some versions of WIN32 Heap32Next is very slow. This can cause
10095 excessive delays in the RAND_poll(): over a minute. As a workaround
10096 include a time check in the inner Heap32Next loop too.
10100 * The code that handled flushing of data in SSL/TLS originally used the
10101 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10102 the problem outlined in PR#1949. The fix suggested there however can
10103 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10104 of Apache). So instead simplify the code to flush unconditionally.
10105 This should be fine since flushing with no data to flush is a no op.
10109 * Handle TLS versions 2.0 and later properly and correctly use the
10110 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10111 off ancient servers have a habit of sticking around for a while...
10115 * Modify compression code so it frees up structures without using the
10116 ex_data callbacks. This works around a problem where some applications
10117 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10118 restarting) then use compression (e.g. SSL with compression) later.
10119 This results in significant per-connection memory leaks and
10120 has caused some security issues including CVE-2008-1678 and
10125 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10126 change when encrypting or decrypting.
10130 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10131 connect and renegotiate with servers which do not support RI.
10132 Until RI is more widely deployed this option is enabled by default.
10136 * Add "missing" ssl ctrls to clear options and mode.
10140 * If client attempts to renegotiate and doesn't support RI respond with
10141 a no_renegotiation alert as required by RFC5746. Some renegotiating
10142 TLS clients will continue a connection gracefully when they receive
10143 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10144 waiting for a server hello which it will never receive. Now we treat a
10145 received no_renegotiation alert as a fatal error. This is because
10146 applications requesting a renegotiation might well expect it to succeed
10147 and would have no code in place to handle the server denying it so the
10148 only safe thing to do is to terminate the connection.
10152 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10153 peer supports secure renegotiation and 0 otherwise. Print out peer
10154 renegotiation support in s_client/s_server.
10158 * Replace the highly broken and deprecated SPKAC certification method with
10159 the updated NID creation version. This should correctly handle UTF8.
10163 * Implement RFC5746. Re-enable renegotiation but require the extension
10164 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10165 turns out to be a bad idea. It has been replaced by
10166 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10167 SSL_CTX_set_options(). This is really not recommended unless you
10168 know what you are doing.
10170 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10172 * Fixes to stateless session resumption handling. Use initial_ctx when
10173 issuing and attempting to decrypt tickets in case it has changed during
10174 servername handling. Use a non-zero length session ID when attempting
10175 stateless session resumption: this makes it possible to determine if
10176 a resumption has occurred immediately after receiving server hello
10177 (several places in OpenSSL subtly assume this) instead of later in
10182 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10183 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10184 fixes for a few places where the return code is not checked
10187 *Julia Lawall <julia@diku.dk>*
10189 * Add --strict-warnings option to Configure script to include devteam
10190 warnings in other configurations.
10194 * Add support for --libdir option and LIBDIR variable in makefiles. This
10195 makes it possible to install openssl libraries in locations which
10196 have names other than "lib", for example "/usr/lib64" which some
10199 *Steve Henson, based on patch from Jeremy Utley*
10201 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10202 X690 8.9.12 and can produce some misleading textual output of OIDs.
10204 *Steve Henson, reported by Dan Kaminsky*
10206 * Delete MD2 from algorithm tables. This follows the recommendation in
10207 several standards that it is not used in new applications due to
10208 several cryptographic weaknesses. For binary compatibility reasons
10209 the MD2 API is still compiled in by default.
10213 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10218 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10219 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10222 *Guenter <lists@gknw.net>*
10224 * Fix the server certificate chain building code to use X509_verify_cert(),
10225 it used to have an ad-hoc builder which was unable to cope with anything
10226 other than a simple chain.
10228 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10230 * Don't check self signed certificate signatures in X509_verify_cert()
10231 by default (a flag can override this): it just wastes time without
10232 adding any security. As a useful side effect self signed root CAs
10233 with non-FIPS digests are now usable in FIPS mode.
10237 * In dtls1_process_out_of_seq_message() the check if the current message
10238 is already buffered was missing. For every new message was memory
10239 allocated, allowing an attacker to perform an denial of service attack
10240 with sending out of seq handshake messages until there is no memory
10241 left. Additionally every future message was buffered, even if the
10242 sequence number made no sense and would be part of another handshake.
10243 So only messages with sequence numbers less than 10 in advance will be
10244 buffered. ([CVE-2009-1378])
10246 *Robin Seggelmann, discovered by Daniel Mentz*
10248 * Records are buffered if they arrive with a future epoch to be
10249 processed after finishing the corresponding handshake. There is
10250 currently no limitation to this buffer allowing an attacker to perform
10251 a DOS attack with sending records with future epochs until there is no
10252 memory left. This patch adds the pqueue_size() function to determine
10253 the size of a buffer and limits the record buffer to 100 entries.
10256 *Robin Seggelmann, discovered by Daniel Mentz*
10258 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10259 parent structure is freed. ([CVE-2009-1379])
10263 * Handle non-blocking I/O properly in SSL_shutdown() call.
10265 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10267 * Add `2.5.4.*` OIDs
10269 *Ilya O. <vrghost@gmail.com>*
10271 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10273 * Disable renegotiation completely - this fixes a severe security
10274 problem ([CVE-2009-3555]) at the cost of breaking all
10275 renegotiation. Renegotiation can be re-enabled by setting
10276 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10277 run-time. This is really not recommended unless you know what
10282 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10284 * Don't set val to NULL when freeing up structures, it is freed up by
10285 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10286 zeroing past the valid field. ([CVE-2009-0789])
10288 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10290 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10291 checked correctly. This would allow some invalid signed attributes to
10292 appear to verify correctly. ([CVE-2009-0591])
10294 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10296 * Reject UniversalString and BMPString types with invalid lengths. This
10297 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10298 a legal length. ([CVE-2009-0590])
10302 * Set S/MIME signing as the default purpose rather than setting it
10303 unconditionally. This allows applications to override it at the store
10308 * Permit restricted recursion of ASN1 strings. This is needed in practice
10309 to handle some structures.
10313 * Improve efficiency of mem_gets: don't search whole buffer each time
10316 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10318 * New -hex option for openssl rand.
10322 * Print out UTF8String and NumericString when parsing ASN1.
10326 * Support NumericString type for name components.
10330 * Allow CC in the environment to override the automatically chosen
10331 compiler. Note that nothing is done to ensure flags work with the
10336 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10338 * Properly check EVP_VerifyFinal() and similar return values
10341 *Ben Laurie, Bodo Moeller, Google Security Team*
10343 * Enable TLS extensions by default.
10347 * Allow the CHIL engine to be loaded, whether the application is
10348 multithreaded or not. (This does not release the developer from the
10349 obligation to set up the dynamic locking callbacks.)
10351 *Sander Temme <sander@temme.net>*
10353 * Use correct exit code if there is an error in dgst command.
10355 *Steve Henson; problem pointed out by Roland Dirlewanger*
10357 * Tweak Configure so that you need to say "experimental-jpake" to enable
10358 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10362 * Add experimental JPAKE support, including demo authentication in
10363 s_client and s_server.
10367 * Set the comparison function in v3_addr_canonize().
10369 *Rob Austein <sra@hactrn.net>*
10371 * Add support for XMPP STARTTLS in s_client.
10373 *Philip Paeps <philip@freebsd.org>*
10375 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10376 to ensure that even with this option, only ciphersuites in the
10377 server's preference list will be accepted. (Note that the option
10378 applies only when resuming a session, so the earlier behavior was
10379 just about the algorithm choice for symmetric cryptography.)
10383 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10385 * Fix NULL pointer dereference if a DTLS server received
10386 ChangeCipherSpec as first record ([CVE-2009-1386]).
10390 * Fix a state transition in s3_srvr.c and d1_srvr.c
10391 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10393 *Nagendra Modadugu*
10395 * The fix in 0.9.8c that supposedly got rid of unsafe
10396 double-checked locking was incomplete for RSA blinding,
10397 addressing just one layer of what turns out to have been
10398 doubly unsafe triple-checked locking.
10400 So now fix this for real by retiring the MONT_HELPER macro
10401 in crypto/rsa/rsa_eay.c.
10403 *Bodo Moeller; problem pointed out by Marius Schilder*
10405 * Various precautionary measures:
10407 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10409 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10410 (NB: This would require knowledge of the secret session ticket key
10411 to exploit, in which case you'd be SOL either way.)
10413 - Change bn_nist.c so that it will properly handle input BIGNUMs
10414 outside the expected range.
10416 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10419 *Neel Mehta, Bodo Moeller*
10421 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10422 the load fails. Useful for distros.
10424 *Ben Laurie and the FreeBSD team*
10426 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10430 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10434 * Expand ENGINE to support engine supplied SSL client certificate functions.
10436 This work was sponsored by Logica.
10440 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10441 keystores. Support for SSL/TLS client authentication too.
10442 Not compiled unless enable-capieng specified to Configure.
10444 This work was sponsored by Logica.
10448 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10449 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10450 attribute creation routines such as certificate requests and PKCS#12
10455 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10457 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10458 handshake which could lead to a client crash as found using the
10459 Codenomicon TLS test suite ([CVE-2008-1672])
10461 *Steve Henson, Mark Cox*
10463 * Fix double free in TLS server name extensions which could lead to
10464 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10468 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10470 Clear the error queue to ensure that error entries left from
10471 older function calls do not interfere with the correct operation.
10473 *Lutz Jaenicke, Erik de Castro Lopo*
10475 * Remove root CA certificates of commercial CAs:
10477 The OpenSSL project does not recommend any specific CA and does not
10478 have any policy with respect to including or excluding any CA.
10479 Therefore, it does not make any sense to ship an arbitrary selection
10480 of root CA certificates with the OpenSSL software.
10484 * RSA OAEP patches to fix two separate invalid memory reads.
10485 The first one involves inputs when 'lzero' is greater than
10486 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10487 before the beginning of from). The second one involves inputs where
10488 the 'db' section contains nothing but zeroes (there is a one-byte
10489 invalid read after the end of 'db').
10491 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10493 * Partial backport from 0.9.9-dev:
10495 Introduce bn_mul_mont (dedicated Montgomery multiplication
10496 procedure) as a candidate for BIGNUM assembler implementation.
10497 While 0.9.9-dev uses assembler for various architectures, only
10498 x86_64 is available by default here in the 0.9.8 branch, and
10499 32-bit x86 is available through a compile-time setting.
10501 To try the 32-bit x86 assembler implementation, use Configure
10502 option "enable-montasm" (which exists only for this backport).
10504 As "enable-montasm" for 32-bit x86 disclaims code stability
10505 anyway, in this constellation we activate additional code
10506 backported from 0.9.9-dev for further performance improvements,
10507 namely BN_from_montgomery_word. (To enable this otherwise,
10508 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10510 *Andy Polyakov (backport partially by Bodo Moeller)*
10512 * Add TLS session ticket callback. This allows an application to set
10513 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10514 values. This is useful for key rollover for example where several key
10515 sets may exist with different names.
10519 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10520 This was broken until now in 0.9.8 releases, such that the only way
10521 a registered ENGINE could be used (assuming it initialises
10522 successfully on the host) was to explicitly set it as the default
10523 for the relevant algorithms. This is in contradiction with 0.9.7
10524 behaviour and the documentation. With this fix, when an ENGINE is
10525 registered into a given algorithm's table of implementations, the
10526 'uptodate' flag is reset so that auto-discovery will be used next
10527 time a new context for that algorithm attempts to select an
10530 *Ian Lister (tweaked by Geoff Thorpe)*
10532 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10533 implementation in the following ways:
10535 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10538 Lack of BER streaming support means one pass streaming processing is
10539 only supported if data is detached: setting the streaming flag is
10540 ignored for embedded content.
10542 CMS support is disabled by default and must be explicitly enabled
10543 with the enable-cms configuration option.
10547 * Update the GMP engine glue to do direct copies between BIGNUM and
10548 mpz_t when openssl and GMP use the same limb size. Otherwise the
10549 existing "conversion via a text string export" trick is still used.
10551 *Paul Sheer <paulsheer@gmail.com>*
10553 * Zlib compression BIO. This is a filter BIO which compressed and
10554 uncompresses any data passed through it.
10558 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10559 RFC3394 compatible AES key wrapping.
10563 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10564 sets string data without copying. X509_ALGOR_set0() and
10565 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10566 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10567 from an X509_ATTRIBUTE structure optionally checking it occurs only
10568 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10573 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10574 to get the expected BN_FLG_CONSTTIME behavior.
10576 *Bodo Moeller (Google)*
10580 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10581 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10582 - added some more tests to do_tests.pl
10583 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10584 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10585 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10586 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10587 - various changes to netware.pl to enable gcc-cross builds on Win32
10589 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10590 - various changes to fix missing prototype warnings
10591 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10592 - added AES, WHIRLPOOL and CPUID assembler code to build files
10593 - added missing AES assembler make rules to mk1mf.pl
10594 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10596 *Guenter Knauf <eflash@gmx.net>*
10598 * Implement certificate status request TLS extension defined in RFC3546.
10599 A client can set the appropriate parameters and receive the encoded
10600 OCSP response via a callback. A server can query the supplied parameters
10601 and set the encoded OCSP response in the callback. Add simplified examples
10602 to s_client and s_server.
10606 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10608 * Fix various bugs:
10609 + Binary incompatibility of ssl_ctx_st structure
10610 + DTLS interoperation with non-compliant servers
10611 + Don't call get_session_cb() without proposed session
10612 + Fix ia64 assembler code
10614 *Andy Polyakov, Steve Henson*
10616 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10618 * DTLS Handshake overhaul. There were longstanding issues with
10619 OpenSSL DTLS implementation, which were making it impossible for
10620 RFC 4347 compliant client to communicate with OpenSSL server.
10621 Unfortunately just fixing these incompatibilities would "cut off"
10622 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10623 server keeps tolerating non RFC compliant syntax. The opposite is
10624 not true, 0.9.8f client can not communicate with earlier server.
10625 This update even addresses CVE-2007-4995.
10629 * Changes to avoid need for function casts in OpenSSL: some compilers
10630 (gcc 4.2 and later) reject their use.
10631 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10634 * Add RFC4507 support to OpenSSL. This includes the corrections in
10635 RFC4507bis. The encrypted ticket format is an encrypted encoded
10636 SSL_SESSION structure, that way new session features are automatically
10639 If a client application caches session in an SSL_SESSION structure
10640 support is transparent because tickets are now stored in the encoded
10643 The SSL_CTX structure automatically generates keys for ticket
10644 protection in servers so again support should be possible
10645 with no application modification.
10647 If a client or server wishes to disable RFC4507 support then the option
10648 SSL_OP_NO_TICKET can be set.
10650 Add a TLS extension debugging callback to allow the contents of any client
10651 or server extensions to be examined.
10653 This work was sponsored by Google.
10657 * Add initial support for TLS extensions, specifically for the server_name
10658 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10659 have new members for a hostname. The SSL data structure has an
10660 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10661 stored in that context to allow for session resumption, even after the
10662 SSL has been switched to a new SSL_CTX in reaction to a client's
10663 server_name extension.
10665 New functions (subject to change):
10667 SSL_get_servername()
10668 SSL_get_servername_type()
10671 New CTRL codes and macros (subject to change):
10673 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10674 - SSL_CTX_set_tlsext_servername_callback()
10675 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10676 - SSL_CTX_set_tlsext_servername_arg()
10677 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10679 openssl s_client has a new '-servername ...' option.
10681 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10682 '-key2 ...', '-servername_fatal' (subject to change). This allows
10683 testing the HostName extension for a specific single hostname ('-cert'
10684 and '-key' remain fallbacks for handshakes without HostName
10685 negotiation). If the unrecognized_name alert has to be sent, this by
10686 default is a warning; it becomes fatal with the '-servername_fatal'
10689 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10691 * Add AES and SSE2 assembly language support to VC++ build.
10695 * Mitigate attack on final subtraction in Montgomery reduction.
10699 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10700 (which previously caused an internal error).
10704 * Squeeze another 10% out of IGE mode when in != out.
10708 * AES IGE mode speedup.
10710 *Dean Gaudet (Google)*
10712 * Add the Korean symmetric 128-bit cipher SEED (see
10713 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10714 add SEED ciphersuites from RFC 4162:
10716 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10717 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10718 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10719 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10721 To minimize changes between patchlevels in the OpenSSL 0.9.8
10722 series, SEED remains excluded from compilation unless OpenSSL
10723 is configured with 'enable-seed'.
10725 *KISA, Bodo Moeller*
10727 * Mitigate branch prediction attacks, which can be practical if a
10728 single processor is shared, allowing a spy process to extract
10729 information. For detailed background information, see
10730 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10731 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10732 and Necessary Software Countermeasures"). The core of the change
10733 are new versions BN_div_no_branch() and
10734 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10735 respectively, which are slower, but avoid the security-relevant
10736 conditional branches. These are automatically called by BN_div()
10737 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10738 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10739 remove a conditional branch.
10741 BN_FLG_CONSTTIME is the new name for the previous
10742 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10743 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10744 in the exponent causes BN_mod_exp_mont() to use the alternative
10745 implementation in BN_mod_exp_mont_consttime().) The old name
10746 remains as a deprecated alias.
10748 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10749 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10750 constant-time implementations for more than just exponentiation.
10751 Here too the old name is kept as a deprecated alias.
10753 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10754 the BN_BLINDING structure gets an independent copy of the
10755 modulus. This means that the previous `BIGNUM *m` argument to
10756 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10757 essentially becomes `const BIGNUM *m`, although we can't actually
10758 change this in the header file before 0.9.9. It allows
10759 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10760 enable BN_FLG_CONSTTIME.
10762 *Matthew D Wood (Intel Corp)*
10764 * In the SSL/TLS server implementation, be strict about session ID
10765 context matching (which matters if an application uses a single
10766 external cache for different purposes). Previously,
10767 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10768 set. This did ensure strict client verification, but meant that,
10769 with applications using a single external cache for quite
10770 different requirements, clients could circumvent ciphersuite
10771 restrictions for a given session ID context by starting a session
10772 in a different context.
10776 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10777 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10778 authentication-only ciphersuites.
10782 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10783 not complete and could lead to a possible single byte overflow
10784 ([CVE-2007-5135]) [Ben Laurie]
10786 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10788 * Since AES128 and AES256 (and similarly Camellia128 and
10789 Camellia256) share a single mask bit in the logic of
10790 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10791 kludge to work properly if AES128 is available and AES256 isn't
10792 (or if Camellia128 is available and Camellia256 isn't).
10796 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10797 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10798 When a point or a seed is encoded in a BIT STRING, we need to
10799 prevent the removal of trailing zero bits to get the proper DER
10800 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10801 of a NamedBitList, for which trailing 0 bits need to be removed.)
10805 * Have SSL/TLS server implementation tolerate "mismatched" record
10806 protocol version while receiving ClientHello even if the
10807 ClientHello is fragmented. (The server can't insist on the
10808 particular protocol version it has chosen before the ServerHello
10809 message has informed the client about his choice.)
10813 * Add RFC 3779 support.
10815 *Rob Austein for ARIN, Ben Laurie*
10817 * Load error codes if they are not already present instead of using a
10818 static variable. This allows them to be cleanly unloaded and reloaded.
10819 Improve header file function name parsing.
10823 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10824 or CAPABILITY handshake as required by RFCs.
10826 *Goetz Babin-Ebell*
10828 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10830 * Introduce limits to prevent malicious keys being able to
10831 cause a denial of service. ([CVE-2006-2940])
10833 *Steve Henson, Bodo Moeller*
10835 * Fix ASN.1 parsing of certain invalid structures that can result
10836 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10838 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10839 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10841 * Fix SSL client code which could crash if connecting to a
10842 malicious SSLv2 server. ([CVE-2006-4343])
10844 *Tavis Ormandy and Will Drewry, Google Security Team*
10846 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10847 match only those. Before that, "AES256-SHA" would be interpreted
10848 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10849 the same strength classification in 0.9.7h) as we currently only
10850 have a single AES bit in the ciphersuite description bitmap.
10851 That change, however, also applied to ciphersuite strings such as
10852 "RC4-MD5" that intentionally matched multiple ciphersuites --
10853 namely, SSL 2.0 ciphersuites in addition to the more common ones
10854 from SSL 3.0/TLS 1.0.
10856 So we change the selection algorithm again: Naming an explicit
10857 ciphersuite selects this one ciphersuite, and any other similar
10858 ciphersuite (same bitmap) from *other* protocol versions.
10859 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10860 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10862 Since SSL 2.0 does not have any ciphersuites for which the
10863 128/256 bit distinction would be relevant, this works for now.
10864 The proper fix will be to use different bits for AES128 and
10865 AES256, which would have avoided the problems from the beginning;
10866 however, bits are scarce, so we can only do this in a new release
10867 (not just a patchlevel) when we can change the SSL_CIPHER
10868 definition to split the single 'unsigned long mask' bitmap into
10869 multiple values to extend the available space.
10873 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10875 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10876 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10878 * Add AES IGE and biIGE modes.
10882 * Change the Unix randomness entropy gathering to use poll() when
10883 possible instead of select(), since the latter has some
10884 undesirable limitations.
10886 *Darryl Miles via Richard Levitte and Bodo Moeller*
10888 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10889 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10890 cannot be implicitly activated as part of, e.g., the "AES" alias.
10891 However, please upgrade to OpenSSL 0.9.9[-dev] for
10892 non-experimental use of the ECC ciphersuites to get TLS extension
10893 support, which is required for curve and point format negotiation
10894 to avoid potential handshake problems.
10898 * Disable rogue ciphersuites:
10900 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10901 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10902 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10904 The latter two were purportedly from
10905 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10908 Also deactivate the remaining ciphersuites from
10909 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10910 unofficial, and the ID has long expired.
10914 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10915 dual-core machines) and other potential thread-safety issues.
10919 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10920 versions), which is now available for royalty-free use
10921 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10922 Also, add Camellia TLS ciphersuites from RFC 4132.
10924 To minimize changes between patchlevels in the OpenSSL 0.9.8
10925 series, Camellia remains excluded from compilation unless OpenSSL
10926 is configured with 'enable-camellia'.
10930 * Disable the padding bug check when compression is in use. The padding
10931 bug check assumes the first packet is of even length, this is not
10932 necessarily true if compression is enabled and can result in false
10933 positives causing handshake failure. The actual bug test is ancient
10934 code so it is hoped that implementations will either have fixed it by
10935 now or any which still have the bug do not support compression.
10939 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10941 * When applying a cipher rule check to see if string match is an explicit
10942 cipher suite and only match that one cipher suite if it is.
10946 * Link in manifests for VC++ if needed.
10948 *Austin Ziegler <halostatue@gmail.com>*
10950 * Update support for ECC-based TLS ciphersuites according to
10951 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10952 TLS extensions, which are supported starting with the 0.9.9
10953 branch, not in the OpenSSL 0.9.8 branch).
10957 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10958 opaque EVP_CIPHER_CTX handling.
10962 * Fixes and enhancements to zlib compression code. We now only use
10963 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10964 to conform with the standards mentioned here:
10965 <http://www.zlib.net/DLL_FAQ.txt>
10966 Static zlib linking now works on Windows and the new --with-zlib-include
10967 --with-zlib-lib options to Configure can be used to supply the location
10968 of the headers and library. Gracefully handle case where zlib library
10973 * Several fixes and enhancements to the OID generation code. The old code
10974 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10975 handle numbers larger than ULONG_MAX, truncated printing and had a
10976 non standard OBJ_obj2txt() behaviour.
10980 * Add support for building of engines under engine/ as shared libraries
10981 under VC++ build system.
10985 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10986 Hopefully, we will not see any false combination of paths any more.
10990 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10992 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10993 (part of SSL_OP_ALL). This option used to disable the
10994 countermeasure against man-in-the-middle protocol-version
10995 rollback in the SSL 2.0 server implementation, which is a bad
10996 idea. ([CVE-2005-2969])
10998 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10999 for Information Security, National Institute of Advanced Industrial
11000 Science and Technology [AIST], Japan)*
11002 * Add two function to clear and return the verify parameter flags.
11006 * Keep cipherlists sorted in the source instead of sorting them at
11007 runtime, thus removing the need for a lock.
11011 * Avoid some small subgroup attacks in Diffie-Hellman.
11013 *Nick Mathewson and Ben Laurie*
11015 * Add functions for well-known primes.
11019 * Extended Windows CE support.
11021 *Satoshi Nakamura and Andy Polyakov*
11023 * Initialize SSL_METHOD structures at compile time instead of during
11024 runtime, thus removing the need for a lock.
11028 * Make PKCS7_decrypt() work even if no certificate is supplied by
11029 attempting to decrypt each encrypted key in turn. Add support to
11034 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11036 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11039 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11043 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11044 key into the same file any more.
11048 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11052 * Add -utf8 command line and config file option to 'ca'.
11054 *Stefan <stf@udoma.org*
11056 * Removed the macro des_crypt(), as it seems to conflict with some
11057 libraries. Use DES_crypt().
11061 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11062 involves renaming the source and generated shared-libs for
11063 both. The engines will accept the corrected or legacy ids
11064 ('ncipher' and '4758_cca' respectively) when binding. NB,
11065 this only applies when building 'shared'.
11067 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11069 * Add attribute functions to EVP_PKEY structure. Modify
11070 PKCS12_create() to recognize a CSP name attribute and
11071 use it. Make -CSP option work again in pkcs12 utility.
11075 * Add new functionality to the bn blinding code:
11076 - automatic re-creation of the BN_BLINDING parameters after
11077 a fixed number of uses (currently 32)
11078 - add new function for parameter creation
11079 - introduce flags to control the update behaviour of the
11080 BN_BLINDING parameters
11081 - hide BN_BLINDING structure
11082 Add a second BN_BLINDING slot to the RSA structure to improve
11083 performance when a single RSA object is shared among several
11088 * Add support for DTLS.
11090 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11092 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11093 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11097 * Remove buggy and incomplete DH cert support from
11098 ssl/ssl_rsa.c and ssl/s3_both.c
11102 * Use SHA-1 instead of MD5 as the default digest algorithm for
11103 the `apps/openssl` commands.
11107 * Compile clean with "-Wall -Wmissing-prototypes
11108 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11109 DEBUG_SAFESTACK must also be set.
11113 * Change ./Configure so that certain algorithms can be disabled by default.
11114 The new counterpiece to "no-xxx" is "enable-xxx".
11116 The patented RC5 and MDC2 algorithms will now be disabled unless
11117 "enable-rc5" and "enable-mdc2", respectively, are specified.
11119 (IDEA remains enabled despite being patented. This is because IDEA
11120 is frequently required for interoperability, and there is no license
11121 fee for non-commercial use. As before, "no-idea" can be used to
11122 avoid this algorithm.)
11126 * Add processing of proxy certificates (see RFC 3820). This work was
11127 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11128 EGEE (Enabling Grids for E-science in Europe).
11132 * RC4 performance overhaul on modern architectures/implementations, such
11133 as Intel P4, IA-64 and AMD64.
11137 * New utility extract-section.pl. This can be used specify an alternative
11138 section number in a pod file instead of having to treat each file as
11139 a separate case in Makefile. This can be done by adding two lines to the
11142 =for comment openssl_section:XXX
11144 The blank line is mandatory.
11148 * New arguments -certform, -keyform and -pass for s_client and s_server
11149 to allow alternative format key and certificate files and passphrase
11154 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11155 update associated structures and add various utility functions.
11157 Add new policy related verify parameters, include policy checking in
11158 standard verify code. Enhance 'smime' application with extra parameters
11159 to support policy checking and print out.
11163 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11164 Nehemiah processors. These extensions support AES encryption in hardware
11165 as well as RNG (though RNG support is currently disabled).
11167 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11169 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11173 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11175 *Andy Polyakov and a number of other people*
11177 * Improved PowerPC platform support. Most notably BIGNUM assembler
11178 implementation contributed by IBM.
11180 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11182 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11183 exponent rather than 'unsigned long'. There is a corresponding change to
11184 the new 'rsa_keygen' element of the RSA_METHOD structure.
11186 *Jelte Jansen, Geoff Thorpe*
11188 * Functionality for creating the initial serial number file is now
11189 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11191 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11192 number file to 1, which is bound to cause problems. To avoid
11193 the problems while respecting compatibility between different 0.9.7
11194 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11195 CA.pl for serial number initialization. With the new release 0.9.8,
11196 we can fix the problem directly in the 'ca' utility.)
11200 * Reduced header interdependencies by declaring more opaque objects in
11201 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11202 give fewer recursive includes, which could break lazy source code - so
11203 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11204 developers should define this symbol when building and using openssl to
11205 ensure they track the recommended behaviour, interfaces, [etc], but
11206 backwards-compatible behaviour prevails when this isn't defined.
11210 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11214 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11215 This will generate a random key of the appropriate length based on the
11216 cipher context. The EVP_CIPHER can provide its own random key generation
11217 routine to support keys of a specific form. This is used in the des and
11218 3des routines to generate a key of the correct parity. Update S/MIME
11219 code to use new functions and hence generate correct parity DES keys.
11220 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11221 valid (weak or incorrect parity).
11225 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11226 as looking them up. This is useful when the verified structure may contain
11227 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11228 present unless the new PKCS7_NO_CRL flag is asserted.
11232 * Extend ASN1 oid configuration module. It now additionally accepts the
11235 shortName = some long name, 1.2.3.4
11239 * Reimplemented the BN_CTX implementation. There is now no more static
11240 limitation on the number of variables it can handle nor the depth of the
11241 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11242 information can now expand as required, and rather than having a single
11243 static array of bignums, BN_CTX now uses a linked-list of such arrays
11244 allowing it to expand on demand whilst maintaining the usefulness of
11245 BN_CTX's "bundling".
11249 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11250 to allow all RSA operations to function using a single BN_CTX.
11254 * Preliminary support for certificate policy evaluation and checking. This
11255 is initially intended to pass the tests outlined in "Conformance Testing
11256 of Relying Party Client Certificate Path Processing Logic" v1.07.
11260 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11261 remained unused and not that useful. A variety of other little bignum
11262 tweaks and fixes have also been made continuing on from the audit (see
11267 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11268 associated ASN1, EVP and SSL functions and old ASN1 macros.
11272 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11273 and this should never fail. So the return value from the use of
11274 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11275 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11279 * BN_CTX_get() should return zero-valued bignums, providing the same
11280 initialised value as BN_new().
11282 *Geoff Thorpe, suggested by Ulf Möller*
11284 * Support for inhibitAnyPolicy certificate extension.
11288 * An audit of the BIGNUM code is underway, for which debugging code is
11289 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11290 is considered valid when processing BIGNUMs, and causes execution to
11291 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11292 further steps are taken to deliberately pollute unused data in BIGNUM
11293 structures to try and expose faulty code further on. For now, openssl will
11294 (in its default mode of operation) continue to tolerate the inconsistent
11295 forms that it has tolerated in the past, but authors and packagers should
11296 consider trying openssl and their own applications when compiled with
11297 these debugging symbols defined. It will help highlight potential bugs in
11298 their own code, and will improve the test coverage for OpenSSL itself. At
11299 some point, these tighter rules will become openssl's default to improve
11300 maintainability, though the assert()s and other overheads will remain only
11301 in debugging configurations. See bn.h for more details.
11303 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11305 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11306 that can only be obtained through BN_CTX_new() (which implicitly
11307 initialises it). The presence of this function only made it possible
11308 to overwrite an existing structure (and cause memory leaks).
11312 * Because of the callback-based approach for implementing LHASH as a
11313 template type, lh_insert() adds opaque objects to hash-tables and
11314 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11315 to clean up those corresponding objects before destroying the hash table
11316 (and losing the object pointers). So some over-zealous constifications in
11317 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11318 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11319 prototyped to have "const" restrictions on the object pointers they are
11320 given (and so aren't required to cast them away any more).
11324 * The tmdiff.h API was so ugly and minimal that our own timing utility
11325 (speed) prefers to use its own implementation. The two implementations
11326 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11327 its object type properly exposed (MS_TM) instead of casting to/from
11328 `char *`. This may still change yet if someone realises MS_TM and
11330 aren't necessarily the greatest nomenclatures - but this is what was used
11331 internally to the implementation so I've used that for now.
11335 * Ensure that deprecated functions do not get compiled when
11336 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11337 the self-tests were still using deprecated key-generation functions so
11338 these have been updated also.
11342 * Reorganise PKCS#7 code to separate the digest location functionality
11343 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11344 New function PKCS7_set_digest() to set the digest type for PKCS#7
11345 digestedData type. Add additional code to correctly generate the
11346 digestedData type and add support for this type in PKCS7 initialization
11351 * New function PKCS7_set0_type_other() this initializes a PKCS7
11352 structure of type "other".
11356 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11357 sure the loop does correctly stop and breaking ("division by zero")
11358 modulus operations are not performed. The (pre-generated) prime
11359 table crypto/bn/bn_prime.h was already correct, but it could not be
11360 re-generated on some platforms because of the "division by zero"
11361 situation in the script.
11363 *Ralf S. Engelschall*
11365 * Update support for ECC-based TLS ciphersuites according to
11366 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11367 SHA-1 now is only used for "small" curves (where the
11368 representation of a field element takes up to 24 bytes); for
11369 larger curves, the field element resulting from ECDH is directly
11370 used as premaster secret.
11372 *Douglas Stebila (Sun Microsystems Laboratories)*
11374 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11375 curve secp160r1 to the tests.
11377 *Douglas Stebila (Sun Microsystems Laboratories)*
11379 * Add the possibility to load symbols globally with DSO.
11381 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11383 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11384 control of the error stack.
11388 * Add support for STORE in ENGINE.
11392 * Add the STORE type. The intention is to provide a common interface
11393 to certificate and key stores, be they simple file-based stores, or
11394 HSM-type store, or LDAP stores, or...
11395 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11399 * Add a generic structure called OPENSSL_ITEM. This can be used to
11400 pass a list of arguments to any function as well as provide a way
11401 for a function to pass data back to the caller.
11405 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11406 works like BUF_strdup() but can be used to duplicate a portion of
11407 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11412 * Add the function sk_find_ex() which works like sk_find(), but will
11413 return an index to an element even if an exact match couldn't be
11414 found. The index is guaranteed to point at the element where the
11415 searched-for key would be inserted to preserve sorting order.
11419 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11420 takes an extra flags argument for optional functionality. Currently,
11421 the following flags are defined:
11423 OBJ_BSEARCH_VALUE_ON_NOMATCH
11424 This one gets OBJ_bsearch_ex() to return a pointer to the first
11425 element where the comparing function returns a negative or zero
11428 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11429 This one gets OBJ_bsearch_ex() to return a pointer to the first
11430 element where the comparing function returns zero. This is useful
11431 if there are more than one element where the comparing function
11436 * Make it possible to create self-signed certificates with 'openssl ca'
11437 in such a way that the self-signed certificate becomes part of the
11438 CA database and uses the same mechanisms for serial number generation
11439 as all other certificate signing. The new flag '-selfsign' enables
11440 this functionality. Adapt CA.sh and CA.pl.in.
11444 * Add functionality to check the public key of a certificate request
11445 against a given private. This is useful to check that a certificate
11446 request can be signed by that key (self-signing).
11450 * Make it possible to have multiple active certificates with the same
11451 subject in the CA index file. This is done only if the keyword
11452 'unique_subject' is set to 'no' in the main CA section (default
11453 if 'CA_default') of the configuration file. The value is saved
11454 with the database itself in a separate index attribute file,
11455 named like the index file with '.attr' appended to the name.
11459 * Generate multi-valued AVAs using '+' notation in config files for
11464 * Support for nameConstraints certificate extension.
11468 * Support for policyConstraints certificate extension.
11472 * Support for policyMappings certificate extension.
11476 * Make sure the default DSA_METHOD implementation only uses its
11477 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11478 and change its own handlers to be NULL so as to remove unnecessary
11479 indirection. This lets alternative implementations fallback to the
11480 default implementation more easily.
11484 * Support for directoryName in GeneralName related extensions
11489 * Make it possible to link applications using Makefile.shared.
11490 Make that possible even when linking against static libraries!
11494 * Support for single pass processing for S/MIME signing. This now
11495 means that S/MIME signing can be done from a pipe, in addition
11496 cleartext signing (multipart/signed type) is effectively streaming
11497 and the signed data does not need to be all held in memory.
11499 This is done with a new flag PKCS7_STREAM. When this flag is set
11500 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11501 is done after the data is output (and digests calculated) in
11502 SMIME_write_PKCS7().
11506 * Add full support for -rpath/-R, both in shared libraries and
11507 applications, at least on the platforms where it's known how
11512 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11513 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11514 will now compute a table of multiples of the generator that
11515 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11516 faster (notably in the case of a single point multiplication,
11517 scalar * generator).
11519 *Nils Larsch, Bodo Moeller*
11521 * IPv6 support for certificate extensions. The various extensions
11522 which use the IP:a.b.c.d can now take IPv6 addresses using the
11523 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11528 * Added an ENGINE that implements RSA by performing private key
11529 exponentiations with the GMP library. The conversions to and from
11530 GMP's mpz_t format aren't optimised nor are any montgomery forms
11531 cached, and on x86 it appears OpenSSL's own performance has caught up.
11532 However there are likely to be other architectures where GMP could
11533 provide a boost. This ENGINE is not built in by default, but it can be
11534 specified at Configure time and should be accompanied by the necessary
11535 linker additions, eg;
11536 ./config -DOPENSSL_USE_GMP -lgmp
11540 * "openssl engine" will not display ENGINE/DSO load failure errors when
11541 testing availability of engines with "-t" - the old behaviour is
11542 produced by increasing the feature's verbosity with "-tt".
11546 * ECDSA routines: under certain error conditions uninitialized BN objects
11547 could be freed. Solution: make sure initialization is performed early
11548 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11553 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11554 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11555 software implementations. For DSA and DH, parameter generation can
11556 also be overridden by providing the appropriate method callbacks.
11560 * Change the "progress" mechanism used in key-generation and
11561 primality testing to functions that take a new BN_GENCB pointer in
11562 place of callback/argument pairs. The new API functions have `_ex`
11563 postfixes and the older functions are reimplemented as wrappers for
11564 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11565 declarations of the old functions to help (graceful) attempts to
11566 migrate to the new functions. Also, the new key-generation API
11567 functions operate on a caller-supplied key-structure and return
11568 success/failure rather than returning a key or NULL - this is to
11569 help make "keygen" another member function of RSA_METHOD etc.
11571 Example for using the new callback interface:
11573 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11574 void *my_arg = ...;
11577 BN_GENCB_set(&my_cb, my_callback, my_arg);
11579 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11580 /* For the meaning of a, b in calls to my_callback(), see the
11581 * documentation of the function that calls the callback.
11582 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11583 * my_callback should return 1 if it wants BN_is_prime_ex()
11584 * to continue, or 0 to stop.
11589 * Change the ZLIB compression method to be stateful, and make it
11590 available to TLS with the number defined in
11591 draft-ietf-tls-compression-04.txt.
11595 * Add the ASN.1 structures and functions for CertificatePair, which
11596 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11598 CertificatePair ::= SEQUENCE {
11599 forward [0] Certificate OPTIONAL,
11600 reverse [1] Certificate OPTIONAL,
11601 -- at least one of the pair shall be present -- }
11603 Also implement the PEM functions to read and write certificate
11604 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11606 This needed to be defined, mostly for the sake of the LDAP
11607 attribute crossCertificatePair, but may prove useful elsewhere as
11612 * Make it possible to inhibit symlinking of shared libraries in
11613 Makefile.shared, for Cygwin's sake.
11617 * Extend the BIGNUM API by creating a function
11618 void BN_set_negative(BIGNUM *a, int neg);
11619 and a macro that behave like
11620 int BN_is_negative(const BIGNUM *a);
11622 to avoid the need to access 'a->neg' directly in applications.
11626 * Implement fast modular reduction for pseudo-Mersenne primes
11627 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11628 EC_GROUP_new_curve_GFp() will now automatically use this
11631 *Nils Larsch <nla@trustcenter.de>*
11633 * Add new lock type (CRYPTO_LOCK_BN).
11637 * Change the ENGINE framework to automatically load engines
11638 dynamically from specific directories unless they could be
11639 found to already be built in or loaded. Move all the
11640 current engines except for the cryptodev one to a new
11641 directory engines/.
11642 The engines in engines/ are built as shared libraries if
11643 the "shared" options was given to ./Configure or ./config.
11644 Otherwise, they are inserted in libcrypto.a.
11645 /usr/local/ssl/engines is the default directory for dynamic
11646 engines, but that can be overridden at configure time through
11647 the usual use of --prefix and/or --openssldir, and at run
11648 time with the environment variable OPENSSL_ENGINES.
11650 *Geoff Thorpe and Richard Levitte*
11652 * Add Makefile.shared, a helper makefile to build shared
11653 libraries. Adapt Makefile.org.
11657 * Add version info to Win32 DLLs.
11659 *Peter 'Luna' Runestig" <peter@runestig.com>*
11661 * Add new 'medium level' PKCS#12 API. Certificates and keys
11662 can be added using this API to created arbitrary PKCS#12
11663 files while avoiding the low-level API.
11665 New options to PKCS12_create(), key or cert can be NULL and
11666 will then be omitted from the output file. The encryption
11667 algorithm NIDs can be set to -1 for no encryption, the mac
11668 iteration count can be set to 0 to omit the mac.
11670 Enhance pkcs12 utility by making the -nokeys and -nocerts
11671 options work when creating a PKCS#12 file. New option -nomac
11672 to omit the mac, NONE can be set for an encryption algorithm.
11673 New code is modified to use the enhanced PKCS12_create()
11674 instead of the low-level API.
11678 * Extend ASN1 encoder to support indefinite length constructed
11679 encoding. This can output sequences tags and octet strings in
11680 this form. Modify pk7_asn1.c to support indefinite length
11681 encoding. This is experimental and needs additional code to
11682 be useful, such as an ASN1 bio and some enhanced streaming
11685 Extend template encode functionality so that tagging is passed
11686 down to the template encoder.
11690 * Let 'openssl req' fail if an argument to '-newkey' is not
11691 recognized instead of using RSA as a default.
11695 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11696 As these are not official, they are not included in "ALL";
11697 the "ECCdraft" ciphersuite group alias can be used to select them.
11699 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11701 * Add ECDH engine support.
11703 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11705 * Add ECDH in new directory crypto/ecdh/.
11707 *Douglas Stebila (Sun Microsystems Laboratories)*
11709 * Let BN_rand_range() abort with an error after 100 iterations
11710 without success (which indicates a broken PRNG).
11714 * Change BN_mod_sqrt() so that it verifies that the input value
11715 is really the square of the return value. (Previously,
11716 BN_mod_sqrt would show GIGO behaviour.)
11720 * Add named elliptic curves over binary fields from X9.62, SECG,
11721 and WAP/WTLS; add OIDs that were still missing.
11723 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11725 * Extend the EC library for elliptic curves over binary fields
11726 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11729 EC_GF2m_simple_method
11733 EC_GROUP_new_curve_GF2m
11734 EC_GROUP_set_curve_GF2m
11735 EC_GROUP_get_curve_GF2m
11736 EC_POINT_set_affine_coordinates_GF2m
11737 EC_POINT_get_affine_coordinates_GF2m
11738 EC_POINT_set_compressed_coordinates_GF2m
11740 Point compression for binary fields is disabled by default for
11741 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11744 As binary polynomials are represented as BIGNUMs, various members
11745 of the EC_GROUP and EC_POINT data structures can be shared
11746 between the implementations for prime fields and binary fields;
11747 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11748 are essentially identical to their `..._GFp` counterparts.
11749 (For simplicity, the `..._GFp` prefix has been dropped from
11750 various internal method names.)
11752 An internal 'field_div' method (similar to 'field_mul' and
11753 'field_sqr') has been added; this is used only for binary fields.
11755 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11757 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11758 through methods ('mul', 'precompute_mult').
11760 The generic implementations (now internally called 'ec_wNAF_mul'
11761 and 'ec_wNAF_precomputed_mult') remain the default if these
11762 methods are undefined.
11764 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11766 * New function EC_GROUP_get_degree, which is defined through
11767 EC_METHOD. For curves over prime fields, this returns the bit
11768 length of the modulus.
11770 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11772 * New functions EC_GROUP_dup, EC_POINT_dup.
11773 (These simply call ..._new and ..._copy).
11775 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11777 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11778 Polynomials are represented as BIGNUMs (where the sign bit is not
11779 used) in the following functions [macros]:
11782 BN_GF2m_sub [= BN_GF2m_add]
11783 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11784 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11785 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11787 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11788 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11789 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11790 BN_GF2m_cmp [= BN_ucmp]
11792 (Note that only the 'mod' functions are actually for fields GF(2^m).
11793 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11795 For some functions, an the irreducible polynomial defining a
11796 field can be given as an 'unsigned int[]' with strictly
11797 decreasing elements giving the indices of those bits that are set;
11798 i.e., p[] represents the polynomial
11799 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11801 p[0] > p[1] > ... > p[k] = 0.
11802 This applies to the following functions:
11805 BN_GF2m_mod_mul_arr
11806 BN_GF2m_mod_sqr_arr
11807 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11808 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11809 BN_GF2m_mod_exp_arr
11810 BN_GF2m_mod_sqrt_arr
11811 BN_GF2m_mod_solve_quad_arr
11815 Conversion can be performed by the following functions:
11820 bntest.c has additional tests for binary polynomial arithmetic.
11822 Two implementations for BN_GF2m_mod_div() are available.
11823 The default algorithm simply uses BN_GF2m_mod_inv() and
11824 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11825 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11826 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11828 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11830 * Add new error code 'ERR_R_DISABLED' that can be used when some
11831 functionality is disabled at compile-time.
11833 *Douglas Stebila <douglas.stebila@sun.com>*
11835 * Change default behaviour of 'openssl asn1parse' so that more
11836 information is visible when viewing, e.g., a certificate:
11838 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11839 mode the content of non-printable OCTET STRINGs is output in a
11840 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11841 avoid the appearance of a printable string.
11843 *Nils Larsch <nla@trustcenter.de>*
11845 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11847 EC_GROUP_set_asn1_flag()
11848 EC_GROUP_get_asn1_flag()
11849 EC_GROUP_set_point_conversion_form()
11850 EC_GROUP_get_point_conversion_form()
11851 These control ASN1 encoding details:
11852 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11853 has been set to OPENSSL_EC_NAMED_CURVE.
11854 - Points are encoded in uncompressed form by default; options for
11855 asn1_for are as for point2oct, namely
11856 POINT_CONVERSION_COMPRESSED
11857 POINT_CONVERSION_UNCOMPRESSED
11858 POINT_CONVERSION_HYBRID
11860 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11862 EC_GROUP_set_seed()
11863 EC_GROUP_get0_seed()
11864 EC_GROUP_get_seed_len()
11865 This is used only for ASN1 purposes (so far).
11867 *Nils Larsch <nla@trustcenter.de>*
11869 * Add 'field_type' member to EC_METHOD, which holds the NID
11870 of the appropriate field type OID. The new function
11871 EC_METHOD_get_field_type() returns this value.
11873 *Nils Larsch <nla@trustcenter.de>*
11876 EC_POINT_point2bn()
11877 EC_POINT_bn2point()
11878 EC_POINT_point2hex()
11879 EC_POINT_hex2point()
11880 providing useful interfaces to EC_POINT_point2oct() and
11881 EC_POINT_oct2point().
11883 *Nils Larsch <nla@trustcenter.de>*
11885 * Change internals of the EC library so that the functions
11886 EC_GROUP_set_generator()
11887 EC_GROUP_get_generator()
11888 EC_GROUP_get_order()
11889 EC_GROUP_get_cofactor()
11890 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11891 to methods, which would lead to unnecessary code duplication when
11892 adding different types of curves.
11894 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11896 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11897 arithmetic, and such that modified wNAFs are generated
11898 (which avoid length expansion in many cases).
11902 * Add a function EC_GROUP_check_discriminant() (defined via
11903 EC_METHOD) that verifies that the curve discriminant is non-zero.
11905 Add a function EC_GROUP_check() that makes some sanity tests
11906 on a EC_GROUP, its generator and order. This includes
11907 EC_GROUP_check_discriminant().
11909 *Nils Larsch <nla@trustcenter.de>*
11911 * Add ECDSA in new directory crypto/ecdsa/.
11913 Add applications 'openssl ecparam' and 'openssl ecdsa'
11914 (these are based on 'openssl dsaparam' and 'openssl dsa').
11916 ECDSA support is also included in various other files across the
11917 library. Most notably,
11918 - 'openssl req' now has a '-newkey ecdsa:file' option;
11919 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11920 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11921 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11922 them suitable for ECDSA where domain parameters must be
11923 extracted before the specific public key;
11924 - ECDSA engine support has been added.
11926 *Nils Larsch <nla@trustcenter.de>*
11928 * Include some named elliptic curves, and add OIDs from X9.62,
11929 SECG, and WAP/WTLS. Each curve can be obtained from the new
11931 EC_GROUP_new_by_curve_name(),
11932 and the list of available named curves can be obtained with
11933 EC_get_builtin_curves().
11934 Also add a 'curve_name' member to EC_GROUP objects, which can be
11936 EC_GROUP_set_curve_name()
11937 EC_GROUP_get_curve_name()
11939 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11941 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11942 was actually never needed) and in BN_mul(). The removal in BN_mul()
11943 required a small change in bn_mul_part_recursive() and the addition
11944 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11945 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11946 bn_sub_words() and bn_add_words() except they take arrays with
11951 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11953 * Cleanse PEM buffers before freeing them since they may contain
11956 *Benjamin Bennett <ben@psc.edu>*
11958 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11959 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11960 authentication-only ciphersuites.
11964 * Since AES128 and AES256 share a single mask bit in the logic of
11965 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11966 kludge to work properly if AES128 is available and AES256 isn't.
11970 * Expand security boundary to match 1.1.1 module.
11974 * Remove redundant features: hash file source, editing of test vectors
11975 modify fipsld to use external fips_premain.c signature.
11979 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11980 run algorithm test programs.
11984 * Make algorithm test programs more tolerant of whitespace.
11988 * Have SSL/TLS server implementation tolerate "mismatched" record
11989 protocol version while receiving ClientHello even if the
11990 ClientHello is fragmented. (The server can't insist on the
11991 particular protocol version it has chosen before the ServerHello
11992 message has informed the client about his choice.)
11996 * Load error codes if they are not already present instead of using a
11997 static variable. This allows them to be cleanly unloaded and reloaded.
12001 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12003 * Introduce limits to prevent malicious keys being able to
12004 cause a denial of service. ([CVE-2006-2940])
12006 *Steve Henson, Bodo Moeller*
12008 * Fix ASN.1 parsing of certain invalid structures that can result
12009 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12011 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12012 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12014 * Fix SSL client code which could crash if connecting to a
12015 malicious SSLv2 server. ([CVE-2006-4343])
12017 *Tavis Ormandy and Will Drewry, Google Security Team*
12019 * Change ciphersuite string processing so that an explicit
12020 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12021 will no longer include "AES128-SHA"), and any other similar
12022 ciphersuite (same bitmap) from *other* protocol versions (so that
12023 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12024 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12025 changes from 0.9.8b and 0.9.8d.
12029 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12031 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12032 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12034 * Change the Unix randomness entropy gathering to use poll() when
12035 possible instead of select(), since the latter has some
12036 undesirable limitations.
12038 *Darryl Miles via Richard Levitte and Bodo Moeller*
12040 * Disable rogue ciphersuites:
12042 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12043 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12044 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12046 The latter two were purportedly from
12047 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12050 Also deactivate the remaining ciphersuites from
12051 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12052 unofficial, and the ID has long expired.
12056 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12057 dual-core machines) and other potential thread-safety issues.
12061 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12063 * Adapt fipsld and the build system to link against the validated FIPS
12064 module in FIPS mode.
12068 * Fixes for VC++ 2005 build under Windows.
12072 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12073 from a Windows bash shell such as MSYS. It is autodetected from the
12074 "config" script when run from a VC++ environment. Modify standard VC++
12075 build to use fipscanister.o from the GNU make build.
12079 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12081 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12082 The value now differs depending on if you build for FIPS or not.
12083 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12084 safely run with a non-FIPSed libcrypto, as it may crash because of
12085 the difference induced by this change.
12089 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12091 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12092 (part of SSL_OP_ALL). This option used to disable the
12093 countermeasure against man-in-the-middle protocol-version
12094 rollback in the SSL 2.0 server implementation, which is a bad
12095 idea. ([CVE-2005-2969])
12097 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12098 for Information Security, National Institute of Advanced Industrial
12099 Science and Technology [AIST, Japan)]*
12101 * Minimal support for X9.31 signatures and PSS padding modes. This is
12102 mainly for FIPS compliance and not fully integrated at this stage.
12106 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12107 the exponentiation using a fixed-length exponent. (Otherwise,
12108 the information leaked through timing could expose the secret key
12109 after many signatures; cf. Bleichenbacher's attack on DSA with
12114 * Make a new fixed-window mod_exp implementation the default for
12115 RSA, DSA, and DH private-key operations so that the sequence of
12116 squares and multiplies and the memory access pattern are
12117 independent of the particular secret key. This will mitigate
12118 cache-timing and potential related attacks.
12120 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12121 and this is automatically used by BN_mod_exp_mont() if the new flag
12122 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12123 will use this BN flag for private exponents unless the flag
12124 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12125 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12127 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12129 * Change the client implementation for SSLv23_method() and
12130 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12131 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12132 (Previously, the SSL 2.0 backwards compatible Client Hello
12133 message format would be used even with SSL_OP_NO_SSLv2.)
12137 * Add support for smime-type MIME parameter in S/MIME messages which some
12142 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12143 a threadsafe manner. Modify rsa code to use new function and add calls
12144 to dsa and dh code (which had race conditions before).
12148 * Include the fixed error library code in the C error file definitions
12149 instead of fixing them up at runtime. This keeps the error code
12150 structures constant.
12154 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12156 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12159 * Fixes for newer kerberos headers. NB: the casts are needed because
12160 the 'length' field is signed on one version and unsigned on another
12161 with no (?) obvious way to tell the difference, without these VC++
12162 complains. Also the "definition" of FAR (blank) is no longer included
12163 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12164 some needed definitions.
12168 * Undo Cygwin change.
12172 * Added support for proxy certificates according to RFC 3820.
12173 Because they may be a security thread to unaware applications,
12174 they must be explicitly allowed in run-time. See
12175 docs/HOWTO/proxy_certificates.txt for further information.
12179 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12181 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12182 server and client random values. Previously
12183 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12184 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12186 This change has negligible security impact because:
12188 1. Server and client random values still have 24 bytes of pseudo random
12191 2. Server and client random values are sent in the clear in the initial
12194 3. The master secret is derived using the premaster secret (48 bytes in
12195 size for static RSA ciphersuites) as well as client server and random
12198 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12201 *Stephen Henson, reported by UK NISCC*
12203 * Use Windows randomness collection on Cygwin.
12207 * Fix hang in EGD/PRNGD query when communication socket is closed
12208 prematurely by EGD/PRNGD.
12210 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12212 * Prompt for pass phrases when appropriate for PKCS12 input format.
12216 * Back-port of selected performance improvements from development
12217 branch, as well as improved support for PowerPC platforms.
12221 * Add lots of checks for memory allocation failure, error codes to indicate
12222 failure and freeing up memory if a failure occurs.
12224 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12226 * Add new -passin argument to dgst.
12230 * Perform some character comparisons of different types in X509_NAME_cmp:
12231 this is needed for some certificates that re-encode DNs into UTF8Strings
12232 (in violation of RFC3280) and can't or won't issue name rollover
12237 * Make an explicit check during certificate validation to see that
12238 the CA setting in each certificate on the chain is correct. As a
12239 side effect always do the following basic checks on extensions,
12240 not just when there's an associated purpose to the check:
12242 - if there is an unhandled critical extension (unless the user
12243 has chosen to ignore this fault)
12244 - if the path length has been exceeded (if one is set at all)
12245 - that certain extensions fit the associated purpose (if one has
12250 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12252 * Avoid a race condition when CRLs are checked in a multi threaded
12253 environment. This would happen due to the reordering of the revoked
12254 entries during signature checking and serial number lookup. Now the
12255 encoding is cached and the serial number sort performed under a lock.
12256 Add new STACK function sk_is_sorted().
12260 * Add Delta CRL to the extension code.
12264 * Various fixes to s3_pkt.c so alerts are sent properly.
12266 *David Holmes <d.holmes@f5.com>*
12268 * Reduce the chances of duplicate issuer name and serial numbers (in
12269 violation of RFC3280) using the OpenSSL certificate creation utilities.
12270 This is done by creating a random 64 bit value for the initial serial
12271 number when a serial number file is created or when a self signed
12272 certificate is created using 'openssl req -x509'. The initial serial
12273 number file is created using 'openssl x509 -next_serial' in CA.pl
12274 rather than being initialized to 1.
12278 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12280 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12281 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12283 *Joe Orton, Steve Henson*
12285 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12288 *Joe Orton, Steve Henson*
12290 * Make it possible to have multiple active certificates with the same
12291 subject in the CA index file. This is done only if the keyword
12292 'unique_subject' is set to 'no' in the main CA section (default
12293 if 'CA_default') of the configuration file. The value is saved
12294 with the database itself in a separate index attribute file,
12295 named like the index file with '.attr' appended to the name.
12299 * X509 verify fixes. Disable broken certificate workarounds when
12300 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12301 keyUsage extension present. Don't accept CRLs with unhandled critical
12302 extensions: since verify currently doesn't process CRL extensions this
12303 rejects a CRL with *any* critical extensions. Add new verify error codes
12308 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12309 A clarification of RFC2560 will require the use of OCTET STRINGs and
12310 some implementations cannot handle the current raw format. Since OpenSSL
12311 copies and compares OCSP nonces as opaque blobs without any attempt at
12312 parsing them this should not create any compatibility issues.
12316 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12317 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12318 this HMAC (and other) operations are several times slower than OpenSSL
12323 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12325 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12327 * Use the correct content when signing type "other".
12331 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12333 * Fix various bugs revealed by running the NISCC test suite:
12335 Stop out of bounds reads in the ASN1 code when presented with
12336 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12338 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12340 If verify callback ignores invalid public key errors don't try to check
12341 certificate signature with the NULL public key.
12345 * New -ignore_err option in ocsp application to stop the server
12346 exiting on the first error in a request.
12350 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12351 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12356 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12357 extra data after the compression methods not only for TLS 1.0
12358 but also for SSL 3.0 (as required by the specification).
12360 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12362 * Change X509_certificate_type() to mark the key as exported/exportable
12363 when it's 512 *bits* long, not 512 bytes.
12367 * Change AES_cbc_encrypt() so it outputs exact multiple of
12368 blocks during encryption.
12372 * Various fixes to base64 BIO and non blocking I/O. On write
12373 flushes were not handled properly if the BIO retried. On read
12374 data was not being buffered properly and had various logic bugs.
12375 This also affects blocking I/O when the data being decoded is a
12380 * Various S/MIME bugfixes and compatibility changes:
12381 output correct application/pkcs7 MIME type if
12382 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12383 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12384 of files as .eml work). Correctly handle very long lines in MIME
12389 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12391 * Countermeasure against the Klima-Pokorny-Rosa extension of
12392 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12393 a protocol version number mismatch like a decryption error
12394 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12398 * Turn on RSA blinding by default in the default implementation
12399 to avoid a timing attack. Applications that don't want it can call
12400 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12401 They would be ill-advised to do so in most cases.
12403 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12405 * Change RSA blinding code so that it works when the PRNG is not
12406 seeded (in this case, the secret RSA exponent is abused as
12407 an unpredictable seed -- if it is not unpredictable, there
12408 is no point in blinding anyway). Make RSA blinding thread-safe
12409 by remembering the creator's thread ID in rsa->blinding and
12410 having all other threads use local one-time blinding factors
12411 (this requires more computation than sharing rsa->blinding, but
12412 avoids excessive locking; and if an RSA object is not shared
12413 between threads, blinding will still be very fast).
12417 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12418 ENGINE as defaults for all supported algorithms irrespective of
12419 the 'flags' parameter. 'flags' is now honoured, so applications
12420 should make sure they are passing it correctly.
12424 * Target "mingw" now allows native Windows code to be generated in
12425 the Cygwin environment as well as with the MinGW compiler.
12429 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12431 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12432 via timing by performing a MAC computation even if incorrect
12433 block cipher padding has been found. This is a countermeasure
12434 against active attacks where the attacker has to distinguish
12435 between bad padding and a MAC verification error. ([CVE-2003-0078])
12437 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12438 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12439 Martin Vuagnoux (EPFL, Ilion)*
12441 * Make the no-err option work as intended. The intention with no-err
12442 is not to have the whole error stack handling routines removed from
12443 libcrypto, it's only intended to remove all the function name and
12444 reason texts, thereby removing some of the footprint that may not
12445 be interesting if those errors aren't displayed anyway.
12447 NOTE: it's still possible for any application or module to have its
12448 own set of error texts inserted. The routines are there, just not
12449 used by default when no-err is given.
12453 * Add support for FreeBSD on IA64.
12455 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12457 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12458 Kerberos function mit_des_cbc_cksum(). Before this change,
12459 the value returned by DES_cbc_cksum() was like the one from
12460 mit_des_cbc_cksum(), except the bytes were swapped.
12462 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12464 * Allow an application to disable the automatic SSL chain building.
12465 Before this a rather primitive chain build was always performed in
12466 ssl3_output_cert_chain(): an application had no way to send the
12467 correct chain if the automatic operation produced an incorrect result.
12469 Now the chain builder is disabled if either:
12471 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12473 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12475 The reasoning behind this is that an application would not want the
12476 auto chain building to take place if extra chain certificates are
12477 present and it might also want a means of sending no additional
12478 certificates (for example the chain has two certificates and the
12483 * Add the possibility to build without the ENGINE framework.
12485 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12487 * Under Win32 gmtime() can return NULL: check return value in
12488 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12492 * DSA routines: under certain error conditions uninitialized BN objects
12493 could be freed. Solution: make sure initialization is performed early
12494 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12495 Nils Larsch <nla@trustcenter.de> via PR#459)
12499 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12500 checked on reconnect on the client side, therefore session resumption
12501 could still fail with a "ssl session id is different" error. This
12502 behaviour is masked when SSL_OP_ALL is used due to
12503 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12504 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12505 followup to PR #377.
12509 * IA-32 assembler support enhancements: unified ELF targets, support
12510 for SCO/Caldera platforms, fix for Cygwin shared build.
12514 * Add support for FreeBSD on sparc64. As a consequence, support for
12515 FreeBSD on non-x86 processors is separate from x86 processors on
12516 the config script, much like the NetBSD support.
12518 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12520 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12522 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12525 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12526 code (06) was taken as the first octet of the session ID and the last
12527 octet was ignored consequently. As a result SSLv2 client side session
12528 caching could not have worked due to the session ID mismatch between
12530 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12535 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12536 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12541 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12542 seems that in spite of existing for more than a year, many application
12543 author have done nothing to provide the necessary callbacks, which
12544 means that this particular engine will not work properly anywhere.
12545 This is a very unfortunate situation which forces us, in the name
12546 of usability, to give the hw_ncipher.c a static lock, which is part
12548 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12549 appear in 0.9.8 or later. We EXPECT application authors to have
12550 dealt properly with this when 0.9.8 is released (unless we actually
12551 make such changes in the libcrypto locking code that changes will
12552 have to be made anyway).
12556 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12557 octets have been read, EOF or an error occurs. Without this change
12558 some truncated ASN1 structures will not produce an error.
12562 * Disable Heimdal support, since it hasn't been fully implemented.
12563 Still give the possibility to force the use of Heimdal, but with
12564 warnings and a request that patches get sent to openssl-dev.
12568 * Add the VC-CE target, introduce the WINCE sysname, and add
12569 INSTALL.WCE and appropriate conditionals to make it build.
12571 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12573 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12574 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12575 edit numbers of the version.
12577 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12579 * Introduce safe string copy and catenation functions
12580 (BUF_strlcpy() and BUF_strlcat()).
12582 *Ben Laurie (CHATS) and Richard Levitte*
12584 * Avoid using fixed-size buffers for one-line DNs.
12586 *Ben Laurie (CHATS)*
12588 * Add BUF_MEM_grow_clean() to avoid information leakage when
12589 resizing buffers containing secrets, and use where appropriate.
12591 *Ben Laurie (CHATS)*
12593 * Avoid using fixed size buffers for configuration file location.
12595 *Ben Laurie (CHATS)*
12597 * Avoid filename truncation for various CA files.
12599 *Ben Laurie (CHATS)*
12601 * Use sizeof in preference to magic numbers.
12603 *Ben Laurie (CHATS)*
12605 * Avoid filename truncation in cert requests.
12607 *Ben Laurie (CHATS)*
12609 * Add assertions to check for (supposedly impossible) buffer
12612 *Ben Laurie (CHATS)*
12614 * Don't cache truncated DNS entries in the local cache (this could
12615 potentially lead to a spoofing attack).
12617 *Ben Laurie (CHATS)*
12619 * Fix various buffers to be large enough for hex/decimal
12620 representations in a platform independent manner.
12622 *Ben Laurie (CHATS)*
12624 * Add CRYPTO_realloc_clean() to avoid information leakage when
12625 resizing buffers containing secrets, and use where appropriate.
12627 *Ben Laurie (CHATS)*
12629 * Add BIO_indent() to avoid much slightly worrying code to do
12632 *Ben Laurie (CHATS)*
12634 * Convert sprintf()/BIO_puts() to BIO_printf().
12636 *Ben Laurie (CHATS)*
12638 * buffer_gets() could terminate with the buffer only half
12641 *Ben Laurie (CHATS)*
12643 * Add assertions to prevent user-supplied crypto functions from
12644 overflowing internal buffers by having large block sizes, etc.
12646 *Ben Laurie (CHATS)*
12648 * New OPENSSL_assert() macro (similar to assert(), but enabled
12651 *Ben Laurie (CHATS)*
12653 * Eliminate unused copy of key in RC4.
12655 *Ben Laurie (CHATS)*
12657 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12659 *Ben Laurie (CHATS)*
12661 * Fix off-by-one error in EGD path.
12663 *Ben Laurie (CHATS)*
12665 * If RANDFILE path is too long, ignore instead of truncating.
12667 *Ben Laurie (CHATS)*
12669 * Eliminate unused and incorrectly sized X.509 structure
12672 *Ben Laurie (CHATS)*
12674 * Eliminate unused and dangerous function knumber().
12676 *Ben Laurie (CHATS)*
12678 * Eliminate unused and dangerous structure, KSSL_ERR.
12680 *Ben Laurie (CHATS)*
12682 * Protect against overlong session ID context length in an encoded
12683 session object. Since these are local, this does not appear to be
12686 *Ben Laurie (CHATS)*
12688 * Change from security patch (see 0.9.6e below) that did not affect
12689 the 0.9.6 release series:
12691 Remote buffer overflow in SSL3 protocol - an attacker could
12692 supply an oversized master key in Kerberos-enabled versions.
12695 *Ben Laurie (CHATS)*
12697 * Change the SSL kerb5 codes to match RFC 2712.
12701 * Make -nameopt work fully for req and add -reqopt switch.
12703 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12705 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12707 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12709 * Make sure tests can be performed even if the corresponding algorithms
12710 have been removed entirely. This was also the last step to make
12711 OpenSSL compilable with DJGPP under all reasonable conditions.
12713 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12715 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12716 to allow version independent disabling of normally unselected ciphers,
12717 which may be activated as a side-effect of selecting a single cipher.
12719 (E.g., cipher list string "RSA" enables ciphersuites that are left
12720 out of "ALL" because they do not provide symmetric encryption.
12721 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12723 *Lutz Jaenicke, Bodo Moeller*
12725 * Add appropriate support for separate platform-dependent build
12726 directories. The recommended way to make a platform-dependent
12727 build directory is the following (tested on Linux), maybe with
12730 # Place yourself outside of the OpenSSL source tree. In
12731 # this example, the environment variable OPENSSL_SOURCE
12732 # is assumed to contain the absolute OpenSSL source directory.
12733 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12734 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12735 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12736 mkdir -p `dirname $F`
12737 ln -s $OPENSSL_SOURCE/$F $F
12740 To be absolutely sure not to disturb the source tree, a "make clean"
12741 is a good thing. If it isn't successful, don't worry about it,
12742 it probably means the source directory is very clean.
12746 * Make sure any ENGINE control commands make local copies of string
12747 pointers passed to them whenever necessary. Otherwise it is possible
12748 the caller may have overwritten (or deallocated) the original string
12749 data when a later ENGINE operation tries to use the stored values.
12751 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12753 * Improve diagnostics in file reading and command-line digests.
12755 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12757 * Add AES modes CFB and OFB to the object database. Correct an
12758 error in AES-CFB decryption.
12762 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12763 allows existing EVP_CIPHER_CTX structures to be reused after
12764 calling `EVP_*Final()`. This behaviour is used by encryption
12765 BIOs and some applications. This has the side effect that
12766 applications must explicitly clean up cipher contexts with
12767 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12771 * Check the values of dna and dnb in bn_mul_recursive before calling
12772 bn_mul_comba (a non zero value means the a or b arrays do not contain
12773 n2 elements) and fallback to bn_mul_normal if either is not zero.
12777 * Fix escaping of non-ASCII characters when using the -subj option
12778 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12782 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12783 form for "surname", serialNumber has no short form.
12784 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12785 therefore remove "mail" short name for "internet 7".
12786 The OID for unique identifiers in X509 certificates is
12787 x500UniqueIdentifier, not uniqueIdentifier.
12788 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12792 * Add an "init" command to the ENGINE config module and auto initialize
12793 ENGINEs. Without any "init" command the ENGINE will be initialized
12794 after all ctrl commands have been executed on it. If init=1 the
12795 ENGINE is initialized at that point (ctrls before that point are run
12796 on the uninitialized ENGINE and after on the initialized one). If
12797 init=0 then the ENGINE will not be initialized at all.
12801 * Fix the 'app_verify_callback' interface so that the user-defined
12802 argument is actually passed to the callback: In the
12803 SSL_CTX_set_cert_verify_callback() prototype, the callback
12804 declaration has been changed from
12807 int (*cb)(X509_STORE_CTX *,void *);
12808 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12809 i=s->ctx->app_verify_callback(&ctx)
12810 has been changed into
12811 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12813 To update applications using SSL_CTX_set_cert_verify_callback(),
12814 a dummy argument can be added to their callback functions.
12816 *D. K. Smetters <smetters@parc.xerox.com>*
12818 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12820 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12822 * Add and OPENSSL_LOAD_CONF define which will cause
12823 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12824 This allows older applications to transparently support certain
12825 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12826 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12827 load the config file and OPENSSL_add_all_algorithms_conf() which will
12828 always load it have also been added.
12832 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12833 Adjust NIDs and EVP layer.
12835 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12837 * Config modules support in openssl utility.
12839 Most commands now load modules from the config file,
12840 though in a few (such as version) this isn't done
12841 because it couldn't be used for anything.
12843 In the case of ca and req the config file used is
12844 the same as the utility itself: that is the -config
12845 command line option can be used to specify an
12850 * Move default behaviour from OPENSSL_config(). If appname is NULL
12851 use "openssl_conf" if filename is NULL use default openssl config file.
12855 * Add an argument to OPENSSL_config() to allow the use of an alternative
12856 config section name. Add a new flag to tolerate a missing config file
12857 and move code to CONF_modules_load_file().
12861 * Support for crypto accelerator cards from Accelerated Encryption
12862 Processing, www.aep.ie. (Use engine 'aep')
12863 The support was copied from 0.9.6c [engine] and adapted/corrected
12864 to work with the new engine framework.
12866 *AEP Inc. and Richard Levitte*
12868 * Support for SureWare crypto accelerator cards from Baltimore
12869 Technologies. (Use engine 'sureware')
12870 The support was copied from 0.9.6c [engine] and adapted
12871 to work with the new engine framework.
12875 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12876 make the newer ENGINE framework commands for the CHIL engine work.
12878 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12880 * Make it possible to produce shared libraries on ReliantUNIX.
12882 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12884 * Add the configuration target debug-linux-ppro.
12885 Make 'openssl rsa' use the general key loading routines
12886 implemented in `apps.c`, and make those routines able to
12887 handle the key format FORMAT_NETSCAPE and the variant
12890 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12892 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12894 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12896 * Add -keyform to rsautl, and document -engine.
12898 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12900 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12901 BIO_R_NO_SUCH_FILE error code rather than the generic
12902 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12906 * Add new functions
12907 ERR_peek_last_error
12908 ERR_peek_last_error_line
12909 ERR_peek_last_error_line_data.
12910 These are similar to
12912 ERR_peek_error_line
12913 ERR_peek_error_line_data,
12914 but report on the latest error recorded rather than the first one
12915 still in the error queue.
12917 *Ben Laurie, Bodo Moeller*
12919 * default_algorithms option in ENGINE config module. This allows things
12921 default_algorithms = ALL
12922 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12926 * Preliminary ENGINE config module.
12930 * New experimental application configuration code.
12934 * Change the AES code to follow the same name structure as all other
12935 symmetric ciphers, and behave the same way. Move everything to
12936 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12938 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12940 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12942 *Ben Laurie and Theo de Raadt*
12944 * Add option to output public keys in req command.
12946 *Massimiliano Pala madwolf@openca.org*
12948 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12949 (up to about 10% better than before for P-192 and P-224).
12953 * New functions/macros
12955 SSL_CTX_set_msg_callback(ctx, cb)
12956 SSL_CTX_set_msg_callback_arg(ctx, arg)
12957 SSL_set_msg_callback(ssl, cb)
12958 SSL_set_msg_callback_arg(ssl, arg)
12960 to request calling a callback function
12962 void cb(int write_p, int version, int content_type,
12963 const void *buf, size_t len, SSL *ssl, void *arg)
12965 whenever a protocol message has been completely received
12966 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12967 protocol version according to which the SSL library interprets
12968 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12969 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12970 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12971 specification (change_cipher_spec(20), alert(21), handshake(22)).
12972 'buf' and 'len' point to the actual message, 'ssl' to the
12973 SSL object, and 'arg' is the application-defined value set by
12974 SSL[_CTX]_set_msg_callback_arg().
12976 'openssl s_client' and 'openssl s_server' have new '-msg' options
12977 to enable a callback that displays all protocol messages.
12981 * Change the shared library support so shared libraries are built as
12982 soon as the corresponding static library is finished, and thereby get
12983 openssl and the test programs linked against the shared library.
12984 This still only happens when the keyword "shard" has been given to
12985 the configuration scripts.
12987 NOTE: shared library support is still an experimental thing, and
12988 backward binary compatibility is still not guaranteed.
12990 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12992 * Add support for Subject Information Access extension.
12994 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12996 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12997 additional bytes when new memory had to be allocated, not just
12998 when reusing an existing buffer.
13002 * New command line and configuration option 'utf8' for the req command.
13003 This allows field values to be specified as UTF8 strings.
13007 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13008 runs for the former and machine-readable output for the latter.
13012 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13013 of the e-mail address in the DN (i.e., it will go into a certificate
13014 extension only). The new configuration file option 'email_in_dn = no'
13015 has the same effect.
13017 *Massimiliano Pala madwolf@openca.org*
13019 * Change all functions with names starting with `des_` to be starting
13020 with `DES_` instead. Add wrappers that are compatible with libdes,
13021 but are named `_ossl_old_des_*`. Finally, add macros that map the
13022 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13023 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13024 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13027 Since we provide two compatibility mappings, the user needs to
13028 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13029 compatibility is desired. The default (i.e., when that macro
13030 isn't defined) is OpenSSL 0.9.6c compatibility.
13032 There are also macros that enable and disable the support of old
13033 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13034 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13035 are defined, the default will apply: to support the old des routines.
13037 In either case, one must include openssl/des.h to get the correct
13038 definitions. Do not try to just include openssl/des_old.h, that
13041 NOTE: This is a major break of an old API into a new one. Software
13042 authors are encouraged to switch to the `DES_` style functions. Some
13043 time in the future, des_old.h and the libdes compatibility functions
13044 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13045 default), and then completely removed.
13049 * Test for certificates which contain unsupported critical extensions.
13050 If such a certificate is found during a verify operation it is
13051 rejected by default: this behaviour can be overridden by either
13052 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13053 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13054 X509_supported_extension() has also been added which returns 1 if a
13055 particular extension is supported.
13059 * Modify the behaviour of EVP cipher functions in similar way to digests
13060 to retain compatibility with existing code.
13064 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13065 compatibility with existing code. In particular the 'ctx' parameter does
13066 not have to be to be initialized before the call to EVP_DigestInit() and
13067 it is tidied up after a call to EVP_DigestFinal(). New function
13068 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13069 EVP_MD_CTX_copy() changed to not require the destination to be
13070 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13071 requires the destination to be valid.
13073 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13074 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13078 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13079 so that complete 'Handshake' protocol structures are kept in memory
13080 instead of overwriting 'msg_type' and 'length' with 'body' data.
13084 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13086 *Massimo Santin via Richard Levitte*
13088 * Major restructuring to the underlying ENGINE code. This includes
13089 reduction of linker bloat, separation of pure "ENGINE" manipulation
13090 (initialisation, etc) from functionality dealing with implementations
13091 of specific crypto interfaces. This change also introduces integrated
13092 support for symmetric ciphers and digest implementations - so ENGINEs
13093 can now accelerate these by providing EVP_CIPHER and EVP_MD
13094 implementations of their own. This is detailed in
13095 [crypto/engine/README.md](crypto/engine/README.md)
13096 as it couldn't be adequately described here. However, there are a few
13097 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13098 were changed in the original introduction of ENGINE code have now
13099 reverted back - the hooking from this code to ENGINE is now a good
13100 deal more passive and at run-time, operations deal directly with
13101 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13102 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13103 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13104 they were not being used by the framework as there is no concept of a
13105 BIGNUM_METHOD and they could not be generalised to the new
13106 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13107 ENGINE_cpy() has been removed as it cannot be consistently defined in
13112 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13116 * Change mkdef.pl to sort symbols that get the same entry number,
13117 and make sure the automatically generated functions `ERR_load_*`
13118 become part of libeay.num as well.
13122 * New function SSL_renegotiate_pending(). This returns true once
13123 renegotiation has been requested (either SSL_renegotiate() call
13124 or HelloRequest/ClientHello received from the peer) and becomes
13125 false once a handshake has been completed.
13126 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13127 sends a HelloRequest, but does not ensure that a handshake takes
13128 place. SSL_renegotiate_pending() is useful for checking if the
13129 client has followed the request.)
13133 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13134 By default, clients may request session resumption even during
13135 renegotiation (if session ID contexts permit); with this option,
13136 session resumption is possible only in the first handshake.
13138 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13139 more bits available for options that should not be part of
13140 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13144 * Add some demos for certificate and certificate request creation.
13148 * Make maximum certificate chain size accepted from the peer application
13149 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13150 "Douglas E. Engert" <deengert@anl.gov>.
13154 * Add support for shared libraries for Unixware-7
13155 (Boyd Lynn Gerber <gerberb@zenez.com>).
13159 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13160 be done prior to destruction. Use this to unload error strings from
13161 ENGINEs that load their own error strings. NB: This adds two new API
13162 functions to "get" and "set" this destroy handler in an ENGINE.
13166 * Alter all existing ENGINE implementations (except "openssl" and
13167 "openbsd") to dynamically instantiate their own error strings. This
13168 makes them more flexible to be built both as statically-linked ENGINEs
13169 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13170 Also, add stub code to each that makes building them as self-contained
13171 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13175 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13176 implementations into applications that are completely implemented in
13177 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13178 commands that can be used to configure what shared-library to load and
13179 to control aspects of the way it is handled. Also, made an update to
13180 the [README-Engine.md](README-Engine.md) file
13181 that brings its information up-to-date and
13182 provides some information and instructions on the "dynamic" ENGINE
13183 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13187 * Make it possible to unload ranges of ERR strings with a new
13188 "ERR_unload_strings" function.
13192 * Add a copy() function to EVP_MD.
13196 * Make EVP_MD routines take a context pointer instead of just the
13197 md_data void pointer.
13201 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13202 that the digest can only process a single chunk of data
13203 (typically because it is provided by a piece of
13204 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13205 is only going to provide a single chunk of data, and hence the
13206 framework needn't accumulate the data for oneshot drivers.
13210 * As with "ERR", make it possible to replace the underlying "ex_data"
13211 functions. This change also alters the storage and management of global
13212 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13213 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13214 index counters. The API functions that use this state have been changed
13215 to take a "class_index" rather than pointers to the class's local STACK
13216 and counter, and there is now an API function to dynamically create new
13217 classes. This centralisation allows us to (a) plug a lot of the
13218 thread-safety problems that existed, and (b) makes it possible to clean
13219 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13220 such data would previously have always leaked in application code and
13221 workarounds were in place to make the memory debugging turn a blind eye
13222 to it. Application code that doesn't use this new function will still
13223 leak as before, but their memory debugging output will announce it now
13224 rather than letting it slide.
13226 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13227 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13228 has a return value to indicate success or failure.
13232 * Make it possible to replace the underlying "ERR" functions such that the
13233 global state (2 LHASH tables and 2 locks) is only used by the "default"
13234 implementation. This change also adds two functions to "get" and "set"
13235 the implementation prior to it being automatically set the first time
13236 any other ERR function takes place. Ie. an application can call "get",
13237 pass the return value to a module it has just loaded, and that module
13238 can call its own "set" function using that value. This means the
13239 module's "ERR" operations will use (and modify) the error state in the
13240 application and not in its own statically linked copy of OpenSSL code.
13244 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13245 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13246 the operation, and provides a more encapsulated way for external code
13247 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13248 to use these functions rather than manually incrementing the counts.
13250 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13254 * Add EVP test program.
13258 * Add symmetric cipher support to ENGINE. Expect the API to change!
13262 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13263 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13264 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13265 These allow a CRL to be built without having to access X509_CRL fields
13266 directly. Modify 'ca' application to use new functions.
13270 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13271 bug workarounds. Rollback attack detection is a security feature.
13272 The problem will only arise on OpenSSL servers when TLSv1 is not
13273 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13274 Software authors not wanting to support TLSv1 will have special reasons
13275 for their choice and can explicitly enable this option.
13277 *Bodo Moeller, Lutz Jaenicke*
13279 * Rationalise EVP so it can be extended: don't include a union of
13280 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13281 (similar to those existing for EVP_CIPHER_CTX).
13286 EVP_MD_CTX_init(&md); /* new function call */
13287 EVP_DigestInit(&md, EVP_sha1());
13288 EVP_DigestUpdate(&md, in, len);
13289 EVP_DigestFinal(&md, out, NULL);
13290 EVP_MD_CTX_cleanup(&md); /* new function call */
13294 * Make DES key schedule conform to the usual scheme, as well as
13295 correcting its structure. This means that calls to DES functions
13296 now have to pass a pointer to a des_key_schedule instead of a
13297 plain des_key_schedule (which was actually always a pointer
13300 des_key_schedule ks;
13302 des_set_key_checked(..., &ks);
13303 des_ncbc_encrypt(..., &ks, ...);
13305 (Note that a later change renames 'des_...' into 'DES_...'.)
13309 * Initial reduction of linker bloat: the use of some functions, such as
13310 PEM causes large amounts of unused functions to be linked in due to
13311 poor organisation. For example pem_all.c contains every PEM function
13312 which has a knock on effect of linking in large amounts of (unused)
13313 ASN1 code. Grouping together similar functions and splitting unrelated
13314 functions prevents this.
13318 * Cleanup of EVP macros.
13322 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13323 correct `_ecb suffix`.
13327 * Add initial OCSP responder support to ocsp application. The
13328 revocation information is handled using the text based index
13329 use by the ca application. The responder can either handle
13330 requests generated internally, supplied in files (for example
13331 via a CGI script) or using an internal minimal server.
13335 * Add configuration choices to get zlib compression for TLS.
13339 * Changes to Kerberos SSL for RFC 2712 compliance:
13340 1. Implemented real KerberosWrapper, instead of just using
13341 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13342 2. Implemented optional authenticator field of KerberosWrapper.
13344 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13345 and authenticator structs; see crypto/krb5/.
13347 Generalized Kerberos calls to support multiple Kerberos libraries.
13348 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13349 via Richard Levitte*
13351 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13352 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13353 values for each of the key sizes rather than having just
13354 parameters (and 'speed' generating keys each time).
13358 * Speed up EVP routines.
13361 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13362 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13363 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13364 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13366 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13367 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13368 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13371 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13373 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13377 * Added the OS2-EMX target.
13379 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13381 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13382 New functions to support `NCONF` routines in extension code.
13383 New function `CONF_set_nconf()`
13384 to allow functions which take an `NCONF` to also handle the old `LHASH`
13385 structure: this means that the old `CONF` compatible routines can be
13386 retained (in particular w.rt. extensions) without having to duplicate the
13387 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13391 * Enhance the general user interface with mechanisms for inner control
13392 and with possibilities to have yes/no kind of prompts.
13396 * Change all calls to low-level digest routines in the library and
13397 applications to use EVP. Add missing calls to HMAC_cleanup() and
13398 don't assume HMAC_CTX can be copied using memcpy().
13400 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13402 * Add the possibility to control engines through control names but with
13403 arbitrary arguments instead of just a string.
13404 Change the key loaders to take a UI_METHOD instead of a callback
13405 function pointer. NOTE: this breaks binary compatibility with earlier
13406 versions of OpenSSL [engine].
13407 Adapt the nCipher code for these new conditions and add a card insertion
13412 * Enhance the general user interface with mechanisms to better support
13413 dialog box interfaces, application-defined prompts, the possibility
13414 to use defaults (for example default passwords from somewhere else)
13415 and interrupts/cancellations.
13419 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13420 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13424 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13425 tidy up some unnecessarily weird code in 'sk_new()').
13427 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13429 * Change the key loading routines for ENGINEs to use the same kind
13430 callback (pem_password_cb) as all other routines that need this
13435 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13436 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13437 than this minimum value is recommended.
13441 * New random seeder for OpenVMS, using the system process statistics
13442 that are easily reachable.
13446 * Windows apparently can't transparently handle global
13447 variables defined in DLLs. Initialisations such as:
13449 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13451 won't compile. This is used by the any applications that need to
13452 declare their own ASN1 modules. This was fixed by adding the option
13453 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13454 needed for static libraries under Win32.
13458 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13459 setting of purpose and trust fields. New X509_STORE trust and
13460 purpose functions and tidy up setting in other SSL functions.
13464 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13465 structure. These are inherited by X509_STORE_CTX when it is
13466 initialised. This allows various defaults to be set in the
13467 X509_STORE structure (such as flags for CRL checking and custom
13468 purpose or trust settings) for functions which only use X509_STORE_CTX
13469 internally such as S/MIME.
13471 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13472 trust settings if they are not set in X509_STORE. This allows X509_STORE
13473 purposes and trust (in S/MIME for example) to override any set by default.
13475 Add command line options for CRL checking to smime, s_client and s_server
13480 * Initial CRL based revocation checking. If the CRL checking flag(s)
13481 are set then the CRL is looked up in the X509_STORE structure and
13482 its validity and signature checked, then if the certificate is found
13483 in the CRL the verify fails with a revoked error.
13485 Various new CRL related callbacks added to X509_STORE_CTX structure.
13487 Command line options added to 'verify' application to support this.
13489 This needs some additional work, such as being able to handle multiple
13490 CRLs with different times, extension based lookup (rather than just
13491 by subject name) and ultimately more complete V2 CRL extension
13496 * Add a general user interface API (crypto/ui/). This is designed
13497 to replace things like des_read_password and friends (backward
13498 compatibility functions using this new API are provided).
13499 The purpose is to remove prompting functions from the DES code
13500 section as well as provide for prompting through dialog boxes in
13501 a window system and the like.
13505 * Add "ex_data" support to ENGINE so implementations can add state at a
13506 per-structure level rather than having to store it globally.
13510 * Make it possible for ENGINE structures to be copied when retrieved by
13511 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13512 This causes the "original" ENGINE structure to act like a template,
13513 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13514 operational state can be localised to each ENGINE structure, despite the
13515 fact they all share the same "methods". New ENGINE structures returned in
13516 this case have no functional references and the return value is the single
13517 structural reference. This matches the single structural reference returned
13518 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13523 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13524 needs to match any other type at all we need to manually clear the
13529 * Changes to the "openssl engine" utility to include;
13530 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13531 about an ENGINE's available control commands.
13532 - executing control commands from command line arguments using the
13533 '-pre' and '-post' switches. '-post' is only used if '-t' is
13534 specified and the ENGINE is successfully initialised. The syntax for
13535 the individual commands are colon-separated, for example;
13536 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13540 * New dynamic control command support for ENGINEs. ENGINEs can now
13541 declare their own commands (numbers), names (strings), descriptions,
13542 and input types for run-time discovery by calling applications. A
13543 subset of these commands are implicitly classed as "executable"
13544 depending on their input type, and only these can be invoked through
13545 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13546 can be based on user input, config files, etc). The distinction is
13547 that "executable" commands cannot return anything other than a boolean
13548 result and can only support numeric or string input, whereas some
13549 discoverable commands may only be for direct use through
13550 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13551 pointers, or other custom uses. The "executable" commands are to
13552 support parameterisations of ENGINE behaviour that can be
13553 unambiguously defined by ENGINEs and used consistently across any
13554 OpenSSL-based application. Commands have been added to all the
13555 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13556 control over shared-library paths without source code alterations.
13560 * Changed all ENGINE implementations to dynamically allocate their
13561 ENGINEs rather than declaring them statically. Apart from this being
13562 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13563 this also allows the implementations to compile without using the
13564 internal engine_int.h header.
13568 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13569 'const' value. Any code that should be able to modify a RAND_METHOD
13570 should already have non-const pointers to it (ie. they should only
13571 modify their own ones).
13575 * Made a variety of little tweaks to the ENGINE code.
13576 - "atalla" and "ubsec" string definitions were moved from header files
13577 to C code. "nuron" string definitions were placed in variables
13578 rather than hard-coded - allowing parameterisation of these values
13579 later on via ctrl() commands.
13580 - Removed unused "#if 0"'d code.
13581 - Fixed engine list iteration code so it uses ENGINE_free() to release
13582 structural references.
13583 - Constified the RAND_METHOD element of ENGINE structures.
13584 - Constified various get/set functions as appropriate and added
13585 missing functions (including a catch-all ENGINE_cpy that duplicates
13586 all ENGINE values onto a new ENGINE except reference counts/state).
13587 - Removed NULL parameter checks in get/set functions. Setting a method
13588 or function to NULL is a way of cancelling out a previously set
13589 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13590 and doesn't justify the extra error symbols and code.
13591 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13592 flags from engine_int.h to engine.h.
13593 - Changed prototypes for ENGINE handler functions (init(), finish(),
13594 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13598 * Implement binary inversion algorithm for BN_mod_inverse in addition
13599 to the algorithm using long division. The binary algorithm can be
13600 used only if the modulus is odd. On 32-bit systems, it is faster
13601 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13602 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13603 up to 450 bits. In 64-bit environments, the binary algorithm
13604 appears to be advantageous for much longer moduli; here we use it
13605 for moduli up to 2048 bits.
13609 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13610 could not support the combine flag in choice fields.
13614 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13615 extensions from a certificate request to the certificate.
13619 * Allow multiple 'certopt' and 'nameopt' options to be separated
13620 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13621 file: this allows the display of the certificate about to be
13622 signed to be customised, to allow certain fields to be included
13623 or excluded and extension details. The old system didn't display
13624 multicharacter strings properly, omitted fields not in the policy
13625 and couldn't display additional details such as extensions.
13629 * Function EC_POINTs_mul for multiple scalar multiplication
13630 of an arbitrary number of elliptic curve points
13631 \sum scalars[i]*points[i],
13632 optionally including the generator defined for the EC_GROUP:
13633 scalar*generator + \sum scalars[i]*points[i].
13635 EC_POINT_mul is a simple wrapper function for the typical case
13636 that the point list has just one item (besides the optional
13641 * First EC_METHODs for curves over GF(p):
13643 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13644 operations and provides various method functions that can also
13645 operate with faster implementations of modular arithmetic.
13647 EC_GFp_mont_method() reuses most functions that are part of
13648 EC_GFp_simple_method, but uses Montgomery arithmetic.
13650 *Bodo Moeller; point addition and point doubling
13651 implementation directly derived from source code provided by
13652 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13654 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13655 crypto/ec/ec_lib.c):
13657 Curves are EC_GROUP objects (with an optional group generator)
13658 based on EC_METHODs that are built into the library.
13660 Points are EC_POINT objects based on EC_GROUP objects.
13662 Most of the framework would be able to handle curves over arbitrary
13663 finite fields, but as there are no obvious types for fields other
13664 than GF(p), some functions are limited to that for now.
13668 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13669 that the file contains a complete HTTP response.
13673 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13674 change the def and num file printf format specifier from "%-40sXXX"
13675 to "%-39s XXX". The latter will always guarantee a space after the
13676 field while the former will cause them to run together if the field
13677 is 40 of more characters long.
13681 * Constify the cipher and digest 'method' functions and structures
13682 and modify related functions to take constant EVP_MD and EVP_CIPHER
13687 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13688 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13692 * Modify `EVP_Digest*()` routines so they now return values. Although the
13693 internal software routines can never fail additional hardware versions
13698 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13700 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13701 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13704 ERR_R_NESTED_ASN1_ERROR
13706 ERR_R_MISSING_ASN1_EOS
13707 were 4 .. 9, conflicting with
13708 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13710 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13711 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13713 Add new error code 'ERR_R_INTERNAL_ERROR'.
13717 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13722 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13723 sets the subject name for a new request or supersedes the
13724 subject name in a given request. Formats that can be parsed are
13725 'CN=Some Name, OU=myOU, C=IT'
13727 'CN=Some Name/OU=myOU/C=IT'.
13729 Add options '-batch' and '-verbose' to 'openssl req'.
13731 *Massimiliano Pala <madwolf@hackmasters.net>*
13733 * Introduce the possibility to access global variables through
13734 functions on platform were that's the best way to handle exporting
13735 global variables in shared libraries. To enable this functionality,
13736 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13737 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13738 is normally done by Configure or something similar).
13740 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13741 in the source file (foo.c) like this:
13743 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13744 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13746 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13747 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13749 OPENSSL_DECLARE_GLOBAL(int,foo);
13750 #define foo OPENSSL_GLOBAL_REF(foo)
13751 OPENSSL_DECLARE_GLOBAL(double,bar);
13752 #define bar OPENSSL_GLOBAL_REF(bar)
13754 The #defines are very important, and therefore so is including the
13755 header file everywhere where the defined globals are used.
13757 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13758 of ASN.1 items, but that structure is a bit different.
13760 The largest change is in util/mkdef.pl which has been enhanced with
13761 better and easier to understand logic to choose which symbols should
13762 go into the Windows .def files as well as a number of fixes and code
13763 cleanup (among others, algorithm keywords are now sorted
13764 lexicographically to avoid constant rewrites).
13768 * In BN_div() keep a copy of the sign of 'num' before writing the
13769 result to 'rm' because if rm==num the value will be overwritten
13770 and produce the wrong result if 'num' is negative: this caused
13771 problems with BN_mod() and BN_nnmod().
13775 * Function OCSP_request_verify(). This checks the signature on an
13776 OCSP request and verifies the signer certificate. The signer
13777 certificate is just checked for a generic purpose and OCSP request
13782 * Add OCSP_check_validity() function to check the validity of OCSP
13783 responses. OCSP responses are prepared in real time and may only
13784 be a few seconds old. Simply checking that the current time lies
13785 between thisUpdate and nextUpdate max reject otherwise valid responses
13786 caused by either OCSP responder or client clock inaccuracy. Instead
13787 we allow thisUpdate and nextUpdate to fall within a certain period of
13788 the current time. The age of the response can also optionally be
13789 checked. Two new options -validity_period and -status_age added to
13794 * If signature or public key algorithm is unrecognized print out its
13795 OID rather that just UNKNOWN.
13799 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13800 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13801 ID to be generated from the issuer certificate alone which can then be
13802 passed to OCSP_id_issuer_cmp().
13806 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13807 ASN1 modules to export functions returning ASN1_ITEM pointers
13808 instead of the ASN1_ITEM structures themselves. This adds several
13809 new macros which allow the underlying ASN1 function/structure to
13810 be accessed transparently. As a result code should not use ASN1_ITEM
13811 references directly (such as &X509_it) but instead use the relevant
13812 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13813 use of the new ASN1 code on platforms where exporting structures
13814 is problematical (for example in shared libraries) but exporting
13815 functions returning pointers to structures is not.
13819 * Add support for overriding the generation of SSL/TLS session IDs.
13820 These callbacks can be registered either in an SSL_CTX or per SSL.
13821 The purpose of this is to allow applications to control, if they wish,
13822 the arbitrary values chosen for use as session IDs, particularly as it
13823 can be useful for session caching in multiple-server environments. A
13824 command-line switch for testing this (and any client code that wishes
13825 to use such a feature) has been added to "s_server".
13827 *Geoff Thorpe, Lutz Jaenicke*
13829 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13830 of the form `#if defined(...) || defined(...) || ...` and
13831 `#if !defined(...) && !defined(...) && ...`. This also avoids
13832 the growing number of special cases it was previously handling.
13836 * Make all configuration macros available for application by making
13837 sure they are available in opensslconf.h, by giving them names starting
13838 with `OPENSSL_` to avoid conflicts with other packages and by making
13839 sure e_os2.h will cover all platform-specific cases together with
13841 Additionally, it is now possible to define configuration/platform-
13842 specific names (called "system identities"). In the C code, these
13843 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13844 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13845 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13850 * New option -set_serial to 'req' and 'x509' this allows the serial
13851 number to use to be specified on the command line. Previously self
13852 signed certificates were hard coded with serial number 0 and the
13853 CA options of 'x509' had to use a serial number in a file which was
13858 * New options to 'ca' utility to support V2 CRL entry extensions.
13859 Currently CRL reason, invalidity date and hold instruction are
13860 supported. Add new CRL extensions to V3 code and some new objects.
13864 * New function EVP_CIPHER_CTX_set_padding() this is used to
13865 disable standard block padding (aka PKCS#5 padding) in the EVP
13866 API, which was previously mandatory. This means that the data is
13867 not padded in any way and so the total length much be a multiple
13868 of the block size, otherwise an error occurs.
13872 * Initial (incomplete) OCSP SSL support.
13876 * New function OCSP_parse_url(). This splits up a URL into its host,
13877 port and path components: primarily to parse OCSP URLs. New -url
13878 option to ocsp utility.
13882 * New nonce behavior. The return value of OCSP_check_nonce() now
13883 reflects the various checks performed. Applications can decide
13884 whether to tolerate certain situations such as an absent nonce
13885 in a response when one was present in a request: the ocsp application
13886 just prints out a warning. New function OCSP_add1_basic_nonce()
13887 this is to allow responders to include a nonce in a response even if
13888 the request is nonce-less.
13892 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13893 skipped when using openssl x509 multiple times on a single input file,
13894 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13898 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13899 set string type: to handle setting ASN1_TIME structures. Fix ca
13900 utility to correctly initialize revocation date of CRLs.
13904 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13905 the clients preferred ciphersuites and rather use its own preferences.
13906 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13907 Internet Explorer by ensuring unchanged hash method during stepup.
13908 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13912 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13913 to aes and add a new 'exist' option to print out symbols that don't
13918 * Additional options to ocsp utility to allow flags to be set and
13919 additional certificates supplied.
13923 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13924 OCSP client a number of certificate to only verify the response
13929 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13930 handle the new API. Currently only ECB, CBC modes supported. Add new
13933 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13934 Encryption Standard (AES) Ciphersuites for Transport Layer
13935 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13936 not enabled by default and were not part of the "ALL" ciphersuite
13937 alias because they were not yet official; they could be
13938 explicitly requested by specifying the "AESdraft" ciphersuite
13939 group alias. In the final release of OpenSSL 0.9.7, the group
13940 alias is called "AES" and is part of "ALL".)
13942 *Ben Laurie, Steve Henson, Bodo Moeller*
13944 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13945 request to response.
13949 * Functions for OCSP responders. OCSP_request_onereq_count(),
13950 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13951 extract information from a certificate request. OCSP_response_create()
13952 creates a response and optionally adds a basic response structure.
13953 OCSP_basic_add1_status() adds a complete single response to a basic
13954 response and returns the OCSP_SINGLERESP structure just added (to allow
13955 extensions to be included for example). OCSP_basic_add1_cert() adds a
13956 certificate to a basic response and OCSP_basic_sign() signs a basic
13957 response with various flags. New helper functions ASN1_TIME_check()
13958 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13959 (converts ASN1_TIME to GeneralizedTime).
13963 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13964 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13965 structure from a certificate. X509_pubkey_digest() digests the public_key
13966 contents: this is used in various key identifiers.
13970 * Make sk_sort() tolerate a NULL argument.
13972 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13974 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13975 passed by the function are trusted implicitly. If any of them signed the
13976 response then it is assumed to be valid and is not verified.
13980 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13981 to data. This was previously part of the PKCS7 ASN1 code. This
13982 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13983 *Steve Henson, reported by Kenneth R. Robinette
13984 <support@securenetterm.com>*
13986 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13987 routines: without these tracing memory leaks is very painful.
13988 Fix leaks in PKCS12 and PKCS7 routines.
13992 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13993 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13994 effectively meant GeneralizedTime would never be used. Now it
13995 is initialised to -1 but X509_time_adj() now has to check the value
13996 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13997 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13998 *Steve Henson, reported by Kenneth R. Robinette
13999 <support@securenetterm.com>*
14001 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14002 result in a zero length in the ASN1_INTEGER structure which was
14003 not consistent with the structure when d2i_ASN1_INTEGER() was used
14004 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14005 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14006 where it did not print out a minus for negative ASN1_INTEGER.
14010 * Add summary printout to ocsp utility. The various functions which
14011 convert status values to strings have been renamed to:
14012 OCSP_response_status_str(), OCSP_cert_status_str() and
14013 OCSP_crl_reason_str() and are no longer static. New options
14014 to verify nonce values and to disable verification. OCSP response
14015 printout format cleaned up.
14019 * Add additional OCSP certificate checks. These are those specified
14020 in RFC2560. This consists of two separate checks: the CA of the
14021 certificate being checked must either be the OCSP signer certificate
14022 or the issuer of the OCSP signer certificate. In the latter case the
14023 OCSP signer certificate must contain the OCSP signing extended key
14024 usage. This check is performed by attempting to match the OCSP
14025 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14026 in the OCSP_CERTID structures of the response.
14030 * Initial OCSP certificate verification added to OCSP_basic_verify()
14031 and related routines. This uses the standard OpenSSL certificate
14032 verify routines to perform initial checks (just CA validity) and
14033 to obtain the certificate chain. Then additional checks will be
14034 performed on the chain. Currently the root CA is checked to see
14035 if it is explicitly trusted for OCSP signing. This is used to set
14036 a root CA as a global signing root: that is any certificate that
14037 chains to that CA is an acceptable OCSP signing certificate.
14041 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14042 extensions from a separate configuration file.
14043 As when reading extensions from the main configuration file,
14044 the '-extensions ...' option may be used for specifying the
14047 *Massimiliano Pala <madwolf@comune.modena.it>*
14049 * New OCSP utility. Allows OCSP requests to be generated or
14050 read. The request can be sent to a responder and the output
14051 parsed, outputted or printed in text form. Not complete yet:
14052 still needs to check the OCSP response validity.
14056 * New subcommands for 'openssl ca':
14057 `openssl ca -status <serial>` prints the status of the cert with
14058 the given serial number (according to the index file).
14059 `openssl ca -updatedb` updates the expiry status of certificates
14062 *Massimiliano Pala <madwolf@comune.modena.it>*
14064 * New '-newreq-nodes' command option to CA.pl. This is like
14065 '-newreq', but calls 'openssl req' with the '-nodes' option
14066 so that the resulting key is not encrypted.
14068 *Damien Miller <djm@mindrot.org>*
14070 * New configuration for the GNU Hurd.
14072 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14074 * Initial code to implement OCSP basic response verify. This
14075 is currently incomplete. Currently just finds the signer's
14076 certificate and verifies the signature on the response.
14080 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14081 value of OPENSSLDIR. This is available via the new '-d' option
14082 to 'openssl version', and is also included in 'openssl version -a'.
14086 * Allowing defining memory allocation callbacks that will be given
14087 file name and line number information in additional arguments
14088 (a `const char*` and an int). The basic functionality remains, as
14089 well as the original possibility to just replace malloc(),
14090 realloc() and free() by functions that do not know about these
14091 additional arguments. To register and find out the current
14092 settings for extended allocation functions, the following
14093 functions are provided:
14095 CRYPTO_set_mem_ex_functions
14096 CRYPTO_set_locked_mem_ex_functions
14097 CRYPTO_get_mem_ex_functions
14098 CRYPTO_get_locked_mem_ex_functions
14100 These work the same way as CRYPTO_set_mem_functions and friends.
14101 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14102 extended allocation function is enabled.
14103 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14104 a conventional allocation function is enabled.
14106 *Richard Levitte, Bodo Moeller*
14108 * Finish off removing the remaining LHASH function pointer casts.
14109 There should no longer be any prototype-casting required when using
14110 the LHASH abstraction, and any casts that remain are "bugs". See
14111 the callback types and macros at the head of lhash.h for details
14112 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14116 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14117 If /dev/[u]random devices are not available or do not return enough
14118 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14120 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14121 /etc/entropy will be queried once each in this sequence, querying stops
14122 when enough entropy was collected without querying more sockets.
14126 * Change the Unix RAND_poll() variant to be able to poll several
14127 random devices, as specified by DEVRANDOM, until a sufficient amount
14128 of data has been collected. We spend at most 10 ms on each file
14129 (select timeout) and read in non-blocking mode. DEVRANDOM now
14130 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14131 (previously it was just the string "/dev/urandom"), so on typical
14132 platforms the 10 ms delay will never occur.
14133 Also separate out the Unix variant to its own file, rand_unix.c.
14134 For VMS, there's a currently-empty rand_vms.c.
14138 * Move OCSP client related routines to ocsp_cl.c. These
14139 provide utility functions which an application needing
14140 to issue a request to an OCSP responder and analyse the
14141 response will typically need: as opposed to those which an
14142 OCSP responder itself would need which will be added later.
14144 OCSP_request_sign() signs an OCSP request with an API similar
14145 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14146 response. OCSP_response_get1_basic() extracts basic response
14147 from response. OCSP_resp_find_status(): finds and extracts status
14148 information from an OCSP_CERTID structure (which will be created
14149 when the request structure is built). These are built from lower
14150 level functions which work on OCSP_SINGLERESP structures but
14151 won't normally be used unless the application wishes to examine
14152 extensions in the OCSP response for example.
14154 Replace nonce routines with a pair of functions.
14155 OCSP_request_add1_nonce() adds a nonce value and optionally
14156 generates a random value. OCSP_check_nonce() checks the
14157 validity of the nonce in an OCSP response.
14161 * Change function OCSP_request_add() to OCSP_request_add0_id().
14162 This doesn't copy the supplied OCSP_CERTID and avoids the
14163 need to free up the newly created id. Change return type
14164 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14165 This can then be used to add extensions to the request.
14166 Deleted OCSP_request_new(), since most of its functionality
14167 is now in OCSP_REQUEST_new() (and the case insensitive name
14168 clash) apart from the ability to set the request name which
14169 will be added elsewhere.
14173 * Update OCSP API. Remove obsolete extensions argument from
14174 various functions. Extensions are now handled using the new
14175 OCSP extension code. New simple OCSP HTTP function which
14176 can be used to send requests and parse the response.
14180 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14181 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14182 uses the special reorder version of SET OF to sort the attributes
14183 and reorder them to match the encoded order. This resolves a long
14184 standing problem: a verify on a PKCS7 structure just after signing
14185 it used to fail because the attribute order did not match the
14186 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14187 it uses the received order. This is necessary to tolerate some broken
14188 software that does not order SET OF. This is handled by encoding
14189 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14190 to produce the required SET OF.
14194 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14195 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14196 files to get correct declarations of the ASN.1 item variables.
14200 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14201 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14202 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14203 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14204 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14205 ASN1_ITEM and no wrapper functions.
14209 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14210 replace the old function pointer based I/O routines. Change most of
14211 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14215 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14216 lines, recognize more "algorithms" that can be deselected, and make
14217 it complain about algorithm deselection that isn't recognised.
14221 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14222 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14223 to use new functions. Add NO_ASN1_OLD which can be set to remove
14224 some old style ASN1 functions: this can be used to determine if old
14225 code will still work when these eventually go away.
14229 * New extension functions for OCSP structures, these follow the
14230 same conventions as certificates and CRLs.
14234 * New function X509V3_add1_i2d(). This automatically encodes and
14235 adds an extension. Its behaviour can be customised with various
14236 flags to append, replace or delete. Various wrappers added for
14237 certificates and CRLs.
14241 * Fix to avoid calling the underlying ASN1 print routine when
14242 an extension cannot be parsed. Correct a typo in the
14243 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14247 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14248 entries for variables.
14252 * Add functionality to `apps/openssl.c` for detecting locking
14253 problems: As the program is single-threaded, all we have
14254 to do is register a locking callback using an array for
14255 storing which locks are currently held by the program.
14259 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14260 SSL_get_ex_data_X509_STORE_idx(), which is used in
14261 ssl_verify_cert_chain() and thus can be called at any time
14262 during TLS/SSL handshakes so that thread-safety is essential.
14263 Unfortunately, the ex_data design is not at all suited
14264 for multi-threaded use, so it probably should be abolished.
14268 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14270 *Broadcom, tweaked and integrated by Geoff Thorpe*
14272 * Move common extension printing code to new function
14273 X509V3_print_extensions(). Reorganise OCSP print routines and
14274 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14278 * New function X509_signature_print() to remove duplication in some
14283 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14284 set (this was treated exactly the same as SET OF previously). This
14285 is used to reorder the STACK representing the structure to match the
14286 encoding. This will be used to get round a problem where a PKCS7
14287 structure which was signed could not be verified because the STACK
14288 order did not reflect the encoded order.
14292 * Reimplement the OCSP ASN1 module using the new code.
14296 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14297 for its ASN1 operations. The old style function pointers still exist
14298 for now but they will eventually go away.
14302 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14303 completely replaces the old ASN1 functionality with a table driven
14304 encoder and decoder which interprets an ASN1_ITEM structure describing
14305 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14306 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14307 has also been converted to the new form.
14311 * Change BN_mod_exp_recp so that negative moduli are tolerated
14312 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14313 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14314 for negative moduli.
14318 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14319 of not touching the result's sign bit.
14323 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14328 * Changed the LHASH code to use prototypes for callbacks, and created
14329 macros to declare and implement thin (optionally static) functions
14330 that provide type-safety and avoid function pointer casting for the
14331 type-specific callbacks.
14335 * Added Kerberos Cipher Suites to be used with TLS, as written in
14337 *Veers Staats <staatsvr@asc.hpc.mil>,
14338 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14340 * Reformat the FAQ so the different questions and answers can be divided
14341 in sections depending on the subject.
14345 * Have the zlib compression code load ZLIB.DLL dynamically under
14350 * New function BN_mod_sqrt for computing square roots modulo a prime
14351 (using the probabilistic Tonelli-Shanks algorithm unless
14352 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14353 be handled deterministically).
14355 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14357 * Make BN_mod_inverse faster by explicitly handling small quotients
14358 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14359 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14363 * New function BN_kronecker.
14367 * Fix BN_gcd so that it works on negative inputs; the result is
14368 positive unless both parameters are zero.
14369 Previously something reasonably close to an infinite loop was
14370 possible because numbers could be growing instead of shrinking
14371 in the implementation of Euclid's algorithm.
14375 * Fix BN_is_word() and BN_is_one() macros to take into account the
14376 sign of the number in question.
14378 Fix BN_is_word(a,w) to work correctly for w == 0.
14380 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14381 because its test if the absolute value of 'a' equals 'w'.
14382 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14383 it exists mostly for use in the implementations of BN_is_zero(),
14384 BN_is_one(), and BN_is_word().
14388 * New function BN_swap.
14392 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14393 the exponentiation functions are more likely to produce reasonable
14394 results on negative inputs.
14398 * Change BN_mod_mul so that the result is always non-negative.
14399 Previously, it could be negative if one of the factors was negative;
14400 I don't think anyone really wanted that behaviour.
14404 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14405 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14406 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14407 and add new functions:
14416 BN_mod_lshift1_quick
14418 BN_mod_lshift_quick
14420 These functions always generate non-negative results.
14422 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14423 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14425 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14426 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14427 be reduced modulo `m`.
14429 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14432 The following entry accidentally appeared in the CHANGES file
14433 distributed with OpenSSL 0.9.7. The modifications described in
14434 it do *not* apply to OpenSSL 0.9.7.
14436 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14437 was actually never needed) and in BN_mul(). The removal in BN_mul()
14438 required a small change in bn_mul_part_recursive() and the addition
14439 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14440 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14441 bn_sub_words() and bn_add_words() except they take arrays with
14447 * In 'openssl passwd', verify passwords read from the terminal
14448 unless the '-salt' option is used (which usually means that
14449 verification would just waste user's time since the resulting
14450 hash is going to be compared with some given password hash)
14451 or the new '-noverify' option is used.
14453 This is an incompatible change, but it does not affect
14454 non-interactive use of 'openssl passwd' (passwords on the command
14455 line, '-stdin' option, '-in ...' option) and thus should not
14456 cause any problems.
14460 * Remove all references to RSAref, since there's no more need for it.
14464 * Make DSO load along a path given through an environment variable
14465 (SHLIB_PATH) with shl_load().
14469 * Constify the ENGINE code as a result of BIGNUM constification.
14470 Also constify the RSA code and most things related to it. In a
14471 few places, most notable in the depth of the ASN.1 code, ugly
14472 casts back to non-const were required (to be solved at a later
14477 * Make it so the openssl application has all engines loaded by default.
14481 * Constify the BIGNUM routines a little more.
14485 * Add the following functions:
14487 ENGINE_load_cswift()
14489 ENGINE_load_atalla()
14490 ENGINE_load_nuron()
14491 ENGINE_load_builtin_engines()
14493 That way, an application can itself choose if external engines that
14494 are built-in in OpenSSL shall ever be used or not. The benefit is
14495 that applications won't have to be linked with libdl or other dso
14496 libraries unless it's really needed.
14498 Changed 'openssl engine' to load all engines on demand.
14499 Changed the engine header files to avoid the duplication of some
14500 declarations (they differed!).
14504 * 'openssl engine' can now list capabilities.
14508 * Better error reporting in 'openssl engine'.
14512 * Never call load_dh_param(NULL) in s_server.
14516 * Add engine application. It can currently list engines by name and
14517 identity, and test if they are actually available.
14521 * Improve RPM specification file by forcing symbolic linking and making
14522 sure the installed documentation is also owned by root.root.
14524 *Damien Miller <djm@mindrot.org>*
14526 * Give the OpenSSL applications more possibilities to make use of
14527 keys (public as well as private) handled by engines.
14531 * Add OCSP code that comes from CertCo.
14535 * Add VMS support for the Rijndael code.
14539 * Added untested support for Nuron crypto accelerator.
14543 * Add support for external cryptographic devices. This code was
14544 previously distributed separately as the "engine" branch.
14546 *Geoff Thorpe, Richard Levitte*
14548 * Rework the filename-translation in the DSO code. It is now possible to
14549 have far greater control over how a "name" is turned into a filename
14550 depending on the operating environment and any oddities about the
14551 different shared library filenames on each system.
14555 * Support threads on FreeBSD-elf in Configure.
14559 * Fix for SHA1 assembly problem with MASM: it produces
14560 warnings about corrupt line number information when assembling
14561 with debugging information. This is caused by the overlapping
14564 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14567 NCONF_get_number() has no error checking at all. As a replacement,
14568 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14569 promoted strongly. The old NCONF_get_number is kept around for
14570 binary backward compatibility.
14571 Make it possible for methods to load from something other than a BIO,
14572 by providing a function pointer that is given a name instead of a BIO.
14573 For example, this could be used to load configuration data from an
14578 * Fix for non blocking accept BIOs. Added new I/O special reason
14579 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14580 with non blocking I/O was not possible because no retry code was
14581 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14586 * Added the beginnings of Rijndael support.
14590 * Fix for bug in DirectoryString mask setting. Add support for
14591 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14592 to allow certificate printing to more controllable, additional
14593 'certopt' option to 'x509' to allow new printing options to be
14598 * Clean old EAY MD5 hack from e_os.h.
14602 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14604 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14605 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14607 *Joe Orton, Steve Henson*
14609 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14611 * Fix additional bug revealed by the NISCC test suite:
14613 Stop bug triggering large recursion when presented with
14614 certain ASN.1 tags ([CVE-2003-0851])
14618 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14620 * Fix various bugs revealed by running the NISCC test suite:
14622 Stop out of bounds reads in the ASN1 code when presented with
14623 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14625 If verify callback ignores invalid public key errors don't try to check
14626 certificate signature with the NULL public key.
14630 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14631 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14636 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14637 extra data after the compression methods not only for TLS 1.0
14638 but also for SSL 3.0 (as required by the specification).
14640 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14642 * Change X509_certificate_type() to mark the key as exported/exportable
14643 when it's 512 *bits* long, not 512 bytes.
14647 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14649 * Countermeasure against the Klima-Pokorny-Rosa extension of
14650 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14651 a protocol version number mismatch like a decryption error
14652 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14656 * Turn on RSA blinding by default in the default implementation
14657 to avoid a timing attack. Applications that don't want it can call
14658 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14659 They would be ill-advised to do so in most cases.
14661 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14663 * Change RSA blinding code so that it works when the PRNG is not
14664 seeded (in this case, the secret RSA exponent is abused as
14665 an unpredictable seed -- if it is not unpredictable, there
14666 is no point in blinding anyway). Make RSA blinding thread-safe
14667 by remembering the creator's thread ID in rsa->blinding and
14668 having all other threads use local one-time blinding factors
14669 (this requires more computation than sharing rsa->blinding, but
14670 avoids excessive locking; and if an RSA object is not shared
14671 between threads, blinding will still be very fast).
14675 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14677 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14678 via timing by performing a MAC computation even if incorrect
14679 block cipher padding has been found. This is a countermeasure
14680 against active attacks where the attacker has to distinguish
14681 between bad padding and a MAC verification error. ([CVE-2003-0078])
14683 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14684 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14685 Martin Vuagnoux (EPFL, Ilion)*
14687 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14689 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14690 memory from its contents. This is done with a counter that will
14691 place alternating values in each byte. This can be used to solve
14692 two issues: 1) the removal of calls to memset() by highly optimizing
14693 compilers, and 2) cleansing with other values than 0, since those can
14694 be read through on certain media, for example a swap space on disk.
14698 * Bugfix: client side session caching did not work with external caching,
14699 because the session->cipher setting was not restored when reloading
14700 from the external cache. This problem was masked, when
14701 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14702 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14706 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14707 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14709 *Zeev Lieber <zeev-l@yahoo.com>*
14711 * Undo an undocumented change introduced in 0.9.6e which caused
14712 repeated calls to OpenSSL_add_all_ciphers() and
14713 OpenSSL_add_all_digests() to be ignored, even after calling
14718 * Change the default configuration reader to deal with last line not
14719 being properly terminated.
14723 * Change X509_NAME_cmp() so it applies the special rules on handling
14724 DN values that are of type PrintableString, as well as RDNs of type
14725 emailAddress where the value has the type ia5String.
14727 *stefank@valicert.com via Richard Levitte*
14729 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14730 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14731 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14732 the bitwise-OR of the two for use by the majority of applications
14733 wanting this behaviour, and update the docs. The documented
14734 behaviour and actual behaviour were inconsistent and had been
14735 changing anyway, so this is more a bug-fix than a behavioural
14738 *Geoff Thorpe, diagnosed by Nadav Har'El*
14740 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14741 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14745 * Fix initialization code race conditions in
14746 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14747 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14748 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14749 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14750 ssl2_get_cipher_by_char(),
14751 ssl3_get_cipher_by_char().
14753 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14755 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14756 the cached sessions are flushed, as the remove_cb() might use ex_data
14757 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14758 (see [openssl.org #212]).
14760 *Geoff Thorpe, Lutz Jaenicke*
14762 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14763 length, instead of the encoding length to d2i_ASN1_OBJECT.
14767 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14769 * [In 0.9.6g-engine release:]
14770 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14772 *Lynn Gazis <lgazis@rainbow.com>*
14774 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14776 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14777 and get fix the header length calculation.
14778 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14779 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14781 * Use proper error handling instead of 'assertions' in buffer
14782 overflow checks added in 0.9.6e. This prevents DoS (the
14783 assertions could call abort()).
14785 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14787 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14789 * Add various sanity checks to asn1_get_length() to reject
14790 the ASN1 length bytes if they exceed sizeof(long), will appear
14791 negative or the content length exceeds the length of the
14794 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14796 * Fix cipher selection routines: ciphers without encryption had no flags
14797 for the cipher strength set and where therefore not handled correctly
14798 by the selection routines (PR #130).
14802 * Fix EVP_dsa_sha macro.
14807 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14808 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14809 that was added in OpenSSL 0.9.6d.
14811 As the countermeasure turned out to be incompatible with some
14812 broken SSL implementations, the new option is part of SSL_OP_ALL.
14813 SSL_OP_ALL is usually employed when compatibility with weird SSL
14814 implementations is desired (e.g. '-bugs' option to 's_client' and
14815 's_server'), so the new option is automatically set in many
14820 * Changes in security patch:
14822 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14823 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14824 Air Force Materiel Command, USAF, under agreement number
14827 * Add various sanity checks to asn1_get_length() to reject
14828 the ASN1 length bytes if they exceed sizeof(long), will appear
14829 negative or the content length exceeds the length of the
14830 supplied buffer. ([CVE-2002-0659])
14832 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14834 * Assertions for various potential buffer overflows, not known to
14835 happen in practice.
14837 *Ben Laurie (CHATS)*
14839 * Various temporary buffers to hold ASCII versions of integers were
14840 too small for 64 bit platforms. ([CVE-2002-0655])
14841 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14843 * Remote buffer overflow in SSL3 protocol - an attacker could
14844 supply an oversized session ID to a client. ([CVE-2002-0656])
14846 *Ben Laurie (CHATS)*
14848 * Remote buffer overflow in SSL2 protocol - an attacker could
14849 supply an oversized client master key. ([CVE-2002-0656])
14851 *Ben Laurie (CHATS)*
14853 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14855 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14856 encoded as NULL) with id-dsa-with-sha1.
14858 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14860 * Check various `X509_...()` return values in `apps/req.c`.
14862 *Nils Larsch <nla@trustcenter.de>*
14864 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14865 an end-of-file condition would erroneously be flagged, when the CRLF
14866 was just at the end of a processed block. The bug was discovered when
14867 processing data through a buffering memory BIO handing the data to a
14868 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14869 <ptsekov@syntrex.com> and Nedelcho Stanev.
14873 * Implement a countermeasure against a vulnerability recently found
14874 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14875 before application data chunks to avoid the use of known IVs
14876 with data potentially chosen by the attacker.
14880 * Fix length checks in ssl3_get_client_hello().
14884 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14885 to prevent ssl3_read_internal() from incorrectly assuming that
14886 ssl3_read_bytes() found application data while handshake
14887 processing was enabled when in fact s->s3->in_read_app_data was
14888 merely automatically cleared during the initial handshake.
14890 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14892 * Fix object definitions for Private and Enterprise: they were not
14893 recognized in their shortname (=lowercase) representation. Extend
14894 obj_dat.pl to issue an error when using undefined keywords instead
14895 of silently ignoring the problem (Svenning Sorensen
14896 <sss@sss.dnsalias.net>).
14900 * Fix DH_generate_parameters() so that it works for 'non-standard'
14901 generators, i.e. generators other than 2 and 5. (Previously, the
14902 code did not properly initialise the 'add' and 'rem' values to
14903 BN_generate_prime().)
14905 In the new general case, we do not insist that 'generator' is
14906 actually a primitive root: This requirement is rather pointless;
14907 a generator of the order-q subgroup is just as good, if not
14912 * Map new X509 verification errors to alerts. Discovered and submitted by
14913 Tom Wu <tom@arcot.com>.
14917 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14918 returning non-zero before the data has been completely received
14919 when using non-blocking I/O.
14921 *Bodo Moeller; problem pointed out by John Hughes*
14923 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14925 *Ben Laurie, Lutz Jaenicke*
14927 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14928 Yoram Zahavi <YoramZ@gilian.com>).
14932 * Add information about CygWin 1.3 and on, and preserve proper
14933 configuration for the versions before that.
14935 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14937 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14938 check whether we deal with a copy of a session and do not delete from
14939 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14940 <izhar@checkpoint.com>.
14944 * Do not store session data into the internal session cache, if it
14945 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14946 flag is set). Proposed by Aslam <aslam@funk.com>.
14950 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14955 * [In 0.9.6d-engine release:]
14956 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14958 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14960 * Add the configuration target linux-s390x.
14962 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14964 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14965 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14966 variable as an indication that a ClientHello message has been
14967 received. As the flag value will be lost between multiple
14968 invocations of ssl3_accept when using non-blocking I/O, the
14969 function may not be aware that a handshake has actually taken
14970 place, thus preventing a new session from being added to the
14973 To avoid this problem, we now set s->new_session to 2 instead of
14974 using a local variable.
14976 *Lutz Jaenicke, Bodo Moeller*
14978 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14979 if the SSL_R_LENGTH_MISMATCH error is detected.
14981 *Geoff Thorpe, Bodo Moeller*
14983 * New 'shared_ldflag' column in Configure platform table.
14987 * Fix EVP_CIPHER_mode macro.
14989 *"Dan S. Camper" <dan@bti.net>*
14991 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14992 type, we must throw them away by setting rr->length to 0.
14994 *D P Chang <dpc@qualys.com>*
14996 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14998 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14999 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15000 worked incorrectly for those cases where range = `10..._2` and
15001 `3*range` is two bits longer than range.)
15005 * Only add signing time to PKCS7 structures if it is not already
15010 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15011 OBJ_ld_ce should be OBJ_id_ce.
15012 Also some ip-pda OIDs in crypto/objects/objects.txt were
15013 incorrect (cf. RFC 3039).
15015 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15017 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15018 returns early because it has nothing to do.
15020 *Andy Schneider <andy.schneider@bjss.co.uk>*
15022 * [In 0.9.6c-engine release:]
15023 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15025 *Andy Schneider <andy.schneider@bjss.co.uk>*
15027 * [In 0.9.6c-engine release:]
15028 Add support for Cryptographic Appliance's keyserver technology.
15029 (Use engine 'keyclient')
15031 *Cryptographic Appliances and Geoff Thorpe*
15033 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15034 is called via tools/c89.sh because arguments have to be
15035 rearranged (all '-L' options must appear before the first object
15038 *Richard Shapiro <rshapiro@abinitio.com>*
15040 * [In 0.9.6c-engine release:]
15041 Add support for Broadcom crypto accelerator cards, backported
15044 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15046 * [In 0.9.6c-engine release:]
15047 Add support for SureWare crypto accelerator cards from
15048 Baltimore Technologies. (Use engine 'sureware')
15050 *Baltimore Technologies and Mark Cox*
15052 * [In 0.9.6c-engine release:]
15053 Add support for crypto accelerator cards from Accelerated
15054 Encryption Processing, www.aep.ie. (Use engine 'aep')
15056 *AEP Inc. and Mark Cox*
15058 * Add a configuration entry for gcc on UnixWare.
15060 *Gary Benson <gbenson@redhat.com>*
15062 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15063 messages are stored in a single piece (fixed-length part and
15064 variable-length part combined) and fix various bugs found on the way.
15068 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15069 instead. BIO_gethostbyname() does not know what timeouts are
15070 appropriate, so entries would stay in cache even when they have
15072 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15074 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15075 faced with a pathologically small ClientHello fragment that does
15076 not contain client_version: Instead of aborting with an error,
15077 simply choose the highest available protocol version (i.e.,
15078 TLS 1.0 unless it is disabled). In practice, ClientHello
15079 messages are never sent like this, but this change gives us
15080 strictly correct behaviour at least for TLS.
15084 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15085 never resets s->method to s->ctx->method when called from within
15086 one of the SSL handshake functions.
15088 *Bodo Moeller; problem pointed out by Niko Baric*
15090 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15091 (sent using the client's version number) if client_version is
15092 smaller than the protocol version in use. Also change
15093 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15094 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15095 the client will at least see that alert.
15099 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15104 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15105 client receives HelloRequest while in a handshake.
15107 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15109 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15110 should end in 'break', not 'goto end' which circumvents various
15111 cleanups done in state SSL_ST_OK. But session related stuff
15112 must be disabled for SSL_ST_OK in the case that we just sent a
15115 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15116 before just sending a HelloRequest.
15118 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15120 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15121 reveal whether illegal block cipher padding was found or a MAC
15122 verification error occurred. (Neither SSLerr() codes nor alerts
15123 are directly visible to potential attackers, but the information
15124 may leak via logfiles.)
15126 Similar changes are not required for the SSL 2.0 implementation
15127 because the number of padding bytes is sent in clear for SSL 2.0,
15128 and the extra bytes are just ignored. However ssl/s2_pkt.c
15129 failed to verify that the purported number of padding bytes is in
15134 * Add OpenUNIX-8 support including shared libraries
15135 (Boyd Lynn Gerber <gerberb@zenez.com>).
15139 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15140 'wristwatch attack' using huge encoding parameters (cf.
15141 James H. Manger's CRYPTO 2001 paper). Note that the
15142 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15143 encoding parameters and hence was not vulnerable.
15147 * BN_sqr() bug fix.
15149 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15151 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15152 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15153 followed by modular reduction.
15155 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15157 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15158 equivalent based on BN_pseudo_rand() instead of BN_rand().
15162 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15163 This function was broken, as the check for a new client hello message
15164 to handle SGC did not allow these large messages.
15165 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15169 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15173 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15174 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15178 * Rework the configuration and shared library support for Tru64 Unix.
15179 The configuration part makes use of modern compiler features and
15180 still retains old compiler behavior for those that run older versions
15181 of the OS. The shared library support part includes a variant that
15182 uses the RPATH feature, and is available through the special
15183 configuration target "alpha-cc-rpath", which will never be selected
15186 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15188 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15189 with the same message size as in ssl3_get_certificate_request().
15190 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15191 messages might inadvertently be reject as too long.
15193 *Petr Lampa <lampa@fee.vutbr.cz>*
15195 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15199 * Modified SSL library such that the verify_callback that has been set
15200 specifically for an SSL object with SSL_set_verify() is actually being
15201 used. Before the change, a verify_callback set with this function was
15202 ignored and the verify_callback() set in the SSL_CTX at the time of
15203 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15204 to allow the necessary settings.
15208 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15209 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15210 done automatically (in contradiction to the requirements of the C
15211 standard). This made problems when used from OpenSSH.
15215 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15216 dh->length and always used
15218 BN_rand_range(priv_key, dh->p).
15220 BN_rand_range() is not necessary for Diffie-Hellman, and this
15221 specific range makes Diffie-Hellman unnecessarily inefficient if
15222 dh->length (recommended exponent length) is much smaller than the
15223 length of dh->p. We could use BN_rand_range() if the order of
15224 the subgroup was stored in the DH structure, but we only have
15229 BN_rand(priv_key, l, ...)
15231 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15238 RSA_eay_public_encrypt
15239 RSA_eay_private_decrypt
15240 RSA_eay_private_encrypt (signing)
15241 RSA_eay_public_decrypt (signature verification)
15243 (default implementations for RSA_public_encrypt,
15244 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15245 always reject numbers >= n.
15249 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15250 to synchronize access to 'locking_thread'. This is necessary on
15251 systems where access to 'locking_thread' (an 'unsigned long'
15252 variable) is not atomic.
15256 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15257 *before* setting the 'crypto_lock_rand' flag. The previous code had
15258 a race condition if 0 is a valid thread ID.
15260 *Travis Vitek <vitek@roguewave.com>*
15262 * Add support for shared libraries under Irix.
15264 *Albert Chin-A-Young <china@thewrittenword.com>*
15266 * Add configuration option to build on Linux on both big-endian and
15267 little-endian MIPS.
15269 *Ralf Baechle <ralf@uni-koblenz.de>*
15271 * Add the possibility to create shared libraries on HP-UX.
15275 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15277 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15278 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15279 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15280 PRNG state recovery was possible based on the output of
15281 one PRNG request appropriately sized to gain knowledge on
15282 'md' followed by enough consecutive 1-byte PRNG requests
15283 to traverse all of 'state'.
15285 1. When updating 'md_local' (the current thread's copy of 'md')
15286 during PRNG output generation, hash all of the previous
15287 'md_local' value, not just the half used for PRNG output.
15289 2. Make the number of bytes from 'state' included into the hash
15290 independent from the number of PRNG bytes requested.
15292 The first measure alone would be sufficient to avoid
15293 Markku-Juhani's attack. (Actually it had never occurred
15294 to me that the half of 'md_local' used for chaining was the
15295 half from which PRNG output bytes were taken -- I had always
15296 assumed that the secret half would be used.) The second
15297 measure makes sure that additional data from 'state' is never
15298 mixed into 'md_local' in small portions; this heuristically
15299 further strengthens the PRNG.
15303 * Fix crypto/bn/asm/mips3.s.
15307 * When only the key is given to "enc", the IV is undefined. Print out
15308 an error message in this case.
15312 * Handle special case when X509_NAME is empty in X509 printing routines.
15316 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15317 positive and less than q.
15321 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15322 used: it isn't thread safe and the add_lock_callback should handle
15325 *Paul Rose <Paul.Rose@bridge.com>*
15327 * Verify that incoming data obeys the block size in
15328 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15334 *Ulf Möller, Bodo Möller*
15336 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15337 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15338 when fixing the server behaviour for backwards-compatible 'client
15339 hello' messages. (Note that the attack is impractical against
15340 SSL 3.0 and TLS 1.0 anyway because length and version checking
15341 means that the probability of guessing a valid ciphertext is
15342 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15345 Before 0.9.5, the countermeasure (hide the error by generating a
15346 random 'decryption result') did not work properly because
15347 ERR_clear_error() was missing, meaning that SSL_get_error() would
15348 detect the supposedly ignored error.
15350 Both problems are now fixed.
15354 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15355 (previously it was 1024).
15359 * Fix for compatibility mode trust settings: ignore trust settings
15360 unless some valid trust or reject settings are present.
15364 * Fix for blowfish EVP: its a variable length cipher.
15368 * Fix various bugs related to DSA S/MIME verification. Handle missing
15369 parameters in DSA public key structures and return an error in the
15370 DSA routines if parameters are absent.
15374 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15375 in the current directory if neither $RANDFILE nor $HOME was set.
15376 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15377 caused some confusion to Windows users who haven't defined $HOME.
15378 Thus RAND_file_name() is changed again: e_os.h can define a
15379 DEFAULT_HOME, which will be used if $HOME is not set.
15380 For Windows, we use "C:"; on other platforms, we still require
15381 environment variables.
15383 * Move 'if (!initialized) RAND_poll()' into regions protected by
15384 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15385 having multiple threads call RAND_poll() concurrently.
15389 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15390 combination of a flag and a thread ID variable.
15391 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15392 flag), *other* threads can enter ssleay_add_bytes without obeying
15393 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15394 that they do not hold after the first thread unsets add_do_not_lock).
15398 * Change bctest again: '-x' expressions are not available in all
15399 versions of 'test'.
15403 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15405 * Fix a couple of memory leaks in PKCS7_dataDecode()
15407 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15409 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15410 the default extension for executables, if any. Also, make the perl
15411 scripts that use symlink() to test if it really exists and use "cp"
15412 if it doesn't. All this made OpenSSL compilable and installable in
15417 * Fix for asn1_GetSequence() for indefinite length constructed data.
15418 If SEQUENCE is length is indefinite just set c->slen to the total
15419 amount of data available.
15421 *Steve Henson, reported by shige@FreeBSD.org*
15423 *This change does not apply to 0.9.7.*
15425 * Change bctest to avoid here-documents inside command substitution
15426 (workaround for FreeBSD /bin/sh bug).
15427 For compatibility with Ultrix, avoid shell functions (introduced
15428 in the bctest version that searches along $PATH).
15432 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15433 with des_encrypt() defined on some operating systems, like Solaris
15438 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15439 On the Importance of Eliminating Errors in Cryptographic
15440 Computations, J. Cryptology 14 (2001) 2, 101-119,
15441 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15445 * MIPS assembler BIGNUM division bug fix.
15449 * Disabled incorrect Alpha assembler code.
15453 * Fix PKCS#7 decode routines so they correctly update the length
15454 after reading an EOC for the EXPLICIT tag.
15458 *This change does not apply to 0.9.7.*
15460 * Fix bug in PKCS#12 key generation routines. This was triggered
15461 if a 3DES key was generated with a 0 initial byte. Include
15462 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15463 (but broken) behaviour.
15467 * Enhance bctest to search for a working bc along $PATH and print
15470 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15472 * Fix memory leaks in err.c: free err_data string if necessary;
15473 don't write to the wrong index in ERR_set_error_data.
15477 * Implement ssl23_peek (analogous to ssl23_read), which previously
15482 * Replace rdtsc with `_emit` statements for VC++ version 5.
15484 *Jeremy Cooper <jeremy@baymoo.org>*
15486 * Make it possible to reuse SSLv2 sessions.
15490 * In copy_email() check for >= 0 as a return value for
15491 X509_NAME_get_index_by_NID() since 0 is a valid index.
15493 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15495 * Avoid coredump with unsupported or invalid public keys by checking if
15496 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15497 PKCS7_verify() fails with non detached data.
15501 * Don't use getenv in library functions when run as setuid/setgid.
15502 New function OPENSSL_issetugid().
15506 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15507 due to incorrect handling of multi-threading:
15509 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15511 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15513 3. Count how many times MemCheck_off() has been called so that
15514 nested use can be treated correctly. This also avoids
15515 inband-signalling in the previous code (which relied on the
15516 assumption that thread ID 0 is impossible).
15520 * Add "-rand" option also to s_client and s_server.
15524 * Fix CPU detection on Irix 6.x.
15525 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15526 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15528 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15533 *This change does not apply to 0.9.7.*
15535 * Use the cached encoding of an X509_NAME structure rather than
15536 copying it. This is apparently the reason for the libsafe "errors"
15537 but the code is actually correct.
15541 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15542 Bleichenbacher's DSA attack.
15543 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15544 to be set and top=0 forces the highest bit to be set; top=-1 is new
15545 and leaves the highest bit random.
15547 *Ulf Moeller, Bodo Moeller*
15549 * In the `NCONF_...`-based implementations for `CONF_...` queries
15550 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15551 a temporary CONF structure with the data component set to NULL
15552 (which gives segmentation faults in lh_retrieve).
15553 Instead, use NULL for the CONF pointer in CONF_get_string and
15554 CONF_get_number (which may use environment variables) and directly
15555 return NULL from CONF_get_section.
15559 * Fix potential buffer overrun for EBCDIC.
15563 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15564 keyUsage if basicConstraints absent for a CA.
15568 * Make SMIME_write_PKCS7() write mail header values with a format that
15569 is more generally accepted (no spaces before the semicolon), since
15570 some programs can't parse those values properly otherwise. Also make
15571 sure BIO's that break lines after each write do not create invalid
15576 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15577 macros previously used would not encode an empty SEQUENCE OF
15578 and break the signature.
15582 *This change does not apply to 0.9.7.*
15584 * Zero the premaster secret after deriving the master secret in
15589 * Add some EVP_add_digest_alias registrations (as found in
15590 OpenSSL_add_all_digests()) to SSL_library_init()
15591 aka OpenSSL_add_ssl_algorithms(). This provides improved
15592 compatibility with peers using X.509 certificates
15593 with unconventional AlgorithmIdentifier OIDs.
15597 * Fix for Irix with NO_ASM.
15599 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15601 * ./config script fixes.
15603 *Ulf Moeller, Richard Levitte*
15605 * Fix 'openssl passwd -1'.
15609 * Change PKCS12_key_gen_asc() so it can cope with non null
15610 terminated strings whose length is passed in the passlen
15611 parameter, for example from PEM callbacks. This was done
15612 by adding an extra length parameter to asc2uni().
15614 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15616 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15617 call failed, free the DSA structure.
15621 * Fix to uni2asc() to cope with zero length Unicode strings.
15622 These are present in some PKCS#12 files.
15626 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15627 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15628 when writing a 32767 byte record.
15630 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15632 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15633 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15635 (RSA objects have a reference count access to which is protected
15636 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15637 so they are meant to be shared between threads.)
15638 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15639 "Reddie, Steven" <Steven.Reddie@ca.com>*
15641 * Fix a deadlock in CRYPTO_mem_leaks().
15645 * Use better test patterns in bntest.
15649 * rand_win.c fix for Borland C.
15653 * BN_rshift bugfix for n == 0.
15657 * Add a 'bctest' script that checks for some known 'bc' bugs
15658 so that 'make test' does not abort just because 'bc' is broken.
15662 * Store verify_result within SSL_SESSION also for client side to
15663 avoid potential security hole. (Re-used sessions on the client side
15664 always resulted in verify_result==X509_V_OK, not using the original
15665 result of the server certificate verification.)
15669 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15670 SSL3_RT_APPLICATION_DATA, return 0.
15671 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15676 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15677 releases, have been re-implemented by renaming the previous
15678 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15679 and ssl3_read_internal, respectively, and adding 'peek' parameters
15680 to them. The new ssl[23]_{read,peek} functions are calls to
15681 ssl[23]_read_internal with the 'peek' flag set appropriately.
15682 A 'peek' parameter has also been added to ssl3_read_bytes, which
15683 does the actual work for ssl3_read_internal.
15687 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15688 the method-specific "init()" handler. Also clean up ex_data after
15689 calling the method-specific "finish()" handler. Previously, this was
15690 happening the other way round.
15694 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15695 The previous value, 12, was not always sufficient for BN_mod_exp().
15699 * Make sure that shared libraries get the internal name engine with
15700 the full version number and not just 0. This should mark the
15701 shared libraries as not backward compatible. Of course, this should
15702 be changed again when we can guarantee backward binary compatibility.
15706 * Fix typo in get_cert_by_subject() in by_dir.c
15708 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15710 * Rework the system to generate shared libraries:
15712 - Make note of the expected extension for the shared libraries and
15713 if there is a need for symbolic links from for example libcrypto.so.0
15714 to libcrypto.so.0.9.7. There is extended info in Configure for
15717 - Make as few rebuilds of the shared libraries as possible.
15719 - Still avoid linking the OpenSSL programs with the shared libraries.
15721 - When installing, install the shared libraries separately from the
15726 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15728 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15729 and not in SSL_clear because the latter is also used by the
15730 accept/connect functions; previously, the settings made by
15731 SSL_set_read_ahead would be lost during the handshake.
15733 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15735 * Correct util/mkdef.pl to be selective about disabled algorithms.
15736 Previously, it would create entries for disabled algorithms no
15741 * Added several new manual pages for SSL_* function.
15745 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15747 * In ssl23_get_client_hello, generate an error message when faced
15748 with an initial SSL 3.0/TLS record that is too small to contain the
15749 first two bytes of the ClientHello message, i.e. client_version.
15750 (Note that this is a pathologic case that probably has never happened
15751 in real life.) The previous approach was to use the version number
15752 from the record header as a substitute; but our protocol choice
15753 should not depend on that one because it is not authenticated
15754 by the Finished messages.
15758 * More robust randomness gathering functions for Windows.
15760 *Jeffrey Altman <jaltman@columbia.edu>*
15762 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15763 not set then we don't setup the error code for issuer check errors
15764 to avoid possibly overwriting other errors which the callback does
15765 handle. If an application does set the flag then we assume it knows
15766 what it is doing and can handle the new informational codes
15771 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15772 a general "ANY" type, as such it should be able to decode anything
15773 including tagged types. However it didn't check the class so it would
15774 wrongly interpret tagged types in the same way as their universal
15775 counterpart and unknown types were just rejected. Changed so that the
15776 tagged and unknown types are handled in the same way as a SEQUENCE:
15777 that is the encoding is stored intact. There is also a new type
15778 "V_ASN1_OTHER" which is used when the class is not universal, in this
15779 case we have no idea what the actual type is so we just lump them all
15784 * On VMS, stdout may very well lead to a file that is written to
15785 in a record-oriented fashion. That means that every write() will
15786 write a separate record, which will be read separately by the
15787 programs trying to read from it. This can be very confusing.
15789 The solution is to put a BIO filter in the way that will buffer
15790 text until a linefeed is reached, and then write everything a
15791 line at a time, so every record written will be an actual line,
15792 not chunks of lines and not (usually doesn't happen, but I've
15793 seen it once) several lines in one record. BIO_f_linebuffer() is
15796 Currently, it's a VMS-only method, because that's where it has
15797 been tested well enough.
15801 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15802 it can return incorrect results.
15803 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15804 but it was in 0.9.6-beta[12].)
15808 * Disable the check for content being present when verifying detached
15809 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15810 include zero length content when signing messages.
15814 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15815 BIO_ctrl (for BIO pairs).
15819 * Add DSO method for VMS.
15823 * Bug fix: Montgomery multiplication could produce results with the
15828 * Add RPM specification openssl.spec and modify it to build three
15829 packages. The default package contains applications, application
15830 documentation and run-time libraries. The devel package contains
15831 include files, static libraries and function documentation. The
15832 doc package contains the contents of the doc directory. The original
15833 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15837 * Add a large number of documentation files for many SSL routines.
15839 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15841 * Add a configuration entry for Sony News 4.
15843 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15845 * Don't set the two most significant bits to one when generating a
15846 random number < q in the DSA library.
15850 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15851 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15852 the underlying transport is blocking) if a handshake took place.
15853 (The default behaviour is needed by applications such as s_client
15854 and s_server that use select() to determine when to use SSL_read;
15855 but for applications that know in advance when to expect data, it
15856 just makes things more complicated.)
15860 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15865 * Add a few more EBCDIC conditionals that make `req` and `x509`
15866 work better on such systems.
15868 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15870 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15871 Update PKCS12_parse() so it copies the friendlyName and the
15872 keyid to the certificates aux info.
15876 * Fix bug in PKCS7_verify() which caused an infinite loop
15877 if there was more than one signature.
15879 *Sven Uszpelkat <su@celocom.de>*
15881 * Major change in util/mkdef.pl to include extra information
15882 about each symbol, as well as presenting variables as well
15883 as functions. This change means that there's n more need
15884 to rebuild the .num files when some algorithms are excluded.
15888 * Allow the verify time to be set by an application,
15889 rather than always using the current time.
15893 * Phase 2 verify code reorganisation. The certificate
15894 verify code now looks up an issuer certificate by a
15895 number of criteria: subject name, authority key id
15896 and key usage. It also verifies self signed certificates
15897 by the same criteria. The main comparison function is
15898 X509_check_issued() which performs these checks.
15900 Lot of changes were necessary in order to support this
15901 without completely rewriting the lookup code.
15903 Authority and subject key identifier are now cached.
15905 The LHASH 'certs' is X509_STORE has now been replaced
15906 by a STACK_OF(X509_OBJECT). This is mainly because an
15907 LHASH can't store or retrieve multiple objects with
15908 the same hash value.
15910 As a result various functions (which were all internal
15911 use only) have changed to handle the new X509_STORE
15912 structure. This will break anything that messed round
15913 with X509_STORE internally.
15915 The functions X509_STORE_add_cert() now checks for an
15916 exact match, rather than just subject name.
15918 The X509_STORE API doesn't directly support the retrieval
15919 of multiple certificates matching a given criteria, however
15920 this can be worked round by performing a lookup first
15921 (which will fill the cache with candidate certificates)
15922 and then examining the cache for matches. This is probably
15923 the best we can do without throwing out X509_LOOKUP
15924 entirely (maybe later...).
15926 The X509_VERIFY_CTX structure has been enhanced considerably.
15928 All certificate lookup operations now go via a get_issuer()
15929 callback. Although this currently uses an X509_STORE it
15930 can be replaced by custom lookups. This is a simple way
15931 to bypass the X509_STORE hackery necessary to make this
15932 work and makes it possible to use more efficient techniques
15933 in future. A very simple version which uses a simple
15934 STACK for its trusted certificate store is also provided
15935 using X509_STORE_CTX_trusted_stack().
15937 The verify_cb() and verify() callbacks now have equivalents
15938 in the X509_STORE_CTX structure.
15940 X509_STORE_CTX also has a 'flags' field which can be used
15941 to customise the verify behaviour.
15945 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15946 excludes S/MIME capabilities.
15950 * When a certificate request is read in keep a copy of the
15951 original encoding of the signed data and use it when outputting
15952 again. Signatures then use the original encoding rather than
15953 a decoded, encoded version which may cause problems if the
15954 request is improperly encoded.
15958 * For consistency with other BIO_puts implementations, call
15959 buffer_write(b, ...) directly in buffer_puts instead of calling
15962 In BIO_puts, increment b->num_write as in BIO_write.
15964 *Peter.Sylvester@EdelWeb.fr*
15966 * Fix BN_mul_word for the case where the word is 0. (We have to use
15967 BN_zero, we may not return a BIGNUM with an array consisting of
15968 words set to zero.)
15972 * Avoid calling abort() from within the library when problems are
15973 detected, except if preprocessor symbols have been defined
15974 (such as REF_CHECK, BN_DEBUG etc.).
15978 * New openssl application 'rsautl'. This utility can be
15979 used for low-level RSA operations. DER public key
15980 BIO/fp routines also added.
15984 * New Configure entry and patches for compiling on QNX 4.
15986 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15988 * A demo state-machine implementation was sponsored by
15989 Nuron (<http://www.nuron.com/>) and is now available in
15990 demos/state_machine.
15994 * New options added to the 'dgst' utility for signature
15995 generation and verification.
15999 * Unrecognized PKCS#7 content types are now handled via a
16000 catch all ASN1_TYPE structure. This allows unsupported
16001 types to be stored as a "blob" and an application can
16002 encode and decode it manually.
16006 * Fix various signed/unsigned issues to make a_strex.c
16007 compile under VC++.
16009 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16011 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16012 length if passed a buffer. ASN1_INTEGER_to_BN failed
16013 if passed a NULL BN and its argument was negative.
16015 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16017 * Modification to PKCS#7 encoding routines to output definite
16018 length encoding. Since currently the whole structures are in
16019 memory there's not real point in using indefinite length
16020 constructed encoding. However if OpenSSL is compiled with
16021 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16025 * Added BIO_vprintf() and BIO_vsnprintf().
16029 * Added more prefixes to parse for in the strings written
16030 through a logging bio, to cover all the levels that are available
16031 through syslog. The prefixes are now:
16033 PANIC, EMERG, EMR => LOG_EMERG
16034 ALERT, ALR => LOG_ALERT
16035 CRIT, CRI => LOG_CRIT
16036 ERROR, ERR => LOG_ERR
16037 WARNING, WARN, WAR => LOG_WARNING
16038 NOTICE, NOTE, NOT => LOG_NOTICE
16039 INFO, INF => LOG_INFO
16040 DEBUG, DBG => LOG_DEBUG
16042 and as before, if none of those prefixes are present at the
16043 beginning of the string, LOG_ERR is chosen.
16045 On Win32, the `LOG_*` levels are mapped according to this:
16047 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16048 LOG_WARNING => EVENTLOG_WARNING_TYPE
16049 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16053 * Made it possible to reconfigure with just the configuration
16054 argument "reconf" or "reconfigure". The command line arguments
16055 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16056 and are retrieved from there when reconfiguring.
16062 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16064 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16068 * The obj_dat.pl script was messing up the sorting of object
16069 names. The reason was that it compared the quoted version
16070 of strings as a result "OCSP" > "OCSP Signing" because
16071 " > SPACE. Changed script to store unquoted versions of
16072 names and add quotes on output. It was also omitting some
16073 names from the lookup table if they were given a default
16074 value (that is if SN is missing it is given the same
16075 value as LN and vice versa), these are now added on the
16076 grounds that if an object has a name we should be able to
16077 look it up. Finally added warning output when duplicate
16078 short or long names are found.
16082 * Changes needed for Tandem NSK.
16084 *Scott Uroff <scott@xypro.com>*
16086 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16087 RSA_padding_check_SSLv23(), special padding was never detected
16088 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16089 version rollback attacks was not effective.
16091 In s23_clnt.c, don't use special rollback-attack detection padding
16092 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16093 client; similarly, in s23_srvr.c, don't do the rollback check if
16094 SSL 2.0 is the only protocol enabled in the server.
16098 * Make it possible to get hexdumps of unprintable data with 'openssl
16099 asn1parse'. By implication, the functions ASN1_parse_dump() and
16100 BIO_dump_indent() are added.
16104 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16105 these print out strings and name structures based on various
16106 flags including RFC2253 support and proper handling of
16107 multibyte characters. Added options to the 'x509' utility
16108 to allow the various flags to be set.
16112 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16113 Also change the functions X509_cmp_current_time() and
16114 X509_gmtime_adj() work with an ASN1_TIME structure,
16115 this will enable certificates using GeneralizedTime in validity
16116 dates to be checked.
16120 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16121 negative public key encodings) on by default,
16122 NO_NEG_PUBKEY_BUG can be set to disable it.
16126 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16127 content octets. An i2c_ASN1_OBJECT is unnecessary because
16128 the encoding can be trivially obtained from the structure.
16132 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16133 not read locks (`CRYPTO_r_[un]lock`).
16137 * A first attempt at creating official support for shared
16138 libraries through configuration. I've kept it so the
16139 default is static libraries only, and the OpenSSL programs
16140 are always statically linked for now, but there are
16141 preparations for dynamic linking in place.
16142 This has been tested on Linux and Tru64.
16146 * Randomness polling function for Win9x, as described in:
16147 Peter Gutmann, Software Generation of Practically Strong
16152 * Fix so PRNG is seeded in req if using an already existing
16157 * New options to smime application. -inform and -outform
16158 allow alternative formats for the S/MIME message including
16159 PEM and DER. The -content option allows the content to be
16160 specified separately. This should allow things like Netscape
16161 form signing output easier to verify.
16165 * Fix the ASN1 encoding of tags using the 'long form'.
16169 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16170 STRING types. These convert content octets to and from the
16171 underlying type. The actual tag and length octets are
16172 already assumed to have been read in and checked. These
16173 are needed because all other string types have virtually
16174 identical handling apart from the tag. By having versions
16175 of the ASN1 functions that just operate on content octets
16176 IMPLICIT tagging can be handled properly. It also allows
16177 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16178 and ASN1_INTEGER are identical apart from the tag.
16182 * Change the handling of OID objects as follows:
16184 - New object identifiers are inserted in objects.txt, following
16185 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16186 - objects.pl is used to process obj_mac.num and create a new
16188 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16191 This is currently kind of a hack, and the perl code in objects.pl
16192 isn't very elegant, but it works as I intended. The simplest way
16193 to check that it worked correctly is to look in obj_dat.h and
16194 check the array nid_objs and make sure the objects haven't moved
16195 around (this is important!). Additions are OK, as well as
16196 consistent name changes.
16200 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16204 * Addition of the command line parameter '-rand file' to 'openssl req'.
16205 The given file adds to whatever has already been seeded into the
16206 random pool through the RANDFILE configuration file option or
16207 environment variable, or the default random state file.
16211 * mkstack.pl now sorts each macro group into lexical order.
16212 Previously the output order depended on the order the files
16213 appeared in the directory, resulting in needless rewriting
16218 * Patches to make OpenSSL compile under Win32 again. Mostly
16219 work arounds for the VC++ problem that it treats func() as
16220 func(void). Also stripped out the parts of mkdef.pl that
16221 added extra typesafe functions: these no longer exist.
16225 * Reorganisation of the stack code. The macros are now all
16226 collected in safestack.h . Each macro is defined in terms of
16227 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16228 DEBUG_SAFESTACK is now handled in terms of function casts,
16229 this has the advantage of retaining type safety without the
16230 use of additional functions. If DEBUG_SAFESTACK is not defined
16231 then the non typesafe macros are used instead. Also modified the
16232 mkstack.pl script to handle the new form. Needs testing to see
16233 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16234 the default if no major problems. Similar behaviour for ASN1_SET_OF
16235 and PKCS12_STACK_OF.
16239 * When some versions of IIS use the 'NET' form of private key the
16240 key derivation algorithm is different. Normally MD5(password) is
16241 used as a 128 bit RC4 key. In the modified case
16242 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16243 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16244 as the old Netscape_RSA functions except they have an additional
16245 'sgckey' parameter which uses the modified algorithm. Also added
16246 an -sgckey command line option to the rsa utility. Thanks to
16247 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16248 algorithm to openssl-dev.
16252 * The evp_local.h macros were using 'c.##kname' which resulted in
16253 invalid expansion on some systems (SCO 5.0.5 for example).
16254 Corrected to 'c.kname'.
16256 *Phillip Porch <root@theporch.com>*
16258 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16259 a STACK of email addresses from a certificate or request, these look
16260 in the subject name and the subject alternative name extensions and
16261 omit any duplicate addresses.
16265 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16266 This makes DSA verification about 2 % faster.
16270 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16271 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16272 plus overhead for 1024 bit moduli).
16273 This makes exponentiations about 0.5 % faster for 1024 bit
16274 exponents (as measured by "openssl speed rsa2048").
16278 * Rename memory handling macros to avoid conflicts with other
16280 Malloc => OPENSSL_malloc
16281 Malloc_locked => OPENSSL_malloc_locked
16282 Realloc => OPENSSL_realloc
16283 Free => OPENSSL_free
16287 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16288 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16292 * CygWin32 support.
16294 *John Jarvie <jjarvie@newsguy.com>*
16296 * The type-safe stack code has been rejigged. It is now only compiled
16297 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16298 by default all type-specific stack functions are "#define"d back to
16299 standard stack functions. This results in more streamlined output
16300 but retains the type-safety checking possibilities of the original
16305 * The STACK code has been cleaned up, and certain type declarations
16306 that didn't make a lot of sense have been brought in line. This has
16307 also involved a cleanup of sorts in safestack.h to more correctly
16308 map type-safe stack functions onto their plain stack counterparts.
16309 This work has also resulted in a variety of "const"ifications of
16310 lots of the code, especially `_cmp` operations which should normally
16311 be prototyped with "const" parameters anyway.
16315 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16316 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16317 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16318 where all of 'md' is used each time the PRNG is used, but 'state'
16319 is used only indexed by a cyclic counter. As entropy may not be
16320 well distributed from the beginning, 'md' is important as a
16321 chaining variable. However, the output function chains only half
16322 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16323 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16324 in all of 'state' being rewritten, with the new values depending
16325 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16329 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16330 the handshake is continued after ssl_verify_cert_chain();
16331 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16332 can lead to 'unexplainable' connection aborts later.
16334 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16336 * Major EVP API cipher revision.
16337 Add hooks for extra EVP features. This allows various cipher
16338 parameters to be set in the EVP interface. Support added for variable
16339 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16340 setting of RC2 and RC5 parameters.
16342 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16345 Remove lots of duplicated code from the EVP library. For example *every*
16346 cipher init() function handles the 'iv' in the same way according to the
16347 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16348 for CFB and OFB modes they zero ctx->num.
16350 New functionality allows removal of S/MIME code RC2 hack.
16352 Most of the routines have the same form and so can be declared in terms
16355 By shifting this to the top level EVP_CipherInit() it can be removed from
16356 all individual ciphers. If the cipher wants to handle IVs or keys
16357 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16360 Change lots of functions like EVP_EncryptUpdate() to now return a
16361 value: although software versions of the algorithms cannot fail
16362 any installed hardware versions can.
16366 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16367 this option is set, tolerate broken clients that send the negotiated
16368 protocol version number instead of the requested protocol version
16373 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16374 i.e. non-zero for export ciphersuites, zero otherwise.
16375 Previous versions had this flag inverted, inconsistent with
16376 rsa_tmp_cb (..._TMP_RSA_CB).
16378 *Bodo Moeller; problem reported by Amit Chopra*
16380 * Add missing DSA library text string. Work around for some IIS
16381 key files with invalid SEQUENCE encoding.
16385 * Add a document (doc/standards.txt) that list all kinds of standards
16386 and so on that are implemented in OpenSSL.
16390 * Enhance c_rehash script. Old version would mishandle certificates
16391 with the same subject name hash and wouldn't handle CRLs at all.
16392 Added -fingerprint option to crl utility, to support new c_rehash
16397 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16401 * Fix for SSL server purpose checking. Server checking was
16402 rejecting certificates which had extended key usage present
16403 but no ssl client purpose.
16405 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16407 * Make PKCS#12 code work with no password. The PKCS#12 spec
16408 is a little unclear about how a blank password is handled.
16409 Since the password in encoded as a BMPString with terminating
16410 double NULL a zero length password would end up as just the
16411 double NULL. However no password at all is different and is
16412 handled differently in the PKCS#12 key generation code. NS
16413 treats a blank password as zero length. MSIE treats it as no
16414 password on export: but it will try both on import. We now do
16415 the same: PKCS12_parse() tries zero length and no password if
16416 the password is set to "" or NULL (NULL is now a valid password:
16417 it wasn't before) as does the pkcs12 application.
16421 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16422 perror when PEM_read_bio_X509_REQ fails, the error message must
16423 be obtained from the error queue.
16427 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16428 it in ERR_remove_state if appropriate, and change ERR_get_state
16429 accordingly to avoid race conditions (this is necessary because
16430 thread_hash is no longer constant once set).
16434 * Bugfix for linux-elf makefile.one.
16438 * RSA_get_default_method() will now cause a default
16439 RSA_METHOD to be chosen if one doesn't exist already.
16440 Previously this was only set during a call to RSA_new()
16441 or RSA_new_method(NULL) meaning it was possible for
16442 RSA_get_default_method() to return NULL.
16446 * Added native name translation to the existing DSO code
16447 that will convert (if the flag to do so is set) filenames
16448 that are sufficiently small and have no path information
16449 into a canonical native form. Eg. "blah" converted to
16450 "libblah.so" or "blah.dll" etc.
16454 * New function ERR_error_string_n(e, buf, len) which is like
16455 ERR_error_string(e, buf), but writes at most 'len' bytes
16456 including the 0 terminator. For ERR_error_string_n, 'buf'
16459 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16461 * CONF library reworked to become more general. A new CONF
16462 configuration file reader "class" is implemented as well as a
16463 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16464 old `CONF_*` functions are still there, but are reimplemented to
16465 work in terms of the new functions. Also, a set of functions
16466 to handle the internal storage of the configuration data is
16467 provided to make it easier to write new configuration file
16468 reader "classes" (I can definitely see something reading a
16469 configuration file in XML format, for example), called `_CONF_*`,
16470 or "the configuration storage API"...
16472 The new configuration file reading functions are:
16474 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16475 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16477 NCONF_default, NCONF_WIN32
16479 NCONF_dump_fp, NCONF_dump_bio
16481 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16482 NCONF_new creates a new CONF object. This works in the same way
16483 as other interfaces in OpenSSL, like the BIO interface.
16484 `NCONF_dump_*` dump the internal storage of the configuration file,
16485 which is useful for debugging. All other functions take the same
16486 arguments as the old `CONF_*` functions with the exception of the
16487 first that must be a `CONF *` instead of a `LHASH *`.
16489 To make it easier to use the new classes with the old `CONF_*` functions,
16490 the function CONF_set_default_method is provided.
16494 * Add '-tls1' option to 'openssl ciphers', which was already
16495 mentioned in the documentation but had not been implemented.
16496 (This option is not yet really useful because even the additional
16497 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16501 * Initial DSO code added into libcrypto for letting OpenSSL (and
16502 OpenSSL-based applications) load shared libraries and bind to
16503 them in a portable way.
16505 *Geoff Thorpe, with contributions from Richard Levitte*
16507 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16509 * Make sure _lrotl and _lrotr are only used with MSVC.
16511 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16512 (the default implementation of RAND_status).
16514 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16515 to '-clrext' (= clear extensions), as intended and documented.
16516 *Bodo Moeller; inconsistency pointed out by Michael Attili
16517 <attili@amaxo.com>*
16519 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16520 was larger than the MD block size.
16522 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16524 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16525 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16526 using the passed key: if the passed key was a private key the result
16527 of X509_print(), for example, would be to print out all the private key
16532 * des_quad_cksum() byte order bug fix.
16533 *Ulf Möller, using the problem description in krb4-0.9.7, where
16534 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16536 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16539 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16541 * For easily testing in shell scripts whether some command
16542 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16543 returns with exit code 0 iff no command of the given name is available.
16544 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16545 the output goes to stdout and nothing is printed to stderr.
16546 Additional arguments are always ignored.
16548 Since for each cipher there is a command of the same name,
16549 the 'no-cipher' compilation switches can be tested this way.
16551 ('openssl no-XXX' is not able to detect pseudo-commands such
16552 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16556 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16560 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16561 is set; it will be thrown away anyway because each handshake creates
16563 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16564 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16565 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16566 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16570 * New s_client option -ign_eof: EOF at stdin is ignored, and
16571 'Q' and 'R' lose their special meanings (quit/renegotiate).
16572 This is part of what -quiet does; unlike -quiet, -ign_eof
16573 does not suppress any output.
16577 * Add compatibility options to the purpose and trust code. The
16578 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16579 accepts a certificate or CA, this was the previous behaviour,
16580 with all the associated security issues.
16582 X509_TRUST_COMPAT is the old trust behaviour: only and
16583 automatically trust self signed roots in certificate store. A
16584 new trust setting X509_TRUST_DEFAULT is used to specify that
16585 a purpose has no associated trust setting and it should instead
16586 use the value in the default purpose.
16590 * Fix the PKCS#8 DSA private key code so it decodes keys again
16591 and fix a memory leak.
16595 * In util/mkerr.pl (which implements 'make errors'), preserve
16596 reason strings from the previous version of the .c file, as
16597 the default to have only downcase letters (and digits) in
16598 automatically generated reasons codes is not always appropriate.
16602 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16603 using strerror. Previously, ERR_reason_error_string() returned
16604 library names as reason strings for SYSerr; but SYSerr is a special
16605 case where small numbers are errno values, not library numbers.
16609 * Add '-dsaparam' option to 'openssl dhparam' application. This
16610 converts DSA parameters into DH parameters. (When creating parameters,
16611 DSA_generate_parameters is used.)
16615 * Include 'length' (recommended exponent length) in C code generated
16616 by 'openssl dhparam -C'.
16620 * The second argument to set_label in perlasm was already being used
16621 so couldn't be used as a "file scope" flag. Moved to third argument
16626 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16627 instead of RAND_bytes for encryption IVs and salts.
16631 * Include RAND_status() into RAND_METHOD instead of implementing
16632 it only for md_rand.c Otherwise replacing the PRNG by calling
16633 RAND_set_rand_method would be impossible.
16637 * Don't let DSA_generate_key() enter an infinite loop if the random
16638 number generation fails.
16642 * New 'rand' application for creating pseudo-random output.
16646 * Added configuration support for Linux/IA64
16648 *Rolf Haberrecker <rolf@suse.de>*
16650 * Assembler module support for Mingw32.
16654 * Shared library support for HPUX (in shlib/).
16656 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16658 * Shared library support for Solaris gcc.
16660 *Lutz Behnke <behnke@trustcenter.de>*
16662 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16664 * PKCS7_encrypt() was adding text MIME headers twice because they
16665 were added manually and by SMIME_crlf_copy().
16669 * In bntest.c don't call BN_rand with zero bits argument.
16671 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16673 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16674 case was implemented. This caused BN_div_recp() to fail occasionally.
16678 * Add an optional second argument to the set_label() in the perl
16679 assembly language builder. If this argument exists and is set
16680 to 1 it signals that the assembler should use a symbol whose
16681 scope is the entire file, not just the current function. This
16682 is needed with MASM which uses the format label:: for this scope.
16684 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16686 * Change the ASN1 types so they are typedefs by default. Before
16687 almost all types were #define'd to ASN1_STRING which was causing
16688 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16693 * Change names of new functions to the new get1/get0 naming
16694 convention: After 'get1', the caller owns a reference count
16695 and has to call `..._free`; 'get0' returns a pointer to some
16696 data structure without incrementing reference counters.
16697 (Some of the existing 'get' functions increment a reference
16698 counter, some don't.)
16699 Similarly, 'set1' and 'add1' functions increase reference
16700 counters or duplicate objects.
16704 * Allow for the possibility of temp RSA key generation failure:
16705 the code used to assume it always worked and crashed on failure.
16709 * Fix potential buffer overrun problem in BIO_printf().
16710 *Ulf Möller, using public domain code by Patrick Powell; problem
16711 pointed out by David Sacerdote <das33@cornell.edu>*
16713 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16714 RAND_egd() and RAND_status(). In the command line application,
16715 the EGD socket can be specified like a seed file using RANDFILE
16720 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16721 Some CAs (e.g. Verisign) distribute certificates in this form.
16725 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16726 list to exclude them. This means that no special compilation option
16727 is needed to use anonymous DH: it just needs to be included in the
16732 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16733 EVP_MD_type. The old functionality is available in a new macro called
16734 EVP_MD_md(). Change code that uses it and update docs.
16738 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16739 where the `void *` argument is replaced by a function pointer argument.
16740 Previously `void *` was abused to point to functions, which works on
16741 many platforms, but is not correct. As these functions are usually
16742 called by macros defined in OpenSSL header files, most source code
16743 should work without changes.
16747 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16748 sections with information on -D... compiler switches used for
16749 compiling the library so that applications can see them. To enable
16750 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16751 must be defined. E.g.,
16752 #define OPENSSL_ALGORITHM_DEFINES
16753 #include <openssl/opensslconf.h>
16754 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16756 *Richard Levitte, Ulf and Bodo Möller*
16758 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16763 * Change the 'other' type in certificate aux info to a STACK_OF
16764 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16765 the required ASN1 format: arbitrary types determined by an OID.
16769 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16770 argument to 'req'. This is not because the function is newer or
16771 better than others it just uses the work 'NEW' in the certificate
16772 request header lines. Some software needs this.
16776 * Reorganise password command line arguments: now passwords can be
16777 obtained from various sources. Delete the PEM_cb function and make
16778 it the default behaviour: i.e. if the callback is NULL and the
16779 usrdata argument is not NULL interpret it as a null terminated pass
16780 phrase. If usrdata and the callback are NULL then the pass phrase
16781 is prompted for as usual.
16785 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16786 the support is automatically enabled. The resulting binaries will
16787 autodetect the card and use it if present.
16789 *Ben Laurie and Compaq Inc.*
16791 * Work around for Netscape hang bug. This sends certificate request
16792 and server done in one record. Since this is perfectly legal in the
16793 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16794 the bugs/SSLv3 entry for more info.
16798 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16802 * Add -rand argument to smime and pkcs12 applications and read/write
16807 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16811 * Add command line password options to the remaining applications.
16815 * Bug fix for BN_div_recp() for numerators with an even number of
16820 * More tests in bntest.c, and changed test_bn output.
16824 * ./config recognizes MacOS X now.
16828 * Bug fix for BN_div() when the first words of num and divisor are
16829 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16833 * Add support for various broken PKCS#8 formats, and command line
16834 options to produce them.
16838 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16839 get temporary BIGNUMs from a BN_CTX.
16843 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16848 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16849 include a #define from the old name to the new. The original intent
16850 was that statically linked binaries could for example just call
16851 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16852 link with digests. This never worked because SSLeay_add_all_digests()
16853 and SSLeay_add_all_ciphers() were in the same source file so calling
16854 one would link with the other. They are now in separate source files.
16858 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16862 * Use a less unusual form of the Miller-Rabin primality test (it used
16863 a binary algorithm for exponentiation integrated into the Miller-Rabin
16864 loop, our standard modexp algorithms are faster).
16868 * Support for the EBCDIC character set completed.
16870 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16872 * Source code cleanups: use const where appropriate, eliminate casts,
16873 use `void *` instead of `char *` in lhash.
16877 * Bugfix: ssl3_send_server_key_exchange was not restartable
16878 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16879 this the server could overwrite ephemeral keys that the client
16884 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16885 using 50 iterations of the Rabin-Miller test.
16887 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16888 iterations of the Rabin-Miller test as required by the appendix
16889 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16890 As BN_is_prime_fasttest includes trial division, DSA parameter
16891 generation becomes much faster.
16893 This implies a change for the callback functions in DSA_is_prime
16894 and DSA_generate_parameters: The callback function is called once
16895 for each positive witness in the Rabin-Miller test, not just
16896 occasionally in the inner loop; and the parameters to the
16897 callback function now provide an iteration count for the outer
16898 loop rather than for the current invocation of the inner loop.
16899 DSA_generate_parameters additionally can call the callback
16900 function with an 'iteration count' of -1, meaning that a
16901 candidate has passed the trial division test (when q is generated
16902 from an application-provided seed, trial division is skipped).
16906 * New function BN_is_prime_fasttest that optionally does trial
16907 division before starting the Rabin-Miller test and has
16908 an additional BN_CTX * argument (whereas BN_is_prime always
16909 has to allocate at least one BN_CTX).
16910 'callback(1, -1, cb_arg)' is called when a number has passed the
16911 trial division stage.
16915 * Fix for bug in CRL encoding. The validity dates weren't being handled
16920 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16924 * New function BN_pseudo_rand().
16928 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16929 bignum version of BN_from_montgomery() with the working code from
16930 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16935 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16936 made it impossible to use the same SSL_SESSION data structure in
16937 SSL2 clients in multiple threads.
16941 * The return value of RAND_load_file() no longer counts bytes obtained
16942 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16943 to seed the PRNG (previously an explicit byte count was required).
16945 *Ulf Möller, Bodo Möller*
16947 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16948 used `char *` instead of `void *` and had casts all over the place.
16952 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16956 * Retain source code compatibility for BN_prime_checks macro:
16957 BN_is_prime(..., BN_prime_checks, ...) now uses
16958 BN_prime_checks_for_size to determine the appropriate number of
16959 Rabin-Miller iterations.
16963 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16964 DH_CHECK_P_NOT_SAFE_PRIME.
16965 (Check if this is true? OpenPGP calls them "strong".)
16969 * Merge the functionality of "dh" and "gendh" programs into a new program
16970 "dhparam". The old programs are retained for now but will handle DH keys
16971 (instead of parameters) in future.
16975 * Make the ciphers, s_server and s_client programs check the return values
16976 when a new cipher list is set.
16980 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16981 ciphers. Before when the 56bit ciphers were enabled the sorting was
16984 The syntax for the cipher sorting has been extended to support sorting by
16985 cipher-strength (using the strength_bits hard coded in the tables).
16986 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16988 Fix a bug in the cipher-command parser: when supplying a cipher command
16989 string with an "undefined" symbol (neither command nor alphanumeric
16990 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16991 an error is flagged.
16993 Due to the strength-sorting extension, the code of the
16994 ssl_create_cipher_list() function was completely rearranged. I hope that
16995 the readability was also increased :-)
16997 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16999 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17000 for the first serial number and places 2 in the serial number file. This
17001 avoids problems when the root CA is created with serial number zero and
17002 the first user certificate has the same issuer name and serial number
17007 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17008 the new code. Add documentation for this stuff.
17012 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17013 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17014 structures and behave in an analogous way to the X509v3 functions:
17015 they shouldn't be called directly but wrapper functions should be used
17018 So we also now have some wrapper functions that call the X509at functions
17019 when passed certificate requests. (TO DO: similar things can be done with
17020 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17021 things. Some of these need some d2i or i2d and print functionality
17022 because they handle more complex structures.)
17026 * Add missing #ifndefs that caused missing symbols when building libssl
17027 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17028 NO_RSA in `ssl/s2*.c`.
17030 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17032 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17033 has a return value which indicates the quality of the random data
17034 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17035 error queue. New function RAND_pseudo_bytes() generates output that is
17036 guaranteed to be unique but not unpredictable. RAND_add is like
17037 RAND_seed, but takes an extra argument for an entropy estimate
17038 (RAND_seed always assumes full entropy).
17042 * Do more iterations of Rabin-Miller probable prime test (specifically,
17043 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17044 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17045 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17046 false-positive rate of at most 2^-80 for random input.
17050 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17054 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17055 in the 0.9.5 release), this returns the chain
17056 from an X509_CTX structure with a dup of the stack and all
17057 the X509 reference counts upped: so the stack will exist
17058 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17061 Also make SSL_SESSION_print() print out the verify return
17066 * Add manpage for the pkcs12 command. Also change the default
17067 behaviour so MAC iteration counts are used unless the new
17068 -nomaciter option is used. This improves file security and
17069 only older versions of MSIE (4.0 for example) need it.
17073 * Honor the no-xxx Configure options when creating .DEF files.
17077 * Add PKCS#10 attributes to field table: challengePassword,
17078 unstructuredName and unstructuredAddress. These are taken from
17079 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17080 international characters are used.
17082 More changes to X509_ATTRIBUTE code: allow the setting of types
17083 based on strings. Remove the 'loc' parameter when adding
17084 attributes because these will be a SET OF encoding which is sorted
17089 * Initial changes to the 'req' utility to allow request generation
17090 automation. This will allow an application to just generate a template
17091 file containing all the field values and have req construct the
17094 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17095 used all over the place including certificate requests and PKCS#7
17096 structures. They are currently handled manually where necessary with
17097 some primitive wrappers for PKCS#7. The new functions behave in a
17098 manner analogous to the X509 extension functions: they allow
17099 attributes to be looked up by NID and added.
17101 Later something similar to the X509V3 code would be desirable to
17102 automatically handle the encoding, decoding and printing of the
17103 more complex types. The string types like challengePassword can
17104 be handled by the string table functions.
17106 Also modified the multi byte string table handling. Now there is
17107 a 'global mask' which masks out certain types. The table itself
17108 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17109 is useful when for example there is only one permissible type
17110 (as in countryName) and using the mask might result in no valid
17115 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17116 SSL_get_peer_finished to allow applications to obtain the latest
17117 Finished messages sent to the peer or expected from the peer,
17118 respectively. (SSL_get_peer_finished is usually the Finished message
17119 actually received from the peer, otherwise the protocol will be aborted.)
17121 As the Finished message are message digests of the complete handshake
17122 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17123 be used for external authentication procedures when the authentication
17124 provided by SSL/TLS is not desired or is not enough.
17128 * Enhanced support for Alpha Linux is added. Now ./config checks if
17129 the host supports BWX extension and if Compaq C is present on the
17130 $PATH. Just exploiting of the BWX extension results in 20-30%
17131 performance kick for some algorithms, e.g. DES and RC4 to mention
17132 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17137 * Add support for MS "fast SGC". This is arguably a violation of the
17138 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17139 weak crypto and after checking the certificate is SGC a second one
17140 with strong crypto. MS SGC stops the first handshake after receiving
17141 the server certificate message and sends a second client hello. Since
17142 a server will typically do all the time consuming operations before
17143 expecting any further messages from the client (server key exchange
17144 is the most expensive) there is little difference between the two.
17146 To get OpenSSL to support MS SGC we have to permit a second client
17147 hello message after we have sent server done. In addition we have to
17148 reset the MAC if we do get this second client hello.
17152 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17153 if a DER encoded private key is RSA or DSA traditional format. Changed
17154 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17155 format DER encoded private key. Newer code should use PKCS#8 format which
17156 has the key type encoded in the ASN1 structure. Added DER private key
17157 support to pkcs8 application.
17161 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17162 ciphersuites has been selected (as required by the SSL 3/TLS 1
17163 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17164 is set, we interpret this as a request to violate the specification
17165 (the worst that can happen is a handshake failure, and 'correct'
17166 behaviour would result in a handshake failure anyway).
17170 * In SSL_CTX_add_session, take into account that there might be multiple
17171 SSL_SESSION structures with the same session ID (e.g. when two threads
17172 concurrently obtain them from an external cache).
17173 The internal cache can handle only one SSL_SESSION with a given ID,
17174 so if there's a conflict, we now throw out the old one to achieve
17179 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17180 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17181 some routines that use cipher OIDs: some ciphers do not have OIDs
17182 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17187 * Simplify the trust setting structure and code. Now we just have
17188 two sequences of OIDs for trusted and rejected settings. These will
17189 typically have values the same as the extended key usage extension
17190 and any application specific purposes.
17192 The trust checking code now has a default behaviour: it will just
17193 check for an object with the same NID as the passed id. Functions can
17194 be provided to override either the default behaviour or the behaviour
17195 for a given id. SSL client, server and email already have functions
17196 in place for compatibility: they check the NID and also return "trusted"
17197 if the certificate is self signed.
17201 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17202 traditional format into an EVP_PKEY structure.
17206 * Add a password callback function PEM_cb() which either prompts for
17207 a password if usr_data is NULL or otherwise assumes it is a null
17208 terminated password. Allow passwords to be passed on command line
17209 environment or config files in a few more utilities.
17213 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17214 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17215 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17216 Update documentation.
17220 * Support for ASN1 "NULL" type. This could be handled before by using
17221 ASN1_TYPE but there wasn't any function that would try to read a NULL
17222 and produce an error if it couldn't. For compatibility we also have
17223 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17224 don't allocate anything because they don't need to.
17228 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17231 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17233 * Rebuild of the memory allocation routines used by OpenSSL code and
17234 possibly others as well. The purpose is to make an interface that
17235 provide hooks so anyone can build a separate set of allocation and
17236 deallocation routines to be used by OpenSSL, for example memory
17237 pool implementations, or something else, which was previously hard
17238 since Malloc(), Realloc() and Free() were defined as macros having
17239 the values malloc, realloc and free, respectively (except for Win32
17240 compilations). The same is provided for memory debugging code.
17241 OpenSSL already comes with functionality to find memory leaks, but
17242 this gives people a chance to debug other memory problems.
17244 With these changes, a new set of functions and macros have appeared:
17246 CRYPTO_set_mem_debug_functions() [F]
17247 CRYPTO_get_mem_debug_functions() [F]
17248 CRYPTO_dbg_set_options() [F]
17249 CRYPTO_dbg_get_options() [F]
17250 CRYPTO_malloc_debug_init() [M]
17252 The memory debug functions are NULL by default, unless the library
17253 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17254 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17255 gives the standard debugging functions that come with OpenSSL) or
17256 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17257 provided by the library user) must be used. When the standard
17258 debugging functions are used, CRYPTO_dbg_set_options can be used to
17259 request additional information:
17260 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17261 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17263 Also, things like CRYPTO_set_mem_functions will always give the
17264 expected result (the new set of functions is used for allocation
17265 and deallocation) at all times, regardless of platform and compiler
17268 To finish it up, some functions that were never use in any other
17269 way than through macros have a new API and new semantic:
17271 CRYPTO_dbg_malloc()
17272 CRYPTO_dbg_realloc()
17275 All macros of value have retained their old syntax.
17277 *Richard Levitte and Bodo Moeller*
17279 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17280 ordering of SMIMECapabilities wasn't in "strength order" and there
17281 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17286 * Some ASN1 types with illegal zero length encoding (INTEGER,
17287 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17289 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17291 * Merge in my S/MIME library for OpenSSL. This provides a simple
17292 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17293 functionality to handle multipart/signed properly) and a utility
17294 called 'smime' to call all this stuff. This is based on code I
17295 originally wrote for Celo who have kindly allowed it to be
17296 included in OpenSSL.
17300 * Add variants des_set_key_checked and des_set_key_unchecked of
17301 des_set_key (aka des_key_sched). Global variable des_check_key
17302 decides which of these is called by des_set_key; this way
17303 des_check_key behaves as it always did, but applications and
17304 the library itself, which was buggy for des_check_key == 1,
17305 have a cleaner way to pick the version they need.
17309 * New function PKCS12_newpass() which changes the password of a
17314 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17315 dynamic mix. In both cases the ids can be used as an index into the
17316 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17317 functions so they accept a list of the field values and the
17318 application doesn't need to directly manipulate the X509_TRUST
17323 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17328 * Modify the way the V3 extension code looks up extensions. This now
17329 works in a similar way to the object code: we have some "standard"
17330 extensions in a static table which is searched with OBJ_bsearch()
17331 and the application can add dynamic ones if needed. The file
17332 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17333 updated whenever a new extension is added to the core code and kept
17334 in ext_nid order. There is a simple program 'tabtest.c' which checks
17335 this. New extensions are not added too often so this file can readily
17336 be maintained manually.
17338 There are two big advantages in doing things this way. The extensions
17339 can be looked up immediately and no longer need to be "added" using
17340 X509V3_add_standard_extensions(): this function now does nothing.
17341 Side note: I get *lots* of email saying the extension code doesn't
17342 work because people forget to call this function.
17343 Also no dynamic allocation is done unless new extensions are added:
17344 so if we don't add custom extensions there is no need to call
17345 X509V3_EXT_cleanup().
17349 * Modify enc utility's salting as follows: make salting the default. Add a
17350 magic header, so unsalted files fail gracefully instead of just decrypting
17351 to garbage. This is because not salting is a big security hole, so people
17352 should be discouraged from doing it.
17356 * Fixes and enhancements to the 'x509' utility. It allowed a message
17357 digest to be passed on the command line but it only used this
17358 parameter when signing a certificate. Modified so all relevant
17359 operations are affected by the digest parameter including the
17360 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17361 DSA key was used because it didn't fix the digest.
17365 * Initial certificate chain verify code. Currently tests the untrusted
17366 certificates for consistency with the verify purpose (which is set
17367 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17369 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17370 this is because it will reject chains with invalid extensions whereas
17371 every previous version of OpenSSL and SSLeay made no checks at all.
17373 Trust code: checks the root CA for the relevant trust settings. Trust
17374 settings have an initial value consistent with the verify purpose: e.g.
17375 if the verify purpose is for SSL client use it expects the CA to be
17376 trusted for SSL client use. However the default value can be changed to
17377 permit custom trust settings: one example of this would be to only trust
17378 certificates from a specific "secure" set of CAs.
17380 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17381 which should be used for version portability: especially since the
17382 verify structure is likely to change more often now.
17384 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17385 to set them. If not set then assume SSL clients will verify SSL servers
17388 Two new options to the verify program: -untrusted allows a set of
17389 untrusted certificates to be passed in and -purpose which sets the
17390 intended purpose of the certificate. If a purpose is set then the
17391 new chain verify code is used to check extension consistency.
17395 * Support for the authority information access extension.
17399 * Modify RSA and DSA PEM read routines to transparently handle
17400 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17401 public keys in a format compatible with certificate
17402 SubjectPublicKeyInfo structures. Unfortunately there were already
17403 functions called *_PublicKey_* which used various odd formats so
17404 these are retained for compatibility: however the DSA variants were
17405 never in a public release so they have been deleted. Changed dsa/rsa
17406 utilities to handle the new format: note no releases ever handled public
17407 keys so we should be OK.
17409 The primary motivation for this change is to avoid the same fiasco
17410 that dogs private keys: there are several incompatible private key
17411 formats some of which are standard and some OpenSSL specific and
17412 require various evil hacks to allow partial transparent handling and
17413 even then it doesn't work with DER formats. Given the option anything
17414 other than PKCS#8 should be dumped: but the other formats have to
17415 stay in the name of compatibility.
17417 With public keys and the benefit of hindsight one standard format
17418 is used which works with EVP_PKEY, RSA or DSA structures: though
17419 it clearly returns an error if you try to read the wrong kind of key.
17421 Added a -pubkey option to the 'x509' utility to output the public key.
17422 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17423 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17424 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17425 that do the same as the `EVP_PKEY_assign_*()` except they up the
17426 reference count of the added key (they don't "swallow" the
17431 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17432 CRLs would fail if the file contained no certificates or no CRLs:
17433 added a new function to read in both types and return the number
17434 read: this means that if none are read it will be an error. The
17435 DER versions of the certificate and CRL reader would always fail
17436 because it isn't possible to mix certificates and CRLs in DER format
17437 without choking one or the other routine. Changed this to just read
17438 a certificate: this is the best we can do. Also modified the code
17439 in `apps/verify.c` to take notice of return codes: it was previously
17440 attempting to read in certificates from NULL pointers and ignoring
17441 any errors: this is one reason why the cert and CRL reader seemed
17442 to work. It doesn't check return codes from the default certificate
17443 routines: these may well fail if the certificates aren't installed.
17447 * Code to support otherName option in GeneralName.
17451 * First update to verify code. Change the verify utility
17452 so it warns if it is passed a self signed certificate:
17453 for consistency with the normal behaviour. X509_verify
17454 has been modified to it will now verify a self signed
17455 certificate if *exactly* the same certificate appears
17456 in the store: it was previously impossible to trust a
17457 single self signed certificate. This means that:
17458 openssl verify ss.pem
17459 now gives a warning about a self signed certificate but
17460 openssl verify -CAfile ss.pem ss.pem
17465 * For servers, store verify_result in SSL_SESSION data structure
17466 (and add it to external session representation).
17467 This is needed when client certificate verifications fails,
17468 but an application-provided verification callback (set by
17469 SSL_CTX_set_cert_verify_callback) allows accepting the session
17470 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17471 but returns 1): When the session is reused, we have to set
17472 ssl->verify_result to the appropriate error code to avoid
17475 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17477 * Fix a bug in the new PKCS#7 code: it didn't consider the
17478 case in PKCS7_dataInit() where the signed PKCS7 structure
17479 didn't contain any existing data because it was being created.
17481 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17483 * Add a salt to the key derivation routines in enc.c. This
17484 forms the first 8 bytes of the encrypted file. Also add a
17485 -S option to allow a salt to be input on the command line.
17489 * New function X509_cmp(). Oddly enough there wasn't a function
17490 to compare two certificates. We do this by working out the SHA1
17491 hash and comparing that. X509_cmp() will be needed by the trust
17496 * SSL_get1_session() is like SSL_get_session(), but increments
17497 the reference count in the SSL_SESSION returned.
17499 *Geoff Thorpe <geoff@eu.c2.net>*
17501 * Fix for 'req': it was adding a null to request attributes.
17502 Also change the X509_LOOKUP and X509_INFO code to handle
17503 certificate auxiliary information.
17507 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17512 * Add the possibility to add extra information to the memory leak
17513 detecting output, to form tracebacks, showing from where each
17514 allocation was originated: CRYPTO_push_info("constant string") adds
17515 the string plus current file name and line number to a per-thread
17516 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17517 is like calling CYRPTO_pop_info() until the stack is empty.
17518 Also updated memory leak detection code to be multi-thread-safe.
17522 * Add options -text and -noout to pkcs7 utility and delete the
17523 encryption options which never did anything. Update docs.
17527 * Add options to some of the utilities to allow the pass phrase
17528 to be included on either the command line (not recommended on
17529 OSes like Unix) or read from the environment. Update the
17530 manpages and fix a few bugs.
17534 * Add a few manpages for some of the openssl commands.
17538 * Fix the -revoke option in ca. It was freeing up memory twice,
17539 leaking and not finding already revoked certificates.
17543 * Extensive changes to support certificate auxiliary information.
17544 This involves the use of X509_CERT_AUX structure and X509_AUX
17545 functions. An X509_AUX function such as PEM_read_X509_AUX()
17546 can still read in a certificate file in the usual way but it
17547 will also read in any additional "auxiliary information". By
17548 doing things this way a fair degree of compatibility can be
17549 retained: existing certificates can have this information added
17550 using the new 'x509' options.
17552 Current auxiliary information includes an "alias" and some trust
17553 settings. The trust settings will ultimately be used in enhanced
17554 certificate chain verification routines: currently a certificate
17555 can only be trusted if it is self signed and then it is trusted
17560 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17561 The problem was that one of the replacement routines had not been working
17562 since SSLeay releases. For now the offending routine has been replaced
17563 with non-optimised assembler. Even so, this now gives around 95%
17564 performance improvement for 1024 bit RSA signs.
17568 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17569 handling. Most clients have the effective key size in bits equal to
17570 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17571 A few however don't do this and instead use the size of the decrypted key
17572 to determine the RC2 key length and the AlgorithmIdentifier to determine
17573 the effective key length. In this case the effective key length can still
17574 be 40 bits but the key length can be 168 bits for example. This is fixed
17575 by manually forcing an RC2 key into the EVP_PKEY structure because the
17576 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17577 the key length and effective key length are equal.
17581 * Add a bunch of functions that should simplify the creation of
17582 X509_NAME structures. Now you should be able to do:
17583 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17584 and have it automatically work out the correct field type and fill in
17585 the structures. The more adventurous can try:
17586 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17587 and it will (hopefully) work out the correct multibyte encoding.
17591 * Change the 'req' utility to use the new field handling and multibyte
17592 copy routines. Before the DN field creation was handled in an ad hoc
17593 way in req, ca, and x509 which was rather broken and didn't support
17594 BMPStrings or UTF8Strings. Since some software doesn't implement
17595 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17596 using the dirstring_type option. See the new comment in the default
17597 openssl.cnf for more info.
17601 * Make crypto/rand/md_rand.c more robust:
17602 - Assure unique random numbers after fork().
17603 - Make sure that concurrent threads access the global counter and
17604 md serializably so that we never lose entropy in them
17605 or use exactly the same state in multiple threads.
17606 Access to the large state is not always serializable because
17607 the additional locking could be a performance killer, and
17608 md should be large enough anyway.
17612 * New file `apps/app_rand.c` with commonly needed functionality
17613 for handling the random seed file.
17615 Use the random seed file in some applications that previously did not:
17617 dsaparam -genkey (which also ignored its '-rand' option),
17620 x509 (when signing).
17621 Except on systems with /dev/urandom, it is crucial to have a random
17622 seed file at least for key creation, DSA signing, and for DH exchanges;
17623 for RSA signatures we could do without one.
17625 gendh and gendsa (unlike genrsa) used to read only the first byte
17626 of each file listed in the '-rand' option. The function as previously
17627 found in genrsa is now in app_rand.c and is used by all programs
17628 that support '-rand'.
17632 * In RAND_write_file, use mode 0600 for creating files;
17633 don't just chmod when it may be too late.
17637 * Report an error from X509_STORE_load_locations
17638 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17642 * New function ASN1_mbstring_copy() this copies a string in either
17643 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17644 into an ASN1_STRING type. A mask of permissible types is passed
17645 and it chooses the "minimal" type to use or an error if not type
17650 * Add function equivalents to the various macros in asn1.h. The old
17651 macros are retained with an `M_` prefix. Code inside the library can
17652 use the `M_` macros. External code (including the openssl utility)
17653 should *NOT* in order to be "shared library friendly".
17657 * Add various functions that can check a certificate's extensions
17658 to see if it usable for various purposes such as SSL client,
17659 server or S/MIME and CAs of these types. This is currently
17660 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17661 verification. Also added a -purpose flag to x509 utility to
17662 print out all the purposes.
17666 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17671 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17672 for, obtain and decode and extension and obtain its critical flag.
17673 This allows all the necessary extension code to be handled in a
17674 single function call.
17678 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17679 platforms. See crypto/rc4/rc4_enc.c for further details.
17683 * New -noout option to asn1parse. This causes no output to be produced
17684 its main use is when combined with -strparse and -out to extract data
17685 from a file (which may not be in ASN.1 format).
17689 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17690 when producing the local key id.
17692 *Richard Levitte <levitte@stacken.kth.se>*
17694 * New option -dhparam in s_server. This allows a DH parameter file to be
17695 stated explicitly. If it is not stated then it tries the first server
17696 certificate file. The previous behaviour hard coded the filename
17701 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17702 a public key to be input or output. For example:
17703 openssl rsa -in key.pem -pubout -out pubkey.pem
17704 Also added necessary DSA public key functions to handle this.
17708 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17709 in the message. This was handled by allowing
17710 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17712 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17714 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17715 to the end of the strings whereas this didn't. This would cause problems
17716 if strings read with d2i_ASN1_bytes() were later modified.
17718 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17720 * Fix for base64 decode bug. When a base64 bio reads only one line of
17721 data and it contains EOF it will end up returning an error. This is
17722 caused by input 46 bytes long. The cause is due to the way base64
17723 BIOs find the start of base64 encoded data. They do this by trying a
17724 trial decode on each line until they find one that works. When they
17725 do a flag is set and it starts again knowing it can pass all the
17726 data directly through the decoder. Unfortunately it doesn't reset
17727 the context it uses. This means that if EOF is reached an attempt
17728 is made to pass two EOFs through the context and this causes the
17729 resulting error. This can also cause other problems as well. As is
17730 usual with these problems it takes *ages* to find and the fix is
17731 trivial: move one line.
17733 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17735 * Ugly workaround to get s_client and s_server working under Windows. The
17736 old code wouldn't work because it needed to select() on sockets and the
17737 tty (for keypresses and to see if data could be written). Win32 only
17738 supports select() on sockets so we select() with a 1s timeout on the
17739 sockets and then see if any characters are waiting to be read, if none
17740 are present then we retry, we also assume we can always write data to
17741 the tty. This isn't nice because the code then blocks until we've
17742 received a complete line of data and it is effectively polling the
17743 keyboard at 1s intervals: however it's quite a bit better than not
17744 working at all :-) A dedicated Windows application might handle this
17745 with an event loop for example.
17749 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17750 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17751 will be called when RSA_sign() and RSA_verify() are used. This is useful
17752 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17753 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17754 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17755 This necessitated the support of an extra signature type NID_md5_sha1
17756 for SSL signatures and modifications to the SSL library to use it instead
17757 of calling RSA_public_decrypt() and RSA_private_encrypt().
17761 * Add new -verify -CAfile and -CApath options to the crl program, these
17762 will lookup a CRL issuers certificate and verify the signature in a
17763 similar way to the verify program. Tidy up the crl program so it
17764 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17765 less strict. It will now permit CRL extensions even if it is not
17766 a V2 CRL: this will allow it to tolerate some broken CRLs.
17770 * Initialize all non-automatic variables each time one of the openssl
17771 sub-programs is started (this is necessary as they may be started
17772 multiple times from the "OpenSSL>" prompt).
17774 *Lennart Bang, Bodo Moeller*
17776 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17777 removing all other RSA functionality (this is what NO_RSA does). This
17778 is so (for example) those in the US can disable those operations covered
17779 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17784 * Non-copying interface to BIO pairs.
17785 (still largely untested)
17789 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17790 ASCII string. This was handled independently in various places before.
17794 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17795 UTF8 strings a character at a time.
17799 * Use client_version from client hello to select the protocol
17800 (s23_srvr.c) and for RSA client key exchange verification
17801 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17805 * Add various utility functions to handle SPKACs, these were previously
17806 handled by poking round in the structure internals. Added new function
17807 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17808 print, verify and generate SPKACs. Based on an original idea from
17809 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17813 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17817 * Allow the config file extension section to be overwritten on the
17818 command line. Based on an original idea from Massimiliano Pala
17819 <madwolf@comune.modena.it>. The new option is called -extensions
17820 and can be applied to ca, req and x509. Also -reqexts to override
17821 the request extensions in req and -crlexts to override the crl extensions
17826 * Add new feature to the SPKAC handling in ca. Now you can include
17827 the same field multiple times by preceding it by "XXXX." for example:
17830 this is the same syntax as used in the req config file.
17834 * Allow certificate extensions to be added to certificate requests. These
17835 are specified in a 'req_extensions' option of the req section of the
17836 config file. They can be printed out with the -text option to req but
17837 are otherwise ignored at present.
17841 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17842 data read consists of only the final block it would not decrypted because
17843 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17844 A misplaced 'break' also meant the decrypted final block might not be
17845 copied until the next read.
17849 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17850 a few extra parameters to the DH structure: these will be useful if
17851 for example we want the value of 'q' or implement X9.42 DH.
17855 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17856 provides hooks that allow the default DSA functions or functions on a
17857 "per key" basis to be replaced. This allows hardware acceleration and
17858 hardware key storage to be handled without major modification to the
17859 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17860 associated functions.
17864 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17865 as "read only": it can't be written to and the buffer it points to will
17866 not be freed. Reading from a read only BIO is much more efficient than
17867 a normal memory BIO. This was added because there are several times when
17868 an area of memory needs to be read from a BIO. The previous method was
17869 to create a memory BIO and write the data to it, this results in two
17870 copies of the data and an O(n^2) reading algorithm. There is a new
17871 function BIO_new_mem_buf() which creates a read only memory BIO from
17872 an area of memory. Also modified the PKCS#7 routines to use read only
17877 * Bugfix: ssl23_get_client_hello did not work properly when called in
17878 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17879 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17880 but a retry condition occurred while trying to read the rest.
17884 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17885 NID_pkcs7_encrypted by default: this was wrong since this should almost
17886 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17887 the encrypted data type: this is a more sensible place to put it and it
17888 allows the PKCS#12 code to be tidied up that duplicated this
17893 * Changed obj_dat.pl script so it takes its input and output files on
17894 the command line. This should avoid shell escape redirection problems
17899 * Initial support for certificate extension requests, these are included
17900 in things like Xenroll certificate requests. Included functions to allow
17901 extensions to be obtained and added.
17905 * -crlf option to s_client and s_server for sending newlines as
17906 CRLF (as required by many protocols).
17910 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17912 * Install libRSAglue.a when OpenSSL is built with RSAref.
17914 *Ralf S. Engelschall*
17916 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17918 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17920 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17925 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17926 DH parameters/keys (q is lost during that conversion, but the resulting
17927 DH parameters contain its length).
17929 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17930 much faster than DH_generate_parameters (which creates parameters
17931 where `p = 2*q + 1`), and also the smaller q makes DH computations
17932 much more efficient (160-bit exponentiation instead of 1024-bit
17933 exponentiation); so this provides a convenient way to support DHE
17934 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17935 utter importance to use
17936 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17938 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17939 when such DH parameters are used, because otherwise small subgroup
17940 attacks may become possible!
17944 * Avoid memory leak in i2d_DHparams.
17948 * Allow the -k option to be used more than once in the enc program:
17949 this allows the same encrypted message to be read by multiple recipients.
17953 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17954 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17955 it will always use the numerical form of the OID, even if it has a short
17960 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17961 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17962 otherwise bn_mod_exp was called. In the case of hardware keys for example
17963 no private key components need be present and it might store extra data
17964 in the RSA structure, which cannot be accessed from bn_mod_exp.
17965 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17966 private key operations.
17970 * Added support for SPARC Linux.
17974 * pem_password_cb function type incompatibly changed from
17975 typedef int pem_password_cb(char *buf, int size, int rwflag);
17977 ....(char *buf, int size, int rwflag, void *userdata);
17978 so that applications can pass data to their callbacks:
17979 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17980 additional void * argument, which is just handed through whenever
17981 the password callback is called.
17983 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17985 New function SSL_CTX_set_default_passwd_cb_userdata.
17987 Compatibility note: As many C implementations push function arguments
17988 onto the stack in reverse order, the new library version is likely to
17989 interoperate with programs that have been compiled with the old
17990 pem_password_cb definition (PEM_whatever takes some data that
17991 happens to be on the stack as its last argument, and the callback
17992 just ignores this garbage); but there is no guarantee whatsoever that
17995 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17996 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17997 problems not only on Windows, but also on some Unix platforms.
17998 To avoid problematic command lines, these definitions are now in an
17999 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18000 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18004 * MIPS III/IV assembler module is reimplemented.
18008 * More DES library cleanups: remove references to srand/rand and
18009 delete an unused file.
18013 * Add support for the free Netwide assembler (NASM) under Win32,
18014 since not many people have MASM (ml) and it can be hard to obtain.
18015 This is currently experimental but it seems to work OK and pass all
18016 the tests. Check out INSTALL.W32 for info.
18020 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18021 without temporary keys kept an extra copy of the server key,
18022 and connections with temporary keys did not free everything in case
18027 * New function RSA_check_key and new openssl rsa option -check
18028 for verifying the consistency of RSA keys.
18030 *Ulf Moeller, Bodo Moeller*
18032 * Various changes to make Win32 compile work:
18033 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18034 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18035 comparison" warnings.
18036 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18040 * Add a debugging option to PKCS#5 v2 key generation function: when
18041 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18042 derived keys are printed to stderr.
18046 * Copy the flags in ASN1_STRING_dup().
18048 *Roman E. Pavlov <pre@mo.msk.ru>*
18050 * The x509 application mishandled signing requests containing DSA
18051 keys when the signing key was also DSA and the parameters didn't match.
18053 It was supposed to omit the parameters when they matched the signing key:
18054 the verifying software was then supposed to automatically use the CA's
18055 parameters if they were absent from the end user certificate.
18057 Omitting parameters is no longer recommended. The test was also
18058 the wrong way round! This was probably due to unusual behaviour in
18059 EVP_cmp_parameters() which returns 1 if the parameters match.
18060 This meant that parameters were omitted when they *didn't* match and
18061 the certificate was useless. Certificates signed with 'ca' didn't have
18064 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18066 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18067 The interface is as follows:
18068 Applications can use
18069 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18070 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18071 "off" is now the default.
18072 The library internally uses
18073 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18074 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18075 to disable memory-checking temporarily.
18077 Some inconsistent states that previously were possible (and were
18078 even the default) are now avoided.
18080 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18081 with each memory chunk allocated; this is occasionally more helpful
18082 than just having a counter.
18084 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18086 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18091 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18092 which largely parallels "options", but is for changing API behaviour,
18093 whereas "options" are about protocol behaviour.
18094 Initial "mode" flags are:
18096 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18097 a single record has been written.
18098 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18099 retries use the same buffer location.
18100 (But all of the contents must be
18105 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18108 * Fix problems with no-hmac etc.
18110 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18112 * New functions RSA_get_default_method(), RSA_set_method() and
18113 RSA_get_method(). These allows replacement of RSA_METHODs without having
18114 to mess around with the internals of an RSA structure.
18118 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18119 Also really enable memory leak checks in openssl.c and in some
18122 *Chad C. Mulligan, Bodo Moeller*
18124 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18125 up the length of negative integers. This has now been simplified to just
18126 store the length when it is first determined and use it later, rather
18127 than trying to keep track of where data is copied and updating it to
18129 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18131 * Add a new function PKCS7_signatureVerify. This allows the verification
18132 of a PKCS#7 signature but with the signing certificate passed to the
18133 function itself. This contrasts with PKCS7_dataVerify which assumes the
18134 certificate is present in the PKCS#7 structure. This isn't always the
18135 case: certificates can be omitted from a PKCS#7 structure and be
18136 distributed by "out of band" means (such as a certificate database).
18140 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18141 function prototypes in pem.h, also change util/mkdef.pl to add the
18142 necessary function names.
18146 * mk1mf.pl (used by Windows builds) did not properly read the
18147 options set by Configure in the top level Makefile, and Configure
18148 was not even able to write more than one option correctly.
18149 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18153 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18154 file to be loaded from a BIO or FILE pointer. The BIO version will
18155 for example allow memory BIOs to contain config info.
18159 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18160 Whoever hopes to achieve shared-library compatibility across versions
18161 must use this, not the compile-time macro.
18162 (Exercise 0.9.4: Which is the minimum library version required by
18164 Note: All this applies only to multi-threaded programs, others don't
18169 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18170 through a BIO pair triggered the default case, i.e.
18171 SSLerr(...,SSL_R_UNKNOWN_STATE).
18175 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18176 can use the SSL library even if none of the specific BIOs is
18181 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18182 for the encoded length.
18184 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18186 * Add initial documentation of the X509V3 functions.
18190 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18191 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18192 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18193 secure PKCS#8 private key format with a high iteration count.
18197 * Fix determination of Perl interpreter: A perl or perl5
18198 *directory* in $PATH was also accepted as the interpreter.
18200 *Ralf S. Engelschall*
18202 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18203 wrong with it but it was very old and did things like calling
18204 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18205 unusual formatting.
18209 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18210 to use the new extension code.
18214 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18215 with macros. This should make it easier to change their form, add extra
18216 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18221 * Add to configuration table a new entry that can specify an alternative
18222 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18223 according to Mark Crispin <MRC@Panda.COM>.
18227 * DES CBC did not update the IV. Weird.
18231 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18232 Changing the behaviour of the former might break existing programs --
18233 where IV updating is needed, des_ncbc_encrypt can be used.
18236 * When bntest is run from "make test" it drives bc to check its
18237 calculations, as well as internally checking them. If an internal check
18238 fails, it needs to cause bc to give a non-zero result or make test carries
18239 on without noticing the failure. Fixed.
18243 * DES library cleanups.
18247 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18248 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18249 ciphers. NOTE: although the key derivation function has been verified
18250 against some published test vectors it has not been extensively tested
18251 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18256 * Instead of "mkdir -p", which is not fully portable, use new
18257 Perl script "util/mkdir-p.pl".
18261 * Rewrite the way password based encryption (PBE) is handled. It used to
18262 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18263 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18264 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18265 the 'parameter' field of the AlgorithmIdentifier is passed to the
18266 underlying key generation function so it must do its own ASN1 parsing.
18267 This has also changed the EVP_PBE_CipherInit() function which now has a
18268 'parameter' argument instead of literal salt and iteration count values
18269 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18273 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18274 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18275 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18276 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18277 value was just used as a "magic string" and not used directly its
18278 value doesn't matter.
18282 * Introduce some semblance of const correctness to BN. Shame C doesn't
18287 * "linux-sparc64" configuration (ultrapenguin).
18289 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18290 "linux-sparc" configuration.
18292 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18294 * config now generates no-xxx options for missing ciphers.
18298 * Support the EBCDIC character set (work in progress).
18299 File ebcdic.c not yet included because it has a different license.
18301 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18303 * Support BS2000/OSD-POSIX.
18305 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18307 * Make callbacks for key generation use `void *` instead of `char *`.
18311 * Make S/MIME samples compile (not yet tested).
18315 * Additional typesafe stacks.
18319 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18323 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18325 * New configuration variant "sco5-gcc".
18327 * Updated some demos.
18329 *Sean O Riordain, Wade Scholine*
18331 * Add missing BIO_free at exit of pkcs12 application.
18335 * Fix memory leak in conf.c.
18339 * Updates for Win32 to assembler version of MD5.
18343 * Set #! path to perl in `apps/der_chop` to where we found it
18344 instead of using a fixed path.
18348 * SHA library changes for irix64-mips4-cc.
18352 * Improvements for VMS support.
18356 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18358 * Bignum library bug fix. IRIX 6 passes "make test" now!
18359 This also avoids the problems with SC4.2 and unpatched SC5.
18361 *Andy Polyakov <appro@fy.chalmers.se>*
18363 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18364 These are required because of the typesafe stack would otherwise break
18365 existing code. If old code used a structure member which used to be STACK
18366 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18367 sk_num or sk_value it would produce an error because the num, data members
18368 are not present in STACK_OF. Now it just produces a warning. sk_set
18369 replaces the old method of assigning a value to sk_value
18370 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18371 that does this will no longer work (and should use sk_set instead) but
18372 this could be regarded as a "questionable" behaviour anyway.
18376 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18377 correctly handle encrypted S/MIME data.
18381 * Change type of various DES function arguments from des_cblock
18382 (which means, in function argument declarations, pointer to char)
18383 to des_cblock * (meaning pointer to array with 8 char elements),
18384 which allows the compiler to do more typechecking; it was like
18385 that back in SSLeay, but with lots of ugly casts.
18387 Introduce new type const_des_cblock.
18391 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18392 problems: find RecipientInfo structure that matches recipient certificate
18393 and initialise the ASN1 structures properly based on passed cipher.
18397 * Belatedly make the BN tests actually check the results.
18401 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18402 to and from BNs: it was completely broken. New compilation option
18403 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18404 key elements as negative integers.
18408 * Reorganize and speed up MD5.
18410 *Andy Polyakov <appro@fy.chalmers.se>*
18414 *Richard Levitte <richard@levitte.org>*
18416 * New option -out to asn1parse to allow the parsed structure to be
18417 output to a file. This is most useful when combined with the -strparse
18418 option to examine the output of things like OCTET STRINGS.
18422 * Make SSL library a little more fool-proof by not requiring any longer
18423 that `SSL_set_{accept,connect}_state` be called before
18424 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18425 in many applications because usually everything *appeared* to work as
18426 intended anyway -- now it really works as intended).
18430 * Move openssl.cnf out of lib/.
18434 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18435 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18436 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18438 *Ralf S. Engelschall*
18440 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18441 handle PKCS#7 enveloped data properly.
18443 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18445 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18446 copying pointers. The cert_st handling is changed by this in
18447 various ways (and thus what used to be known as ctx->default_cert
18448 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18449 any longer when s->cert does not give us what we need).
18450 ssl_cert_instantiate becomes obsolete by this change.
18451 As soon as we've got the new code right (possibly it already is?),
18452 we have solved a couple of bugs of the earlier code where s->cert
18453 was used as if it could not have been shared with other SSL structures.
18455 Note that using the SSL API in certain dirty ways now will result
18456 in different behaviour than observed with earlier library versions:
18457 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18458 does not influence s as it used to.
18460 In order to clean up things more thoroughly, inside SSL_SESSION
18461 we don't use CERT any longer, but a new structure SESS_CERT
18462 that holds per-session data (if available); currently, this is
18463 the peer's certificate chain and, for clients, the server's certificate
18464 and temporary key. CERT holds only those values that can have
18465 meaningful defaults in an SSL_CTX.
18469 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18470 from the internal representation. Various PKCS#7 fixes: remove some
18471 evil casts and set the enc_dig_alg field properly based on the signing
18476 * Allow PKCS#12 password to be set from the command line or the
18477 environment. Let 'ca' get its config file name from the environment
18478 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18483 * Allow certificate policies extension to use an IA5STRING for the
18484 organization field. This is contrary to the PKIX definition but
18485 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18490 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18491 without disallowing inline assembler and the like for non-pedantic builds.
18495 * Support Borland C++ builder.
18497 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18503 * SHA-1 cleanups and performance enhancements.
18505 *Andy Polyakov <appro@fy.chalmers.se>*
18507 * Sparc v8plus assembler for the bignum library.
18509 *Andy Polyakov <appro@fy.chalmers.se>*
18511 * Accept any -xxx and +xxx compiler options in Configure.
18515 * Update HPUX configuration.
18519 * Add missing `sk_<type>_unshift()` function to safestack.h
18521 *Ralf S. Engelschall*
18523 * New function SSL_CTX_use_certificate_chain_file that sets the
18524 "extra_cert"s in addition to the certificate. (This makes sense
18525 only for "PEM" format files, as chains as a whole are not
18530 * Support verify_depth from the SSL API.
18531 x509_vfy.c had what can be considered an off-by-one-error:
18532 Its depth (which was not part of the external interface)
18533 was actually counting the number of certificates in a chain;
18534 now it really counts the depth.
18538 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18539 instead of X509err, which often resulted in confusing error
18540 messages since the error codes are not globally unique
18541 (e.g. an alleged error in ssl3_accept when a certificate
18542 didn't match the private key).
18544 * New function SSL_CTX_set_session_id_context that allows to set a default
18545 value (so that you don't need SSL_set_session_id_context for each
18546 connection using the SSL_CTX).
18550 * OAEP decoding bug fix.
18554 * Support INSTALL_PREFIX for package builders, as proposed by
18559 * New Configure options "threads" and "no-threads". For systems
18560 where the proper compiler options are known (currently Solaris
18561 and Linux), "threads" is the default.
18565 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18569 * Install various scripts to $(OPENSSLDIR)/misc, not to
18570 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18571 such as /usr/local/bin.
18575 * "make linux-shared" to build shared libraries.
18577 *Niels Poppe <niels@netbox.org>*
18579 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18583 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18584 extension adding in x509 utility.
18588 * Remove NOPROTO sections and error code comments.
18592 * Partial rewrite of the DEF file generator to now parse the ANSI
18597 * New Configure options --prefix=DIR and --openssldir=DIR.
18601 * Complete rewrite of the error code script(s). It is all now handled
18602 by one script at the top level which handles error code gathering,
18603 header rewriting and C source file generation. It should be much better
18604 than the old method: it now uses a modified version of Ulf's parser to
18605 read the ANSI prototypes in all header files (thus the old K&R definitions
18606 aren't needed for error creation any more) and do a better job of
18607 translating function codes into names. The old 'ASN1 error code embedded
18608 in a comment' is no longer necessary and it doesn't use .err files which
18609 have now been deleted. Also the error code call doesn't have to appear all
18610 on one line (which resulted in some large lines...).
18614 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18618 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18619 0 (which usually indicates a closed connection), but continue reading.
18623 * Fix some race conditions.
18627 * Add support for CRL distribution points extension. Add Certificate
18628 Policies and CRL distribution points documentation.
18632 * Move the autogenerated header file parts to crypto/opensslconf.h.
18636 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18637 8 of keying material. Merlin has also confirmed interop with this fix
18638 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18640 *Merlin Hughes <merlin@baltimore.ie>*
18642 * Fix lots of warnings.
18644 *Richard Levitte <levitte@stacken.kth.se>*
18646 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18647 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18649 *Richard Levitte <levitte@stacken.kth.se>*
18651 * Fix problems with sizeof(long) == 8.
18653 *Andy Polyakov <appro@fy.chalmers.se>*
18655 * Change functions to ANSI C.
18659 * Fix typos in error codes.
18661 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18663 * Remove defunct assembler files from Configure.
18667 * SPARC v8 assembler BIGNUM implementation.
18669 *Andy Polyakov <appro@fy.chalmers.se>*
18671 * Support for Certificate Policies extension: both print and set.
18672 Various additions to support the r2i method this uses.
18676 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18677 return a const string when you are expecting an allocated buffer.
18681 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18682 types DirectoryString and DisplayText.
18686 * Add code to allow r2i extensions to access the configuration database,
18687 add an LHASH database driver and add several ctx helper functions.
18691 * Fix an evil bug in bn_expand2() which caused various BN functions to
18692 fail when they extended the size of a BIGNUM.
18696 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18697 support typesafe stack.
18701 * Fix typo in SSL_[gs]et_options().
18703 *Nils Frostberg <nils@medcom.se>*
18705 * Delete various functions and files that belonged to the (now obsolete)
18706 old X509V3 handling code.
18710 * New Configure option "rsaref".
18714 * Don't auto-generate pem.h.
18718 * Introduce type-safe ASN.1 SETs.
18722 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18724 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18726 * Introduce type-safe STACKs. This will almost certainly break lots of code
18727 that links with OpenSSL (well at least cause lots of warnings), but fear
18728 not: the conversion is trivial, and it eliminates loads of evil casts. A
18729 few STACKed things have been converted already. Feel free to convert more.
18730 In the fullness of time, I'll do away with the STACK type altogether.
18734 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18735 specified in `<certfile>` by updating the entry in the index.txt file.
18736 This way one no longer has to edit the index.txt file manually for
18737 revoking a certificate. The -revoke option does the gory details now.
18739 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18741 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18742 `-text` option at all and this way the `-noout -text` combination was
18743 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18745 *Ralf S. Engelschall*
18747 * Make sure a corresponding plain text error message exists for the
18748 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18749 verify callback function determined that a certificate was revoked.
18751 *Ralf S. Engelschall*
18753 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18754 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18755 all available ciphers including rc5, which was forgotten until now.
18756 In order to let the testing shell script know which algorithms
18757 are available, a new (up to now undocumented) command
18758 `openssl list-cipher-commands` is used.
18762 * Bugfix: s_client occasionally would sleep in select() when
18763 it should have checked SSL_pending() first.
18767 * New functions DSA_do_sign and DSA_do_verify to provide access to
18768 the raw DSA values prior to ASN.1 encoding.
18772 * Tweaks to Configure
18774 *Niels Poppe <niels@netbox.org>*
18776 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18781 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18785 * New config option to avoid instructions that are illegal on the 80386.
18786 The default code is faster, but requires at least a 486.
18790 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18791 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18792 same as SSL2_VERSION anyway.
18796 * New "-showcerts" option for s_client.
18800 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18801 application. Various cleanups and fixes.
18805 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18806 modify error routines to work internally. Add error codes and PBE init
18807 to library startup routines.
18811 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18812 packing functions to asn1 and evp. Changed function names and error
18813 codes along the way.
18817 * PKCS12 integration: and so it begins... First of several patches to
18818 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18819 objects to objects.h
18823 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18824 and display support for Thawte strong extranet extension.
18828 * Add LinuxPPC support.
18830 *Jeff Dubrule <igor@pobox.org>*
18832 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18833 bn_div_words in alpha.s.
18835 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18837 * Make sure the RSA OAEP test is skipped under -DRSAref because
18838 OAEP isn't supported when OpenSSL is built with RSAref.
18840 *Ulf Moeller <ulf@fitug.de>*
18842 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18843 so they no longer are missing under -DNOPROTO.
18845 *Soren S. Jorvang <soren@t.dk>*
18847 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18849 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18850 doesn't work when the session is reused. Coming soon!
18854 * Fix a security hole, that allows sessions to be reused in the wrong
18855 context thus bypassing client cert protection! All software that uses
18856 client certs and session caches in multiple contexts NEEDS PATCHING to
18857 allow session reuse! A fuller solution is in the works.
18859 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18861 * Some more source tree cleanups (removed obsolete files
18862 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18863 permission on "config" script to be executable) and a fix for the INSTALL
18866 *Ulf Moeller <ulf@fitug.de>*
18868 * Remove some legacy and erroneous uses of malloc, free instead of
18871 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18873 * Make rsa_oaep_test return non-zero on error.
18875 *Ulf Moeller <ulf@fitug.de>*
18877 * Add support for native Solaris shared libraries. Configure
18878 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18879 if someone would make that last step automatic.
18881 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18883 * ctx_size was not built with the right compiler during "make links". Fixed.
18887 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18888 except NULL ciphers". This means the default cipher list will no longer
18889 enable NULL ciphers. They need to be specifically enabled e.g. with
18890 the string "DEFAULT:eNULL".
18894 * Fix to RSA private encryption routines: if p < q then it would
18895 occasionally produce an invalid result. This will only happen with
18896 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18900 * Be less restrictive and allow also `perl util/perlpath.pl
18901 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18902 because this way one can also use an interpreter named `perl5` (which is
18903 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18904 installed as `perl`).
18906 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18908 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18910 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18912 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18913 advapi32.lib to Win32 build and change the pem test comparison
18914 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18915 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18916 and crypto/des/ede_cbcm_enc.c.
18920 * DES quad checksum was broken on big-endian architectures. Fixed.
18924 * Comment out two functions in bio.h that aren't implemented. Fix up the
18925 Win32 test batch file so it (might) work again. The Win32 test batch file
18926 is horrible: I feel ill....
18930 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18931 in e_os.h. Audit of header files to check ANSI and non ANSI
18932 sections: 10 functions were absent from non ANSI section and not exported
18933 from Windows DLLs. Fixed up libeay.num for new functions.
18937 * Make `openssl version` output lines consistent.
18939 *Ralf S. Engelschall*
18941 * Fix Win32 symbol export lists for BIO functions: Added
18942 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18943 to ms/libeay{16,32}.def.
18945 *Ralf S. Engelschall*
18947 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18948 fine under Unix and passes some trivial tests I've now added. But the
18949 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18950 added to make sure no one expects that this stuff really works in the
18951 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18952 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18955 *Ralf S. Engelschall*
18957 * Fix the generation of two part addresses in perl.
18959 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18961 * Add config entry for Linux on MIPS.
18963 *John Tobey <jtobey@channel1.com>*
18965 * Make links whenever Configure is run, unless we are on Windoze.
18969 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18970 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18975 * Add a useful kludge to allow package maintainers to specify compiler and
18976 other platforms details on the command line without having to patch the
18977 Configure script every time: One now can use
18978 `perl Configure <id>:<details>`,
18979 i.e. platform ids are allowed to have details appended
18980 to them (separated by colons). This is treated as there would be a static
18981 pre-configured entry in Configure's %table under key `<id>` with value
18982 `<details>` and `perl Configure <id>` is called. So, when you want to
18983 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18984 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18985 now, which overrides the FreeBSD-elf entry on-the-fly.
18987 *Ralf S. Engelschall*
18989 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18993 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18994 on the `perl Configure ...` command line. This way one can compile
18995 OpenSSL libraries with Position Independent Code (PIC) which is needed
18996 for linking it into DSOs.
18998 *Ralf S. Engelschall*
19000 * Remarkably, export ciphers were totally broken and no-one had noticed!
19005 * Cleaned up the LICENSE document: The official contact for any license
19006 questions now is the OpenSSL core team under openssl-core@openssl.org.
19007 And add a paragraph about the dual-license situation to make sure people
19008 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19009 to the OpenSSL toolkit.
19011 *Ralf S. Engelschall*
19013 * General source tree makefile cleanups: Made `making xxx in yyy...`
19014 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19015 Additionally cleaned up the `make links` target: Remove unnecessary
19016 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19017 to speed processing and no longer clutter the display with confusing
19018 stuff. Instead only the actually done links are displayed.
19020 *Ralf S. Engelschall*
19022 * Permit null encryption ciphersuites, used for authentication only. It used
19023 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19024 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19029 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19030 signed attributes when verifying signatures (this would break them),
19031 the detached data encoding was wrong and public keys obtained using
19032 X509_get_pubkey() weren't freed.
19036 * Add text documentation for the BUFFER functions. Also added a work around
19037 to a Win95 console bug. This was triggered by the password read stuff: the
19038 last character typed gets carried over to the next fread(). If you were
19039 generating a new cert request using 'req' for example then the last
19040 character of the passphrase would be CR which would then enter the first
19045 * Added the new 'Includes OpenSSL Cryptography Software' button as
19046 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19047 button and can be used by applications based on OpenSSL to show the
19048 relationship to the OpenSSL project.
19050 *Ralf S. Engelschall*
19052 * Remove confusing variables in function signatures in files
19053 ssl/ssl_lib.c and ssl/ssl.h.
19055 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19057 * Don't install bss_file.c under PREFIX/include/
19059 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19061 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19062 functions that return function pointers and has support for NT specific
19063 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19064 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19065 unsigned to signed types: this was killing the Win32 compile.
19069 * Add new certificate file to stack functions,
19070 SSL_add_dir_cert_subjects_to_stack() and
19071 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19072 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19073 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19074 This means that Apache-SSL and similar packages don't have to mess around
19075 to add as many CAs as they want to the preferred list.
19079 * Experiment with doxygen documentation. Currently only partially applied to
19081 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19082 openssl.doxy as the configuration file.
19086 * Get rid of remaining C++-style comments which strict C compilers hate.
19088 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19090 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19091 compiled in by default: it has problems with large keys.
19095 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19096 DH private keys and/or callback functions which directly correspond to
19097 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19098 is needed for applications which have to configure certificates on a
19099 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19101 For the RSA certificate situation is makes no difference, but
19102 for the DSA certificate situation this fixes the "no shared cipher"
19103 problem where the OpenSSL cipher selection procedure failed because the
19104 temporary keys were not overtaken from the context and the API provided
19105 no way to reconfigure them.
19106 The new functions now let applications reconfigure the stuff and they
19107 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19108 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19109 non-public-API function ssl_cert_instantiate() is used as a helper
19110 function and also to reduce code redundancy inside ssl_rsa.c.
19112 *Ralf S. Engelschall*
19114 * Move s_server -dcert and -dkey options out of the undocumented feature
19115 area because they are useful for the DSA situation and should be
19116 recognized by the users.
19118 *Ralf S. Engelschall*
19120 * Fix the cipher decision scheme for export ciphers: the export bits are
19121 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19122 SSL_EXP_MASK. So, the original variable has to be used instead of the
19123 already masked variable.
19125 *Richard Levitte <levitte@stacken.kth.se>*
19127 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19129 *Richard Levitte <levitte@stacken.kth.se>*
19131 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19132 from `int` to `unsigned int` because it is a length and initialized by
19133 EVP_DigestFinal() which expects an `unsigned int *`.
19135 *Richard Levitte <levitte@stacken.kth.se>*
19137 * Don't hard-code path to Perl interpreter on shebang line of Configure
19138 script. Instead use the usual Shell->Perl transition trick.
19140 *Ralf S. Engelschall*
19142 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19143 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19144 -noout -modulus` as it's already the case for `openssl rsa -noout
19145 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19146 currently the public key is printed (a decision which was already done by
19147 `openssl dsa -modulus` in the past) which serves a similar purpose.
19148 Additionally the NO_RSA no longer completely removes the whole -modulus
19149 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19152 *Ralf S. Engelschall*
19154 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19155 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19157 *Arne Ansper <arne@ats.cyber.ee>*
19159 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19160 to be added. Now both 'req' and 'ca' can use new objects defined in the
19165 * Add cool BIO that does syslog (or event log on NT).
19167 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19169 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19170 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19171 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19172 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19176 * Add preliminary config info for new extension code.
19180 * Make RSA_NO_PADDING really use no padding.
19182 *Ulf Moeller <ulf@fitug.de>*
19184 * Generate errors when private/public key check is done.
19188 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19189 for some CRL extensions and new objects added.
19193 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19194 key usage extension and fuller support for authority key id.
19198 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19199 padding method for RSA, which is recommended for new applications in PKCS
19200 #1 v2.0 (RFC 2437, October 1998).
19201 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19202 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19203 against Bleichbacher's attack on RSA.
19204 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19207 * Updates to the new SSL compression code
19209 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19211 * Fix so that the version number in the master secret, when passed
19212 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19213 (because the server will not accept higher), that the version number
19214 is 0x03,0x01, not 0x03,0x00
19216 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19218 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19219 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19220 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19224 * Support for RAW extensions where an arbitrary extension can be
19225 created by including its DER encoding. See `apps/openssl.cnf` for
19230 * Make sure latest Perl versions don't interpret some generated C array
19231 code as Perl array code in the crypto/err/err_genc.pl script.
19233 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19235 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19236 not many people have the assembler. Various Win32 compilation fixes and
19237 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19238 build instructions.
19242 * Modify configure script 'Configure' to automatically create crypto/date.h
19243 file under Win32 and also build pem.h from pem.org. New script
19244 util/mkfiles.pl to create the MINFO file on environments that can't do a
19245 'make files': perl util/mkfiles.pl >MINFO should work.
19249 * Major rework of DES function declarations, in the pursuit of correctness
19250 and purity. As a result, many evil casts evaporated, and some weirdness,
19251 too. You may find this causes warnings in your code. Zapping your evil
19252 casts will probably fix them. Mostly.
19256 * Fix for a typo in asn1.h. Bug fix to object creation script
19257 obj_dat.pl. It considered a zero in an object definition to mean
19258 "end of object": none of the objects in objects.h have any zeros
19259 so it wasn't spotted.
19261 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19263 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19264 Masking (CBCM). In the absence of test vectors, the best I have been able
19265 to do is check that the decrypt undoes the encrypt, so far. Send me test
19266 vectors if you have them.
19270 * Correct calculation of key length for export ciphers (too much space was
19271 allocated for null ciphers). This has not been tested!
19275 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19276 message is now correct (it understands "crypto" and "ssl" on its
19277 command line). There is also now an "update" option. This will update
19278 the util/ssleay.num and util/libeay.num files with any new functions.
19280 perl util/mkdef.pl crypto ssl update
19281 it will update them.
19285 * Overhauled the Perl interface:
19286 - ported BN stuff to OpenSSL's different BN library
19287 - made the perl/ source tree CVS-aware
19288 - renamed the package from SSLeay to OpenSSL (the files still contain
19289 their history because I've copied them in the repository)
19290 - removed obsolete files (the test scripts will be replaced
19291 by better Test::Harness variants in the future)
19293 *Ralf S. Engelschall*
19295 * First cut for a very conservative source tree cleanup:
19296 1. merge various obsolete readme texts into doc/ssleay.txt
19297 where we collect the old documents and readme texts.
19298 2. remove the first part of files where I'm already sure that we no
19299 longer need them because of three reasons: either they are just temporary
19300 files which were left by Eric or they are preserved original files where
19301 I've verified that the diff is also available in the CVS via "cvs diff
19302 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19303 the crypto/md/ stuff).
19305 *Ralf S. Engelschall*
19307 * More extension code. Incomplete support for subject and issuer alt
19308 name, issuer and authority key id. Change the i2v function parameters
19309 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19310 what that's for :-) Fix to ASN1 macro which messed up
19311 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19315 * Preliminary support for ENUMERATED type. This is largely copied from the
19320 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19322 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19324 * Make sure `make rehash` target really finds the `openssl` program.
19326 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19328 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19329 like to hear about it if this slows down other processors.
19333 * Add CygWin32 platform information to Configure script.
19335 *Alan Batie <batie@aahz.jf.intel.com>*
19337 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19339 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19341 * New program nseq to manipulate netscape certificate sequences
19345 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19350 * Fixes to BN code. Previously the default was to define BN_RECURSION
19351 but the BN code had some problems that would cause failures when
19352 doing certificate verification and some other functions.
19354 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19356 * Add ASN1 and PEM code to support netscape certificate sequences.
19360 * Add ASN1 and PEM code to support netscape certificate sequences.
19364 * Add several PKIX and private extended key usage OIDs.
19368 * Modify the 'ca' program to handle the new extension code. Modify
19369 openssl.cnf for new extension format, add comments.
19373 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19374 and add a sample to openssl.cnf so req -x509 now adds appropriate
19379 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19380 error code, add initial support to X509_print() and x509 application.
19384 * Takes a deep breath and start adding X509 V3 extension support code. Add
19385 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19386 stuff is currently isolated and isn't even compiled yet.
19390 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19391 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19392 Removed the versions check from X509 routines when loading extensions:
19393 this allows certain broken certificates that don't set the version
19394 properly to be processed.
19398 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19399 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19400 can still be regenerated with "make depend".
19404 * Spelling mistake in C version of CAST-128.
19406 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19408 * Changes to the error generation code. The perl script err-code.pl
19409 now reads in the old error codes and retains the old numbers, only
19410 adding new ones if necessary. It also only changes the .err files if new
19411 codes are added. The makefiles have been modified to only insert errors
19412 when needed (to avoid needlessly modifying header files). This is done
19413 by only inserting errors if the .err file is newer than the auto generated
19414 C file. To rebuild all the error codes from scratch (the old behaviour)
19415 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19416 or delete all the .err files.
19420 * CAST-128 was incorrectly implemented for short keys. The C version has
19421 been fixed, but is untested. The assembler versions are also fixed, but
19422 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19423 to regenerate it if needed.
19424 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19425 Hagino <itojun@kame.net>*
19427 * File was opened incorrectly in randfile.c.
19429 *Ulf Möller <ulf@fitug.de>*
19431 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19432 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19433 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19434 al: it's just almost always a UTCTime. Note this patch adds new error
19435 codes so do a "make errors" if there are problems.
19439 * Correct Linux 1 recognition in config.
19441 *Ulf Möller <ulf@fitug.de>*
19443 * Remove pointless MD5 hash when using DSA keys in ca.
19445 *Anonymous <nobody@replay.com>*
19447 * Generate an error if given an empty string as a cert directory. Also
19448 generate an error if handed NULL (previously returned 0 to indicate an
19449 error, but didn't set one).
19451 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19453 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19457 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19458 parameters. This was causing a warning which killed off the Win32 compile.
19462 * Remove C++ style comments from crypto/bn/bn_local.h.
19464 *Neil Costigan <neil.costigan@celocom.com>*
19466 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19467 based on a text string, looking up short and long names and finally
19468 "dot" format. The "dot" format stuff didn't work. Added new function
19469 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19470 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19471 OID is not part of the table.
19475 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19476 X509_LOOKUP_by_alias().
19480 * Sort openssl functions by name.
19484 * Get the `gendsa` command working and add it to the `list` command. Remove
19485 encryption from sample DSA keys (in case anyone is interested the password
19490 * Make *all* `*_free` functions accept a NULL pointer.
19492 *Frans Heymans <fheymans@isaserver.be>*
19494 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19497 *Anonymous <nobody@replay.com>*
19499 * s_server should send the CAfile as acceptable CAs, not its own cert.
19501 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19503 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19505 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19507 * Temp key "for export" tests were wrong in s3_srvr.c.
19509 *Anonymous <nobody@replay.com>*
19511 * Add prototype for temp key callback functions
19512 SSL_CTX_set_tmp_{rsa,dh}_callback().
19516 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19517 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19521 * X509_name_add_entry() freed the wrong thing after an error.
19523 *Arne Ansper <arne@ats.cyber.ee>*
19525 * rsa_eay.c would attempt to free a NULL context.
19527 *Arne Ansper <arne@ats.cyber.ee>*
19529 * BIO_s_socket() had a broken should_retry() on Windoze.
19531 *Arne Ansper <arne@ats.cyber.ee>*
19533 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19535 *Arne Ansper <arne@ats.cyber.ee>*
19537 * Make sure the already existing X509_STORE->depth variable is initialized
19538 in X509_STORE_new(), but document the fact that this variable is still
19539 unused in the certificate verification process.
19541 *Ralf S. Engelschall*
19543 * Fix the various library and `apps/` files to free up pkeys obtained from
19544 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19548 * Fix reference counting in X509_PUBKEY_get(). This makes
19549 demos/maurice/example2.c work, amongst others, probably.
19551 *Steve Henson and Ben Laurie*
19553 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19554 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19555 are no longer created. This way we have a single and consistent command
19556 line interface `openssl <command>`, similar to `cvs <command>`.
19558 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19560 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19561 BIT STRING wrapper always have zero unused bits.
19565 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19569 * Make the top-level INSTALL documentation easier to understand.
19573 * Makefiles updated to exit if an error occurs in a sub-directory
19574 make (including if user presses ^C) [Paul Sutton]
19576 * Make Montgomery context stuff explicit in RSA data structure.
19580 * Fix build order of pem and err to allow for generated pem.h.
19584 * Fix renumbering bug in X509_NAME_delete_entry().
19588 * Enhanced the err-ins.pl script so it makes the error library number
19589 global and can add a library name. This is needed for external ASN1 and
19590 other error libraries.
19594 * Fixed sk_insert which never worked properly.
19598 * Fix ASN1 macros so they can handle indefinite length constructed
19599 EXPLICIT tags. Some non standard certificates use these: they can now
19604 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19605 into a single doc/ssleay.txt bundle. This way the information is still
19606 preserved but no longer messes up this directory. Now it's new room for
19607 the new set of documentation files.
19609 *Ralf S. Engelschall*
19611 * SETs were incorrectly DER encoded. This was a major pain, because they
19612 shared code with SEQUENCEs, which aren't coded the same. This means that
19613 almost everything to do with SETs or SEQUENCEs has either changed name or
19614 number of arguments.
19616 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19618 * Fix test data to work with the above.
19622 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19623 was already fixed by Eric for 0.9.1 it seems.
19625 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19627 * Autodetect FreeBSD3.
19631 * Fix various bugs in Configure. This affects the following platforms:
19635 unixware-2.0-pentium
19640 * Eliminate generated files from CVS. Reorder tests to regenerate files
19641 before they are needed.
19645 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19649 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19651 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19652 changed SSLeay to OpenSSL in version strings.
19654 *Ralf S. Engelschall*
19656 * Some fixups to the top-level documents.
19660 * Fixed the nasty bug where rsaref.h was not found under compile-time
19661 because the symlink to include/ was missing.
19663 *Ralf S. Engelschall*
19665 * Incorporated the popular no-RSA/DSA-only patches
19666 which allow to compile an RSA-free SSLeay.
19668 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19670 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19671 when "ssleay" is still not found.
19673 *Ralf S. Engelschall*
19675 * Added more platforms to Configure: Cray T3E, HPUX 11,
19677 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19679 * Updated the README file.
19681 *Ralf S. Engelschall*
19683 * Added various .cvsignore files in the CVS repository subdirs
19684 to make a "cvs update" really silent.
19686 *Ralf S. Engelschall*
19688 * Recompiled the error-definition header files and added
19689 missing symbols to the Win32 linker tables.
19691 *Ralf S. Engelschall*
19693 * Cleaned up the top-level documents;
19694 o new files: CHANGES and LICENSE
19695 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19696 o merged COPYRIGHT into LICENSE
19697 o removed obsolete TODO file
19698 o renamed MICROSOFT to INSTALL.W32
19700 *Ralf S. Engelschall*
19702 * Removed dummy files from the 0.9.1b source tree:
19703 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19704 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19705 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19706 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19707 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19709 *Ralf S. Engelschall*
19711 * Added various platform portability fixes.
19715 * The Genesis of the OpenSSL rpject:
19716 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19717 Young and Tim J. Hudson created while they were working for C2Net until
19720 *The OpenSSL Project*
19722 ### Changes between 0.9.0b and 0.9.1b [not released]
19724 * Updated a few CA certificates under certs/
19728 * Changed some BIGNUM api stuff.
19732 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19733 DGUX x86, Linux Alpha, etc.
19737 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19738 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19743 * Add -strparse option to asn1pars program which parses nested
19746 *Dr Stephen Henson <shenson@bigfoot.com>*
19748 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19752 * DSA fix for "ca" program.
19756 * Added "-genkey" option to "dsaparam" program.
19760 * Added RIPE MD160 (rmd160) message digest.
19764 * Added -a (all) option to "ssleay version" command.
19768 * Added PLATFORM define which is the id given to Configure.
19772 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19776 * Extended the ASN.1 parser routines.
19780 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19784 * Added a BN_CTX to the BN library.
19788 * Fixed the weak key values in DES library
19792 * Changed API in EVP library for cipher aliases.
19796 * Added support for RC2/64bit cipher.
19800 * Converted the lhash library to the crypto/mem.c functions.
19804 * Added more recognized ASN.1 object ids.
19808 * Added more RSA padding checks for SSL/TLS.
19812 * Added BIO proxy/filter functionality.
19816 * Added extra_certs to SSL_CTX which can be used
19817 send extra CA certificates to the client in the CA cert chain sending
19818 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19822 * Now Fortezza is denied in the authentication phase because
19823 this is key exchange mechanism is not supported by SSLeay at all.
19827 * Additional PKCS1 checks.
19831 * Support the string "TLSv1" for all TLS v1 ciphers.
19835 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19836 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19840 * Fixed a few memory leaks.
19844 * Fixed various code and comment typos.
19848 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19849 bytes sent in the client random.
19851 *Edward Bishop <ebishop@spyglass.com>*
19855 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
19856 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
19857 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
19858 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
19859 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
19860 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
19861 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
19862 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
19863 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
19864 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19865 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
19866 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19867 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19868 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19869 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19870 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19871 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19872 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19873 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19874 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19875 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19876 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19877 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19878 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19879 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19880 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19881 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19882 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19883 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19884 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19885 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19886 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19887 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19888 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19889 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19890 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19891 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19892 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19893 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19894 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19895 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19896 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19897 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19898 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19899 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19900 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19901 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19902 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19903 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19904 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19905 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19906 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19907 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19908 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19909 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19910 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19911 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19912 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19913 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19914 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19915 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19916 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19917 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19918 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19919 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19920 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19921 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19922 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19923 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19924 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19925 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19926 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19927 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19928 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19929 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19930 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19931 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19932 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19933 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19934 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19935 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19936 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19937 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19938 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19939 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19940 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19941 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19942 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19943 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19944 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19945 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19946 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19947 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19948 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19949 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19950 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19951 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19952 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19953 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19954 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19955 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19956 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19957 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19958 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19959 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19960 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19961 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19962 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19963 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19964 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19965 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19966 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19967 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19968 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19969 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19970 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19971 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19972 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19973 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19974 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19975 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19976 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19977 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19978 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19979 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19980 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19981 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19982 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19983 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19984 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19985 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19986 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19987 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19988 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19989 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19990 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19991 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19992 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19993 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19994 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19995 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19996 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19997 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19998 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19999 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20000 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20001 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20002 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20003 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20004 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20005 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20006 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20007 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20008 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20009 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20010 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20011 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20012 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20013 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20014 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20015 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20016 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20017 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20018 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20019 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20020 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20021 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20022 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20023 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20024 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20025 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20026 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20027 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655