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 client side support for QUIC
32 * Added SHA256/192 algorithm support.
36 * Provide a new configure option `no-http` that can be used to disable HTTP
41 * TLS round-trip time calculation was added by a Brigham Young University
42 Capstone team partnering with Sandia National Laboratories. A new function
43 in ssl_lib titled SSL_get_handshake_rtt will calculate and retrieve this
48 * Added the "-quic" option to s_client to enable connectivity to QUIC servers.
49 QUIC requires the use of ALPN, so this must be specified via the "-alpn"
50 option. Use of the "advanced" s_client command command via the "-adv" option
55 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
56 The previous fix for this timing side channel turned out to cause
57 a severe 2-3x performance regression in the typical use case
58 compared to 3.0.7. The new fix uses existing constant time
59 code paths, and restores the previous performance level while
60 fully eliminating all existing timing side channels.
61 The fix was developed by Bernd Edlinger with testing support
66 * Added an "advanced" command mode to s_client. Use this with the "-adv"
67 option. The old "basic" command mode recognises certain letters that must
68 always appear at the start of a line and cannot be escaped. The advanced
69 command mode enables commands to be entered anywhere and there is an
70 escaping mechanism. After starting s_client with "-adv" type "{help}"
71 to show a list of available commands.
75 * Add Raw Public Key (RFC7250) support. Authentication is supported
76 by matching keys against either local policy (TLSA records synthesised
77 from the expected keys) or DANE (TLSA records obtained by the
78 application from DNS). TLSA records will also match the same key in
79 the server certificate, should RPK use not happen to be negotiated.
83 * Added EC_GROUP_to_params which creates an OSSL_PARAM array
84 from a given EC_GROUP.
88 * Added support for pluggable (provider-based) TLS signature algorithms.
89 This enables TLS 1.3 authentication operations with algorithms embedded
90 in providers not included by default in OpenSSL. In combination with
91 the already available pluggable KEM and X.509 support, this enables
92 for example suitable providers to deliver post-quantum or quantum-safe
93 cryptography to OpenSSL users.
97 * Added support for Hybrid Public Key Encryption (HPKE) as defined
98 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
99 Message Layer Security (MLS) and other IETF specifications.
100 HPKE can also be used by other applications that require
101 encrypting "to" an ECDH public key. External APIs are defined in
102 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
106 * Add support for certificate compression (RFC8879), including
107 library support for Brotli and Zstandard compression.
111 * Add the ability to add custom attributes to PKCS12 files. Add a new API
112 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
113 for a user specified callback and optional argument.
114 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
115 added to the existing STACK_OF attrs.
119 * Major refactor of the libssl record layer
123 * Added a new BIO_s_dgram_mem() to read/write datagrams to memory
127 * Add a mac salt length option for the pkcs12 command.
131 * Add more SRTP protection profiles from RFC8723 and RFC8269.
135 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
137 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
139 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
140 supported and enabled.
144 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
145 to the list of ciphersuites providing Perfect Forward Secrecy as
146 required by SECLEVEL >= 3.
148 *Dmitry Belyavskiy, Nicola Tuveri*
150 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
151 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
152 SSL_get0_iana_groups() function-like macro, retrieves the list of
153 supported groups sent by the peer.
154 The function SSL_client_hello_get_extension_order() populates
155 a caller-supplied array with the list of extension types present in the
156 ClientHello, in order of appearance.
160 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
161 to make it possible to use empty passphrase strings.
165 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
166 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
167 of 160 bits and above and less than 224 bits were previously accepted by
168 default but are now no longer allowed. By default TLS compression was
169 already disabled in previous OpenSSL versions. At security level 2 it cannot
174 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
179 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
180 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
181 will need to load the legacy crypto provider.
184 * CCM8 cipher suites in TLS have been downgraded to security level zero
185 because they use a short authentication tag which lowers their strength.
189 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
194 * Add X.509 certificate codeSigning purpose and related checks on key usage and
195 extended key usage of the leaf certificate according to the CA/Browser Forum.
199 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
200 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
201 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
202 X.509 version 3 if the certificate information includes X.509 extensions.
206 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
207 such as adding a trace facility for debugging certificate chain building.
211 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
212 in particular supporting requests for central key generation, generalized
213 polling, and various types of genm/genp exchanges defined in CMP Updates.
217 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
218 like correcting the TLS and proxy support and adding tracing for debugging.
222 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
226 * `CMS_add0_cert()` and `CMS_add1_cert()` no more throw an error
227 if a certificate to be added is already present.
228 * `CMS_sign_ex()` and `CMS_sign()` now ignore any duplicate certificates
229 in their `certs` argument and no longer throw an error for them.
233 * Fixed and extended `util/check-format.pl` for checking adherence to the
234 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
235 The checks are meanwhile more complete and yield fewer false positives.
239 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
240 sending and receiving multiple messages in a single call. An implementation
241 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
245 * Support for loading root certificates from the Windows certificate store
246 has been added. The support is in the form of a store which recognises the
247 URI string of `org.openssl.winstore://`. This URI scheme currently takes no
248 arguments. This store is built by default and can be disabled using the new
249 compile-time option `no-winstore`. This store is not currently used by
250 default and must be loaded explicitly using the above store URI. It is
251 expected to be loaded by default in the future.
255 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
256 kernel versions that support KTLS have a known bug in CCM processing. That
257 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
258 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
263 * Zerocopy KTLS sendfile() support on Linux.
267 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
268 decryption as a protection against Bleichenbacher-like attacks.
269 The RSA decryption API will now return a randomly generated deterministic
270 message instead of an error in case it detects an error when checking
271 padding during PKCS#1 v1.5 decryption. This is a general protection against
272 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
274 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
275 on the RSA decryption context.
279 * Support for Argon2d, Argon2i, Argon2id KDFs has been added along with
280 basic thread pool implementation for select platforms.
287 ### Changes between 3.1.1 and 3.1.2 [xx XXX xxxx]
289 * Do not ignore empty associated data entries with AES-SIV.
291 The AES-SIV algorithm allows for authentication of multiple associated
292 data entries along with the encryption. To authenticate empty data the
293 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
294 with NULL pointer as the output buffer and 0 as the input buffer length.
295 The AES-SIV implementation in OpenSSL just returns success for such call
296 instead of performing the associated data authentication operation.
297 The empty data thus will not be authenticated. ([CVE-2023-2975])
299 Thanks to Juerg Wullschleger (Google) for discovering the issue.
301 The fix changes the authentication tag value and the ciphertext for
302 applications that use empty associated data entries with AES-SIV.
303 To decrypt data encrypted with previous versions of OpenSSL the application
304 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
309 * When building with the `enable-fips` option and using the resulting
310 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
311 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
312 not operate with truncated digests (FIPS 140-3 IG G.R).
316 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
318 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
319 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
321 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
322 numeric text form. For gigantic sub-identifiers, this would take a very
323 long time, the time complexity being O(n^2) where n is the size of that
324 sub-identifier. ([CVE-2023-2650])
326 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
327 IDENTIFIER to canonical numeric text form if the size of that OBJECT
328 IDENTIFIER is 586 bytes or less, and fail otherwise.
330 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
331 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
332 most 128 sub-identifiers, and that the maximum value that each sub-
333 identifier may have is 2^32-1 (4294967295 decimal).
335 For each byte of every sub-identifier, only the 7 lower bits are part of
336 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
337 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
342 * Multiple algorithm implementation fixes for ARM BE platforms.
346 * Added a -pedantic option to fipsinstall that adjusts the various
347 settings to ensure strict FIPS compliance rather than backwards
352 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
353 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
354 trigger a crash of an application using AES-XTS decryption if the memory
355 just after the buffer being decrypted is not mapped.
356 Thanks to Anton Romanov (Amazon) for discovering the issue.
361 * Add FIPS provider configuration option to disallow the use of
362 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
363 The option '-no_drbg_truncated_digests' can optionally be
364 supplied to 'openssl fipsinstall'.
368 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
369 that it does not enable policy checking. Thanks to David Benjamin for
370 discovering this issue.
375 * Fixed an issue where invalid certificate policies in leaf certificates are
376 silently ignored by OpenSSL and other certificate policy checks are skipped
377 for that certificate. A malicious CA could use this to deliberately assert
378 invalid certificate policies in order to circumvent policy checking on the
379 certificate altogether.
384 * Limited the number of nodes created in a policy tree to mitigate
385 against CVE-2023-0464. The default limit is set to 1000 nodes, which
386 should be sufficient for most installations. If required, the limit
387 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
388 time define to a desired maximum number of nodes or zero to allow
394 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
396 * Add FIPS provider configuration option to enforce the
397 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
398 The option '-ems-check' can optionally be supplied to
399 'openssl fipsinstall'.
403 * The FIPS provider includes a few non-approved algorithms for
404 backward compatibility purposes and the "fips=yes" property query
405 must be used for all algorithm fetches to ensure FIPS compliance.
407 The algorithms that are included but not approved are Triple DES ECB,
408 Triple DES CBC and EdDSA.
412 * Added support for KMAC in KBKDF.
416 * RNDR and RNDRRS support in provider functions to provide
417 random number generation for Arm CPUs (aarch64).
421 * s_client and s_server apps now explicitly say when the TLS version
422 does not include the renegotiation mechanism. This avoids confusion
423 between that scenario versus when the TLS version includes secure
424 renegotiation but the peer lacks support for it.
428 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
430 *Tomasz Kantecki, Andrey Matyukov*
432 * The various OBJ_* functions have been made thread safe.
436 * Parallel dual-prime 1536/2048-bit modular exponentiation for
437 AVX512_IFMA capable processors.
439 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
441 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
442 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
443 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
444 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
445 `OPENSSL_NO_DEPRECATED_3_1`.
447 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
448 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
449 definitions for these functions regardless of whether
450 `OPENSSL_NO_DEPRECATED_3_1` is defined.
452 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
453 functions regardless of whether they are using them. It is recommended that
454 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
458 * When generating safe-prime DH parameters set the recommended private key
459 length equivalent to minimum key lengths as in RFC 7919.
463 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
464 maximum size that is smaller or equal to the digest length to comply with
465 FIPS 186-4 section 5. This is implemented by a new option
466 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
467 `rsa_pss_saltlen` parameter, which is now the default. Signature
468 verification is not affected by this change and continues to work as before.
475 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
476 listed here are only a brief description.
477 The migration guide contains more detailed information related to new features,
478 breaking changes, and mappings for the large list of deprecated functions.
480 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
482 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
484 * Fixed NULL dereference during PKCS7 data verification.
486 A NULL pointer can be dereferenced when signatures are being
487 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
488 algorithm used for the signature is known to the OpenSSL library but
489 the implementation of the hash algorithm is not available the digest
490 initialization will fail. There is a missing check for the return
491 value from the initialization function which later leads to invalid
492 usage of the digest API most likely leading to a crash.
495 PKCS7 data is processed by the SMIME library calls and also by the
496 time stamp (TS) library calls. The TLS implementation in OpenSSL does
497 not call these functions however third party applications would be
498 affected if they call these functions to verify signatures on untrusted
503 * Fixed X.400 address type confusion in X.509 GeneralName.
505 There is a type confusion vulnerability relating to X.400 address processing
506 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
507 but the public structure definition for GENERAL_NAME incorrectly specified
508 the type of the x400Address field as ASN1_TYPE. This field is subsequently
509 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
512 When CRL checking is enabled (i.e. the application sets the
513 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
514 pass arbitrary pointers to a memcmp call, enabling them to read memory
515 contents or enact a denial of service.
520 * Fixed NULL dereference validating DSA public key.
522 An invalid pointer dereference on read can be triggered when an
523 application tries to check a malformed DSA public key by the
524 EVP_PKEY_public_check() function. This will most likely lead
525 to an application crash. This function can be called on public
526 keys supplied from untrusted sources which could allow an attacker
527 to cause a denial of service attack.
529 The TLS implementation in OpenSSL does not call this function
530 but applications might call the function if there are additional
531 security requirements imposed by standards such as FIPS 140-3.
534 *Shane Lontis, Tomáš Mráz*
536 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
538 An invalid pointer dereference on read can be triggered when an
539 application tries to load malformed PKCS7 data with the
540 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
542 The result of the dereference is an application crash which could
543 lead to a denial of service attack. The TLS implementation in OpenSSL
544 does not call this function however third party applications might
545 call these functions on untrusted data.
550 * Fixed Use-after-free following BIO_new_NDEF.
552 The public API function BIO_new_NDEF is a helper function used for
553 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
554 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
555 be called directly by end user applications.
557 The function receives a BIO from the caller, prepends a new BIO_f_asn1
558 filter BIO onto the front of it to form a BIO chain, and then returns
559 the new head of the BIO chain to the caller. Under certain conditions,
560 for example if a CMS recipient public key is invalid, the new filter BIO
561 is freed and the function returns a NULL result indicating a failure.
562 However, in this case, the BIO chain is not properly cleaned up and the
563 BIO passed by the caller still retains internal pointers to the previously
564 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
565 then a use-after-free will occur. This will most likely result in a crash.
568 *Viktor Dukhovni, Matt Caswell*
570 * Fixed Double free after calling PEM_read_bio_ex.
572 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
573 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
574 data. If the function succeeds then the "name_out", "header" and "data"
575 arguments are populated with pointers to buffers containing the relevant
576 decoded data. The caller is responsible for freeing those buffers. It is
577 possible to construct a PEM file that results in 0 bytes of payload data.
578 In this case PEM_read_bio_ex() will return a failure code but will populate
579 the header argument with a pointer to a buffer that has already been freed.
580 If the caller also frees this buffer then a double free will occur. This
581 will most likely lead to a crash.
583 The functions PEM_read_bio() and PEM_read() are simple wrappers around
584 PEM_read_bio_ex() and therefore these functions are also directly affected.
586 These functions are also called indirectly by a number of other OpenSSL
587 functions including PEM_X509_INFO_read_bio_ex() and
588 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
589 internal uses of these functions are not vulnerable because the caller does
590 not free the header argument if PEM_read_bio_ex() returns a failure code.
593 *Kurt Roeckx, Matt Caswell*
595 * Fixed Timing Oracle in RSA Decryption.
597 A timing based side channel exists in the OpenSSL RSA Decryption
598 implementation which could be sufficient to recover a plaintext across
599 a network in a Bleichenbacher style attack. To achieve a successful
600 decryption an attacker would have to be able to send a very large number
601 of trial messages for decryption. The vulnerability affects all RSA padding
602 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
605 *Dmitry Belyavsky, Hubert Kario*
607 * Fixed X.509 Name Constraints Read Buffer Overflow.
609 A read buffer overrun can be triggered in X.509 certificate verification,
610 specifically in name constraint checking. The read buffer overrun might
611 result in a crash which could lead to a denial of service attack.
612 In a TLS client, this can be triggered by connecting to a malicious
613 server. In a TLS server, this can be triggered if the server requests
614 client authentication and a malicious client connects.
619 * Fixed X.509 Policy Constraints Double Locking security issue.
621 If an X.509 certificate contains a malformed policy constraint and
622 policy processing is enabled, then a write lock will be taken twice
623 recursively. On some operating systems (most widely: Windows) this
624 results in a denial of service when the affected process hangs. Policy
625 processing being enabled on a publicly facing server is not considered
626 to be a common setup.
631 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
632 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
633 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
634 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
635 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
636 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
637 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
638 for legacy EC and SM2 keys is also changed similarly to honor the
639 equivalent conversion format flag as specified in the underlying
640 `EC_KEY` object being exported to a provider, when this function is
641 called through `EVP_PKEY_export()`.
645 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
647 * Fixed two buffer overflows in punycode decoding functions.
649 A buffer overrun can be triggered in X.509 certificate verification,
650 specifically in name constraint checking. Note that this occurs after
651 certificate chain signature verification and requires either a CA to
652 have signed the malicious certificate or for the application to continue
653 certificate verification despite failure to construct a path to a trusted
656 In a TLS client, this can be triggered by connecting to a malicious
657 server. In a TLS server, this can be triggered if the server requests
658 client authentication and a malicious client connects.
660 An attacker can craft a malicious email address to overflow
661 an arbitrary number of bytes containing the `.` character (decimal 46)
662 on the stack. This buffer overflow could result in a crash (causing a
666 An attacker can craft a malicious email address to overflow four
667 attacker-controlled bytes on the stack. This buffer overflow could
668 result in a crash (causing a denial of service) or potentially remote code
669 execution depending on stack layout for any given platform/compiler.
674 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
675 parameters in OpenSSL code.
676 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
677 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
678 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
679 Using these invalid names may cause algorithms to use slower methods
680 that ignore the CRT parameters.
684 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
689 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
690 data to be signed before signing the certificate.
694 * Added RIPEMD160 to the default provider.
698 * Ensured that the key share group sent or accepted for the key exchange
699 is allowed for the protocol version.
703 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
705 * OpenSSL supports creating a custom cipher via the legacy
706 EVP_CIPHER_meth_new() function and associated function calls. This function
707 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
708 to use the new provider mechanism in order to implement custom ciphers.
710 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
711 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
712 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
713 and decryption initialisation functions). Instead of using the custom cipher
714 directly it incorrectly tries to fetch an equivalent cipher from the
715 available providers. An equivalent cipher is found based on the NID passed to
716 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
717 given cipher. However it is possible for an application to incorrectly pass
718 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
719 is used in this way the OpenSSL encryption/decryption initialisation function
720 will match the NULL cipher as being equivalent and will fetch this from the
721 available providers. This will succeed if the default provider has been
722 loaded (or if a third party provider has been loaded that offers this
723 cipher). Using the NULL cipher means that the plaintext is emitted as the
726 Applications are only affected by this issue if they call
727 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
728 encryption/decryption initialisation function. Applications that only use
729 SSL/TLS are not impacted by this issue.
734 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
739 * Fixed the linux-mips64 Configure target which was missing the
740 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
745 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
750 * Correctly handle a retransmitted ClientHello in DTLS
754 * Fixed detection of ktls support in cross-compile environment on Linux
758 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
763 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
764 report correct results in some cases
768 * Fix UWP builds by defining VirtualLock
772 * For known safe primes use the minimum key length according to RFC 7919.
773 Longer private key sizes unnecessarily raise the cycles needed to compute the
774 shared secret without any increase of the real security. This fixes a
775 regression from 1.1.1 where these shorter keys were generated for the known
780 * Added the loongarch64 target
784 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
785 only passed to the FIPS provider and not to the default or legacy provider.
789 * Fixed reported performance degradation on aarch64. Restored the
790 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
791 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
792 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
793 The new algorithm is still used for 32 bit targets.
797 * Added a missing header for memcmp that caused compilation failure on some
802 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
804 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
805 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
806 This issue makes the RSA implementation with 2048 bit private keys
807 incorrect on such machines and memory corruption will happen during
808 the computation. As a consequence of the memory corruption an attacker
809 may be able to trigger a remote code execution on the machine performing
812 SSL/TLS servers or other servers using 2048 bit RSA private keys running
813 on machines supporting AVX512IFMA instructions of the X86_64 architecture
814 are affected by this issue.
819 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
820 implementation would not encrypt the entirety of the data under some
821 circumstances. This could reveal sixteen bytes of data that was
822 preexisting in the memory that wasn't written. In the special case of
823 "in place" encryption, sixteen bytes of the plaintext would be revealed.
825 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
826 they are both unaffected.
829 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
831 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
833 * In addition to the c_rehash shell command injection identified in
834 CVE-2022-1292, further bugs where the c_rehash script does not
835 properly sanitise shell metacharacters to prevent command injection have been
838 When the CVE-2022-1292 was fixed it was not discovered that there
839 are other places in the script where the file names of certificates
840 being hashed were possibly passed to a command executed through the shell.
842 This script is distributed by some operating systems in a manner where
843 it is automatically executed. On such operating systems, an attacker
844 could execute arbitrary commands with the privileges of the script.
846 Use of the c_rehash script is considered obsolete and should be replaced
847 by the OpenSSL rehash command line tool.
850 *Daniel Fiala, Tomáš Mráz*
852 * Case insensitive string comparison no longer uses locales. It has instead
853 been directly implemented.
857 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
859 * Case insensitive string comparison is reimplemented via new locale-agnostic
860 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
861 comparison. The previous implementation had problems when the Turkish locale
866 * Fixed a bug in the c_rehash script which was not properly sanitising shell
867 metacharacters to prevent command injection. This script is distributed by
868 some operating systems in a manner where it is automatically executed. On
869 such operating systems, an attacker could execute arbitrary commands with the
870 privileges of the script.
872 Use of the c_rehash script is considered obsolete and should be replaced
873 by the OpenSSL rehash command line tool.
878 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
879 certificate on an OCSP response. The bug caused the function in the case
880 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
881 response (meaning a successful verification) even in the case where the
882 response signing certificate fails to verify.
884 It is anticipated that most users of `OCSP_basic_verify` will not use the
885 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
886 a negative value (indicating a fatal error) in the case of a certificate
887 verification failure. The normal expected return value in this case would be
890 This issue also impacts the command line OpenSSL "ocsp" application. When
891 verifying an ocsp response with the "-no_cert_checks" option the command line
892 application will report that the verification is successful even though it
893 has in fact failed. In this case the incorrect successful response will also
894 be accompanied by error messages showing the failure and contradicting the
895 apparently successful result.
900 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
901 AAD data as the MAC key. This made the MAC key trivially predictable.
903 An attacker could exploit this issue by performing a man-in-the-middle attack
904 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
905 that the modified data would still pass the MAC integrity check.
907 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
908 endpoint will always be rejected by the recipient and the connection will
909 fail at that point. Many application protocols require data to be sent from
910 the client to the server first. Therefore, in such a case, only an OpenSSL
911 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
913 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
914 sent in both directions. In this case both clients and servers could be
915 affected, regardless of the application protocol.
917 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
918 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
919 the handshake when using this ciphersuite.
921 The confidentiality of data is not impacted by this issue, i.e. an attacker
922 cannot decrypt data that has been encrypted using this ciphersuite - they can
925 In order for this attack to work both endpoints must legitimately negotiate
926 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
927 OpenSSL 3.0, and is not available within the default provider or the default
928 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
929 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
930 following must have occurred:
932 1) OpenSSL must have been compiled with the (non-default) compile time option
933 enable-weak-ssl-ciphers
935 2) OpenSSL must have had the legacy provider explicitly loaded (either
936 through application code or via configuration)
938 3) The ciphersuite must have been explicitly added to the ciphersuite list
940 4) The libssl security level must have been set to 0 (default is 1)
942 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
944 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
945 others that both endpoints have in common
950 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
951 occupied by the removed hash table entries.
953 This function is used when decoding certificates or keys. If a long lived
954 process periodically decodes certificates or keys its memory usage will
955 expand without bounds and the process might be terminated by the operating
956 system causing a denial of service. Also traversing the empty hash table
957 entries will take increasingly more time.
959 Typically such long lived processes might be TLS clients or TLS servers
960 configured to accept client certificate authentication.
963 *Hugo Landau, Aliaksei Levin*
965 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
966 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
967 statistics are no longer supported. For compatibility, these statistics are
968 still listed in the output but are now always reported as zero.
972 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
974 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
975 for non-prime moduli.
977 Internally this function is used when parsing certificates that contain
978 elliptic curve public keys in compressed form or explicit elliptic curve
979 parameters with a base point encoded in compressed form.
981 It is possible to trigger the infinite loop by crafting a certificate that
982 has invalid explicit curve parameters.
984 Since certificate parsing happens prior to verification of the certificate
985 signature, any process that parses an externally supplied certificate may thus
986 be subject to a denial of service attack. The infinite loop can also be
987 reached when parsing crafted private keys as they can contain explicit
988 elliptic curve parameters.
990 Thus vulnerable situations include:
992 - TLS clients consuming server certificates
993 - TLS servers consuming client certificates
994 - Hosting providers taking certificates or private keys from customers
995 - Certificate authorities parsing certification requests from subscribers
996 - Anything else which parses ASN.1 elliptic curve parameters
998 Also any other applications that use the BN_mod_sqrt() where the attacker
999 can control the parameter values are vulnerable to this DoS issue.
1004 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
1005 to the list of ciphersuites providing Perfect Forward Secrecy as
1006 required by SECLEVEL >= 3.
1008 *Dmitry Belyavskiy, Nicola Tuveri*
1010 * Made the AES constant time code for no-asm configurations
1011 optional due to the resulting 95% performance degradation.
1012 The AES constant time code can be enabled, for no assembly
1013 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1017 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
1022 * The negative return value handling of the certificate verification callback
1023 was reverted. The replacement is to set the verification retry state with
1024 the SSL_set_retry_verify() function.
1028 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
1030 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
1031 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
1032 verify a certificate supplied by a server. That function may return a
1033 negative return value to indicate an internal error (for example out of
1034 memory). Such a negative return value is mishandled by OpenSSL and will cause
1035 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
1036 success and a subsequent call to SSL_get_error() to return the value
1037 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
1038 returned by OpenSSL if the application has previously called
1039 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
1040 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
1041 totally unexpected and applications may not behave correctly as a result. The
1042 exact behaviour will depend on the application but it could result in
1043 crashes, infinite loops or other similar incorrect responses.
1045 This issue is made more serious in combination with a separate bug in OpenSSL
1046 3.0 that will cause X509_verify_cert() to indicate an internal error when
1047 processing a certificate chain. This will occur where a certificate does not
1048 include the Subject Alternative Name extension but where a Certificate
1049 Authority has enforced name constraints. This issue can occur even with valid
1055 * Corrected a few file name and file reference bugs in the build,
1056 installation and setup scripts, which lead to installation verification
1057 failures. Slightly enhanced the installation verification script.
1061 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
1066 * Fixed PVK encoder to properly query for the passphrase.
1070 * Multiple fixes in the OSSL_HTTP API functions.
1074 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
1075 OSSL_PARAM_INTEGER data type and return error on negative numbers
1076 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
1077 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
1081 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
1085 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
1089 * Multiple threading fixes.
1093 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
1097 * Allow fetching an operation from the provider that owns an unexportable key
1098 as a fallback if that is still allowed by the property query.
1102 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
1104 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
1109 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
1110 S390X capability vector to zero. This simplifies testing of different code
1111 paths on S390X architecture.
1115 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
1116 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
1117 SP 800-38D". The communication will fail at this point.
1121 * The EC_GROUP_clear_free() function is deprecated as there is nothing
1122 confidential in EC_GROUP data.
1126 * The byte order mark (BOM) character is ignored if encountered at the
1127 beginning of a PEM-formatted file.
1131 * Added CMS support for the Russian GOST algorithms.
1135 * Due to move of the implementation of cryptographic operations
1136 to the providers, validation of various operation parameters can
1137 be postponed until the actual operation is executed where previously
1138 it happened immediately when an operation parameter was set.
1140 For example when setting an unsupported curve with
1141 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
1142 fail but later keygen operations with the EVP_PKEY_CTX will fail.
1144 *OpenSSL team members and many third party contributors*
1146 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
1147 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
1148 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1149 instead to retrieve these algorithms from a provider.
1153 * On build targets where the multilib postfix is set in the build
1154 configuration the libdir directory was changing based on whether
1155 the lib directory with the multilib postfix exists on the system
1156 or not. This unpredictable behavior was removed and eventual
1157 multilib postfix is now always added to the default libdir. Use
1158 `--libdir=lib` to override the libdir if adding the postfix is
1163 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1164 no longer interoperable with OpenSSL 1.1.1.
1168 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1169 function codes, this function can only cause problems for calling
1174 * Add a configurable flag to output date formats as ISO 8601. Does not
1175 change the default date format.
1179 * Version of MSVC earlier than 1300 could get link warnings, which could
1180 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1181 Support for this flag has been removed.
1185 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1186 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1187 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1188 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1189 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1193 * The signatures of the functions to get and set options on SSL and
1194 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1195 Some source code changes may be required.
1199 * The public definitions of conf_method_st and conf_st have been
1200 deprecated. They will be made opaque in a future release.
1202 *Rich Salz and Tomáš Mráz*
1204 * Client-initiated renegotiation is disabled by default. To allow it, use
1205 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1206 flag, or the "ClientRenegotiation" config parameter as appropriate.
1210 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1211 or modify relative pathname inclusion.
1215 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1216 validated. Please consult the README-FIPS and
1217 README-PROVIDERS files, as well as the migration guide.
1219 *OpenSSL team members and many third party contributors*
1221 * For the key types DH and DHX the allowed settable parameters are now different.
1225 * The openssl commands that read keys, certificates, and CRLs now
1226 automatically detect the PEM or DER format of the input files.
1228 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1230 * Added enhanced PKCS#12 APIs which accept a library context.
1234 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1238 * Added support for Kernel TLS (KTLS).
1240 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1242 * Support for RFC 5746 secure renegotiation is now required by default for
1243 SSL or TLS connections to succeed.
1247 * The signature of the `copy` functional parameter of the
1248 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1249 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1250 the signature of the `pub_decode` functional parameter of the
1251 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1252 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1256 * The error return values from some control calls (ctrl) have changed.
1260 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1264 * Many functions in the EVP_ namespace that are getters of values from
1265 implementations or contexts were renamed to include get or get0 in their
1266 names. Old names are provided as macro aliases for compatibility and
1271 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1272 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1273 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1278 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1281 * The output from the command line applications may have minor
1286 * The output from numerous "printing" may have minor changes.
1290 * Windows thread synchronization uses read/write primitives (SRWLock) when
1291 supported by the OS, otherwise CriticalSection continues to be used.
1295 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1296 work on read only BIO source/sinks that do not support these functions.
1297 This allows piping or redirection of a file BIO using stdin to be buffered
1298 into memory. This is used internally in OSSL_DECODER_from_bio().
1302 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1303 this function would return one of the values OSSL_STORE_INFO_NAME,
1304 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1305 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1306 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1307 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1308 using this function should be amended to handle the changed return value.
1312 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1313 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1314 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1315 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1316 contains more than one certificate identifier: This means that all
1317 certificates referenced there MUST be part of the validation chain.
1321 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1322 RC5, DESX and DES have been moved to the legacy provider.
1326 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1327 RIPEMD-160 have been moved to the legacy provider.
1331 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1336 * The deprecated functions EVP_PKEY_get0_RSA(),
1337 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1338 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1339 well as the similarly named "get1" functions behave differently in
1344 * A number of functions handling low-level keys or engines were deprecated
1345 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1346 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1347 EVP_PKEY_get0_siphash().
1351 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1352 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1353 will need to load the legacy crypto provider. This includes these PBE
1354 algorithms which use this KDF:
1355 - NID_pbeWithMD2AndDES_CBC
1356 - NID_pbeWithMD5AndDES_CBC
1357 - NID_pbeWithSHA1AndRC2_CBC
1358 - NID_pbeWithMD2AndRC2_CBC
1359 - NID_pbeWithMD5AndRC2_CBC
1360 - NID_pbeWithSHA1AndDES_CBC
1364 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1365 BIO_debug_callback() functions.
1369 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1370 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1374 * The RAND_METHOD APIs have been deprecated.
1378 * The SRP APIs have been deprecated.
1382 * Add a compile time option to prevent the caching of provider fetched
1383 algorithms. This is enabled by including the no-cached-fetch option
1384 at configuration time.
1388 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1389 count of PKCS12_DEFAULT_ITER.
1391 *Tomáš Mráz and Sahana Prasad*
1393 * The openssl speed command does not use low-level API calls anymore.
1397 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1400 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1402 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1406 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1407 providers may supply their own group implementations (using either the "key
1408 exchange" or the "key encapsulation" methods) which will automatically be
1409 detected and used by libssl.
1411 *Matt Caswell, Nicola Tuveri*
1413 * The undocumented function X509_certificate_type() has been deprecated;
1417 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1421 * Removed RSA padding mode for SSLv23 (which was only used for
1422 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1423 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1428 * Deprecated the obsolete X9.31 RSA key generation related functions.
1430 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1431 is not allowed to return a value > 1, this is no more taken as failure.
1433 *Viktor Dukhovni and David von Oheimb*
1435 * Deprecated the obsolete X9.31 RSA key generation related functions
1436 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1437 BN_X931_generate_prime_ex().
1441 * The default key generation method for the regular 2-prime RSA keys was
1442 changed to the FIPS 186-4 B.3.6 method.
1446 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1450 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1454 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1455 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1457 *Rich Salz, Richard Levitte, and David von Oheimb*
1459 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1463 * Deprecated `OCSP_parse_url()`.
1467 * Validation of SM2 keys has been separated from the validation of regular EC
1472 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1473 switches: a validation failure triggers an early exit, returning a failure
1474 exit status to the parent process.
1478 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1479 to ignore unknown ciphers.
1483 * The `-cipher-commands` and `-digest-commands` options
1484 of the command line utility `list` have been deprecated.
1485 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1489 * Added convenience functions for generating asymmetric key pairs:
1490 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1491 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1495 * All of the low-level EC_KEY functions have been deprecated.
1497 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1499 * Deprecated all the libcrypto and libssl error string loading
1504 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1505 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1510 * The `-crypt` option to the `passwd` command line tool has been removed.
1514 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1519 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1523 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1524 EVP_PKEY_get1_tls_encodedpoint().
1528 * The security callback, which can be customised by application code, supports
1529 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1530 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1534 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1535 interface. Their functionality remains unchanged.
1539 * Added new option for 'openssl list', '-providers', which will display the
1540 list of loaded providers, their names, version and status. It optionally
1541 displays their gettable parameters.
1545 * Removed EVP_PKEY_set_alias_type().
1549 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1550 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1554 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1555 parameters are still checked for type safety at compile time via helper
1560 * Remove the RAND_DRBG API
1562 *Paul Dale and Matthias St. Pierre*
1564 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1565 as well as actual hostnames.
1569 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1570 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1571 conversely, silently ignore DTLS protocol version bounds when configuring
1572 TLS-based contexts. The commands can be repeated to set bounds of both
1573 types. The same applies with the corresponding "min_protocol" and
1574 "max_protocol" command-line switches, in case some application uses both TLS
1577 SSL_CTX instances that are created for a fixed protocol version (e.g.
1578 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1579 attempts to apply bounds to these protocol versions would result in an
1580 error. Now only the "version-flexible" SSL_CTX instances are subject to
1581 limits in configuration files in command-line options.
1585 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1590 * Reworked the recorded ERR codes to make better space for system errors.
1591 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1592 given code is a system error (true) or an OpenSSL error (false).
1596 * Reworked the test perl framework to better allow parallel testing.
1598 *Nicola Tuveri and David von Oheimb*
1600 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1601 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1605 * 'Configure' has been changed to figure out the configuration target if
1606 none is given on the command line. Consequently, the 'config' script is
1607 now only a mere wrapper. All documentation is changed to only mention
1610 *Rich Salz and Richard Levitte*
1612 * Added a library context `OSSL_LIB_CTX` that applications as well as
1613 other libraries can use to form a separate context within which
1614 libcrypto operations are performed.
1618 * Added various `_ex` functions to the OpenSSL API that support using
1619 a non-default `OSSL_LIB_CTX`.
1623 * Handshake now fails if Extended Master Secret extension is dropped
1628 * Dropped interactive mode from the `openssl` program.
1632 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1634 *David von Oheimb and Shane Lontis*
1636 * Deprecated `EC_METHOD_get_field_type()`.
1640 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1641 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1642 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1646 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1650 * Add CAdES-BES signature verification support, mostly derived
1651 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1653 *Filipe Raimundo da Silva*
1655 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1659 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1660 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1664 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1668 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1669 EC_KEY_precompute_mult().
1673 * Deprecated EC_POINTs_mul().
1677 * Removed FIPS_mode() and FIPS_mode_set().
1681 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1685 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1686 EC_POINT_get_Jprojective_coordinates_GFp().
1690 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1691 arrays to be more easily constructed via a series of utility functions.
1692 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1693 the various push functions and finally convert to a passable OSSL_PARAM
1694 array using OSSL_PARAM_BLD_to_param().
1698 * The security strength of SHA1 and MD5 based signatures in TLS has been
1703 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1704 contain a provider side internal key.
1708 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1712 * Project text documents not yet having a proper file name extension
1713 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1714 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1718 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1719 have been converted to Markdown with the goal to produce documents
1720 which not only look pretty when viewed online in the browser, but
1721 remain well readable inside a plain text editor.
1723 To achieve this goal, a 'minimalistic' Markdown style has been applied
1724 which avoids formatting elements that interfere too much with the
1725 reading flow in the text file. For example, it
1727 * avoids [ATX headings][] and uses [setext headings][] instead
1728 (which works for `<h1>` and `<h2>` headings only).
1729 * avoids [inline links][] and uses [reference links][] instead.
1730 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1732 [ATX headings]: https://github.github.com/gfm/#atx-headings
1733 [setext headings]: https://github.github.com/gfm/#setext-headings
1734 [inline links]: https://github.github.com/gfm/#inline-link
1735 [reference links]: https://github.github.com/gfm/#reference-link
1736 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1737 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1739 *Matthias St. Pierre*
1741 * The test suite is changed to preserve results of each test recipe.
1742 A new directory test-runs/ with subdirectories named like the
1743 test recipes are created in the build tree for this purpose.
1747 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1748 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1749 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1751 *David von Oheimb, Martin Peylo*
1753 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1754 It supports arbitrary request and response content types, GET redirection,
1755 TLS, connections via HTTP(S) proxies, connections and exchange via
1756 user-defined BIOs (allowing implicit connections), persistent connections,
1757 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1758 The legacy OCSP-focused (and only partly documented) API
1759 is retained for backward compatibility, while most of it is deprecated.
1763 * Added `util/check-format.pl`, a tool for checking adherence to the
1764 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1765 The checks performed are incomplete and yield some false positives.
1766 Still the tool should be useful for detecting most typical glitches.
1770 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1771 If domain name resolution yields multiple IP addresses all of them are tried
1772 after `connect()` failures.
1776 * All of the low-level RSA functions have been deprecated.
1780 * X509 certificates signed using SHA1 are no longer allowed at security
1785 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1786 modified to use PKEY APIs. These commands are now in maintenance mode
1787 and no new features will be added to them.
1791 * The command line utility rsautl has been deprecated.
1795 * The command line utilities genrsa and rsa have been modified to use PKEY
1796 APIs. They now write PKCS#8 keys by default. These commands are now in
1797 maintenance mode and no new features will be added to them.
1801 * All of the low-level DH functions have been deprecated.
1803 *Paul Dale and Matt Caswell*
1805 * All of the low-level DSA functions have been deprecated.
1809 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1810 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1814 * Deprecated low-level ECDH and ECDSA functions.
1818 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1822 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1823 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1824 a new formulation to include all the things it can be used for,
1825 as well as words of caution.
1829 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1833 * All of the low-level HMAC functions have been deprecated.
1835 *Paul Dale and David von Oheimb*
1837 * Over two thousand fixes were made to the documentation, including:
1838 - Common options (such as -rand/-writerand, TLS version control, etc)
1839 were refactored and point to newly-enhanced descriptions in openssl.pod.
1840 - Added style conformance for all options (with help from Richard Levitte),
1841 documented all reported missing options, added a CI build to check
1842 that all options are documented and that no unimplemented options
1844 - Documented some internals, such as all use of environment variables.
1845 - Addressed all internal broken L<> references.
1849 * All of the low-level CMAC functions have been deprecated.
1853 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1854 functions have been deprecated.
1856 *Paul Dale and David von Oheimb*
1858 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1859 set of functions. The documentation mentioned negative values for some
1860 errors, but this was never the case, so the mention of negative values
1863 Code that followed the documentation and thereby check with something
1864 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1868 * All of the low-level cipher functions have been deprecated.
1870 *Matt Caswell and Paul Dale*
1872 * Removed include/openssl/opensslconf.h.in and replaced it with
1873 include/openssl/configuration.h.in, which differs in not including
1874 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1875 was added to include both.
1877 This allows internal hacks where one might need to modify the set
1878 of configured macros, for example this if deprecated symbols are
1879 still supposed to be available internally:
1881 #include <openssl/configuration.h>
1883 #undef OPENSSL_NO_DEPRECATED
1884 #define OPENSSL_SUPPRESS_DEPRECATED
1886 #include <openssl/macros.h>
1888 This should not be used by applications that use the exported
1889 symbols, as that will lead to linking errors.
1893 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1894 used in exponentiation with 512-bit moduli. No EC algorithms are
1895 affected. Analysis suggests that attacks against 2-prime RSA1024,
1896 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1897 difficult to perform and are not believed likely. Attacks against DH512
1898 are considered just feasible. However, for an attack the target would
1899 have to reuse the DH512 private key, which is not recommended anyway.
1900 Also applications directly using the low-level API BN_mod_exp may be
1901 affected if they use BN_FLG_CONSTTIME.
1906 * Most memory-debug features have been deprecated, and the functionality
1907 replaced with no-ops.
1911 * Added documentation for the STACK API.
1915 * Introduced a new method type and API, OSSL_ENCODER, to represent
1916 generic encoders. These do the same sort of job that PEM writers
1917 and d2i functions do, but with support for methods supplied by
1918 providers, and the possibility for providers to support other
1923 * Introduced a new method type and API, OSSL_DECODER, to represent
1924 generic decoders. These do the same sort of job that PEM readers
1925 and i2d functions do, but with support for methods supplied by
1926 providers, and the possibility for providers to support other
1931 * Added a .pragma directive to the syntax of configuration files, to
1932 allow varying behavior in a supported and predictable manner.
1933 Currently added pragma:
1937 This allows dollar signs to be a keyword character unless it's
1938 followed by a opening brace or parenthesis. This is useful for
1939 platforms where dollar signs are commonly used in names, such as
1940 volume names and system directory names on VMS.
1944 * Added functionality to create an EVP_PKEY from user data.
1948 * Change the interpretation of the '--api' configuration option to
1949 mean that this is a desired API compatibility level with no
1950 further meaning. The previous interpretation, that this would
1951 also mean to remove all deprecated symbols up to and including
1952 the given version, no requires that 'no-deprecated' is also used
1953 in the configuration.
1955 When building applications, the desired API compatibility level
1956 can be set with the OPENSSL_API_COMPAT macro like before. For
1957 API compatibility version below 3.0, the old style numerical
1958 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1959 For version 3.0 and on, the value is expected to be the decimal
1960 value calculated from the major and minor version like this:
1962 MAJOR * 10000 + MINOR * 100
1966 -DOPENSSL_API_COMPAT=30000 For 3.0
1967 -DOPENSSL_API_COMPAT=30200 For 3.2
1969 To hide declarations that are deprecated up to and including the
1970 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1971 given when building the application as well.
1975 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1976 access to certificate and CRL stores via URIs and OSSL_STORE
1979 This adds the following functions:
1981 - X509_LOOKUP_store()
1982 - X509_STORE_load_file()
1983 - X509_STORE_load_path()
1984 - X509_STORE_load_store()
1985 - SSL_add_store_cert_subjects_to_stack()
1986 - SSL_CTX_set_default_verify_store()
1987 - SSL_CTX_load_verify_file()
1988 - SSL_CTX_load_verify_dir()
1989 - SSL_CTX_load_verify_store()
1993 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1994 The presence of this system service is determined at run-time.
1998 * Added functionality to create an EVP_PKEY context based on data
1999 for methods from providers. This takes an algorithm name and a
2000 property query string and simply stores them, with the intent
2001 that any operation that uses this context will use those strings
2002 to fetch the needed methods implicitly, thereby making the port
2003 of application written for pre-3.0 OpenSSL easier.
2007 * The undocumented function NCONF_WIN32() has been deprecated; for
2008 conversion details see the HISTORY section of doc/man5/config.pod
2012 * Introduced the new functions EVP_DigestSignInit_ex() and
2013 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
2014 EVP_DigestVerifyUpdate() have been converted to functions. See the man
2015 pages for further details.
2019 * Over two thousand fixes were made to the documentation, including:
2020 adding missing command flags, better style conformance, documentation
2023 *Rich Salz, Richard Levitte*
2025 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
2026 X25519, X448, Ed25519 and Ed448.
2030 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2035 * Deprecated the public definition of `ERR_STATE` as well as the function
2036 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
2041 * Added ERR functionality to give callers access to the stored function
2042 names that have replaced the older function code based functions.
2044 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
2045 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
2046 ERR_peek_error_all() and ERR_peek_last_error_all().
2048 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
2049 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
2050 ERR_func_error_string().
2054 * Extended testing to be verbose for failing tests only. The make variables
2055 VERBOSE_FAILURE or VF can be used to enable this:
2057 $ make VF=1 test # Unix
2058 $ mms /macro=(VF=1) test ! OpenVMS
2059 $ nmake VF=1 test # Windows
2063 * Added the `-copy_extensions` option to the `x509` command for use with
2064 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
2065 all extensions in the request are copied to the certificate or vice versa.
2067 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
2069 * Added the `-copy_extensions` option to the `req` command for use with
2070 `-x509`. When given with the `copy` or `copyall` argument,
2071 all extensions in the certification request are copied to the certificate.
2075 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
2076 they generate are by default RFC 5280 compliant in the following sense:
2077 There is a subjectKeyIdentifier extension with a hash value of the public key
2078 and for not self-signed certs there is an authorityKeyIdentifier extension
2079 with a keyIdentifier field or issuer information identifying the signing key.
2080 This is done unless some configuration overrides the new default behavior,
2081 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
2085 * Added several checks to `X509_verify_cert()` according to requirements in
2086 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
2087 (which may be done by using the CLI option `-x509_strict`):
2088 * The basicConstraints of CA certificates must be marked critical.
2089 * CA certificates must explicitly include the keyUsage extension.
2090 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
2091 * The issuer name of any certificate must not be empty.
2092 * The subject name of CA certs, certs with keyUsage crlSign,
2093 and certs without subjectAlternativeName must not be empty.
2094 * If a subjectAlternativeName extension is given it must not be empty.
2095 * The signatureAlgorithm field and the cert signature must be consistent.
2096 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
2097 must not be marked critical.
2098 * The authorityKeyIdentifier must be given for X.509v3 certs
2099 unless they are self-signed.
2100 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
2104 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
2105 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
2109 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2110 used even when parsing explicit parameters, when loading a encoded key
2111 or calling `EC_GROUP_new_from_ecpkparameters()`/
2112 `EC_GROUP_new_from_ecparameters()`.
2113 This prevents bypass of security hardening and performance gains,
2114 especially for curves with specialized EC_METHODs.
2115 By default, if a key encoded with explicit parameters is loaded and later
2116 encoded, the output is still encoded with explicit parameters, even if
2117 internally a "named" EC_GROUP is used for computation.
2121 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2122 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2123 NULL. After this change, only the cofactor parameter can be NULL. It also
2124 does some minimal sanity checks on the passed order.
2129 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2130 An attack is simple, if the first CMS_recipientInfo is valid but the
2131 second CMS_recipientInfo is chosen ciphertext. If the second
2132 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2133 encryption key will be replaced by garbage, and the message cannot be
2134 decoded, but if the RSA decryption fails, the correct encryption key is
2135 used and the recipient will not notice the attack.
2136 As a work around for this potential attack the length of the decrypted
2137 key must be equal to the cipher default key length, in case the
2138 certificate is not given and all recipientInfo are tried out.
2139 The old behaviour can be re-enabled in the CMS code by setting the
2140 CMS_DEBUG_DECRYPT flag.
2144 * Early start up entropy quality from the DEVRANDOM seed source has been
2145 improved for older Linux systems. The RAND subsystem will wait for
2146 /dev/random to be producing output before seeding from /dev/urandom.
2147 The seeded state is stored for future library initialisations using
2148 a system global shared memory segment. The shared memory identifier
2149 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2150 the desired value. The default identifier is 114.
2154 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2155 when primes for RSA keys are computed.
2156 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2157 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2158 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2159 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2160 This avoids possible fingerprinting of newly generated RSA modules.
2164 * Correct the extended master secret constant on EBCDIC systems. Without this
2165 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2166 negotiate EMS will fail. Unfortunately this also means that TLS connections
2167 between EBCDIC systems with this fix, and EBCDIC systems without this
2168 fix will fail if they negotiate EMS.
2172 * Changed the library initialisation so that the config file is now loaded
2173 by default. This was already the case for libssl. It now occurs for both
2174 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2175 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2179 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2180 where the former acts as a replacement for `ERR_put_error()`, and the
2181 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2182 `ERR_raise_data()` adds more flexibility by taking a format string and
2183 an arbitrary number of arguments following it, to be processed with
2188 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2189 to check if a named provider is loaded and available. When called, it
2190 will also activate all fallback providers if such are still present.
2194 * Enforce a minimum DH modulus size of 512 bits.
2198 * Changed DH parameters to generate the order q subgroup instead of 2q.
2199 Previously generated DH parameters are still accepted by DH_check
2200 but DH_generate_key works around that by clearing bit 0 of the
2201 private key for those. This avoids leaking bit 0 of the private key.
2205 * Significantly reduce secure memory usage by the randomness pools.
2209 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2214 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2215 algorithms. An implementation of a key exchange algorithm can be obtained
2216 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2217 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2218 the older EVP_PKEY_derive_init() function. See the man pages for the new
2219 functions for further details.
2223 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2227 * Removed the function names from error messages and deprecated the
2232 * Removed NextStep support and the macro OPENSSL_UNISTD
2236 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2237 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2238 Also removed "export var as function" capability; we do not export
2239 variables, only functions.
2243 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2244 an error and 1 indicating success. In previous versions of OpenSSL this
2245 was a void type. If a key was set longer than the maximum possible this
2250 * Support SM2 signing and verification schemes with X509 certificate.
2254 * Use SHA256 as the default digest for TS query in the `ts` app.
2258 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2262 * Default cipher lists/suites are now available via a function, the
2263 #defines are deprecated.
2267 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2268 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2269 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2273 * Join the directories crypto/x509 and crypto/x509v3
2277 * Added command 'openssl kdf' that uses the EVP_KDF API.
2281 * Added command 'openssl mac' that uses the EVP_MAC API.
2285 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2286 as default directories. Also added the command 'openssl info'
2287 for scripting purposes.
2291 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2296 * Add prediction resistance to the DRBG reseeding process.
2300 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2301 mandated by IEEE Std 1619-2018.
2305 * Added newline escaping functionality to a filename when using openssl dgst.
2306 This output format is to replicate the output format found in the `*sum`
2307 checksum programs. This aims to preserve backward compatibility.
2309 *Matt Eaton, Richard Levitte, and Paul Dale*
2311 * Removed the heartbeat message in DTLS feature, as it has very
2312 little usage and doesn't seem to fulfill a valuable purpose.
2313 The configuration option is now deprecated.
2317 * Changed the output of 'openssl {digestname} < file' to display the
2318 digest name in its output.
2322 * Added a new generic trace API which provides support for enabling
2323 instrumentation through trace output.
2325 *Richard Levitte & Matthias St. Pierre*
2327 * Added build tests for C++. These are generated files that only do one
2328 thing, to include one public OpenSSL head file each. This tests that
2329 the public header files can be usefully included in a C++ application.
2331 This test isn't enabled by default. It can be enabled with the option
2332 'enable-buildtest-c++'.
2336 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2340 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2344 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2348 * Added KMAC to EVP_MAC.
2352 * Added property based algorithm implementation selection framework to
2357 * Added SCA hardening for modular field inversion in EC_GROUP through
2358 a new dedicated field_inv() pointer in EC_METHOD.
2359 This also addresses a leakage affecting conversions from projective
2360 to affine coordinates.
2362 *Billy Bob Brumley, Nicola Tuveri*
2364 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2365 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2366 those algorithms that were already supported through the EVP_PKEY API
2367 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2368 and scrypt are now wrappers that call EVP_KDF.
2372 * Build devcrypto engine as a dynamic engine.
2374 *Eneas U de Queiroz*
2376 * Add keyed BLAKE2 to EVP_MAC.
2380 * Fix a bug in the computation of the endpoint-pair shared secret used
2381 by DTLS over SCTP. This breaks interoperability with older versions
2382 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2383 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2384 interoperability with such broken implementations. However, enabling
2385 this switch breaks interoperability with correct implementations.
2387 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2388 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2392 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2396 * Changed the license to the Apache License v2.0.
2400 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2402 - Major releases (indicated by incrementing the MAJOR release number)
2403 may introduce incompatible API/ABI changes.
2404 - Minor releases (indicated by incrementing the MINOR release number)
2405 may introduce new features but retain API/ABI compatibility.
2406 - Patch releases (indicated by incrementing the PATCH number)
2407 are intended for bug fixes and other improvements of existing
2408 features only (like improving performance or adding documentation)
2409 and retain API/ABI compatibility.
2413 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2417 * Remove the 'dist' target and add a tarball building script. The
2418 'dist' target has fallen out of use, and it shouldn't be
2419 necessary to configure just to create a source distribution.
2423 * Recreate the OS390-Unix config target. It no longer relies on a
2424 special script like it did for OpenSSL pre-1.1.0.
2428 * Instead of having the source directories listed in Configure, add
2429 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2434 * Add GMAC to EVP_MAC.
2438 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2442 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2443 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2444 to facilitate the continued use of MACs through raw private keys in
2445 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2449 * Deprecate ECDH_KDF_X9_62().
2453 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2454 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2455 are retained for backwards compatibility.
2459 * AES-XTS mode now enforces that its two keys are different to mitigate
2460 the attacked described in "Efficient Instantiations of Tweakable
2461 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2462 Details of this attack can be obtained from:
2463 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2467 * Rename the object files, i.e. give them other names than in previous
2468 versions. Their names now include the name of the final product, as
2469 well as its type mnemonic (bin, lib, shlib).
2473 * Added new option for 'openssl list', '-objects', which will display the
2474 list of built in objects, i.e. OIDs with names.
2478 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2479 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2484 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2485 improves application performance by removing data copies and providing
2486 applications with zero-copy system calls such as sendfile and splice.
2490 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2494 * `PKCS12_parse` now maintains the order of the parsed certificates
2495 when outputting them via `*ca` (rather than reversing it).
2499 * Deprecated pthread fork support methods.
2503 * Added support for FFDHE key exchange in TLS 1.3.
2507 * Added a new concept for OpenSSL plugability: providers. This
2508 functionality is designed to replace the ENGINE API and ENGINE
2509 implementations, and to be much more dynamic, allowing provider
2510 authors to introduce new algorithms among other things, as long as
2511 there's an API that supports the algorithm type.
2513 With this concept comes a new core API for interaction between
2514 libcrypto and provider implementations. Public libcrypto functions
2515 that want to use providers do so through this core API.
2517 The main documentation for this core API is found in
2518 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2519 refer to other manuals describing the API specific for supported
2520 algorithm types (also called operations).
2527 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2529 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2531 * Avoid loading of a dynamic engine twice.
2535 * Prioritise DANE TLSA issuer certs over peer certs
2539 * Fixed random API for MacOS prior to 10.12
2541 These MacOS versions don't support the CommonCrypto APIs
2545 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2547 * Fixed an SM2 Decryption Buffer Overflow.
2549 In order to decrypt SM2 encrypted data an application is expected to
2550 call the API function EVP_PKEY_decrypt(). Typically an application will
2551 call this function twice. The first time, on entry, the "out" parameter
2552 can be NULL and, on exit, the "outlen" parameter is populated with the
2553 buffer size required to hold the decrypted plaintext. The application
2554 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2555 again, but this time passing a non-NULL value for the "out" parameter.
2557 A bug in the implementation of the SM2 decryption code means that the
2558 calculation of the buffer size required to hold the plaintext returned
2559 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2560 size required by the second call. This can lead to a buffer overflow
2561 when EVP_PKEY_decrypt() is called by the application a second time with
2562 a buffer that is too small.
2564 A malicious attacker who is able present SM2 content for decryption to
2565 an application could cause attacker chosen data to overflow the buffer
2566 by up to a maximum of 62 bytes altering the contents of other data held
2567 after the buffer, possibly changing application behaviour or causing
2568 the application to crash. The location of the buffer is application
2569 dependent but is typically heap allocated.
2574 * Fixed various read buffer overruns processing ASN.1 strings
2576 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2577 structure which contains a buffer holding the string data and a field
2578 holding the buffer length. This contrasts with normal C strings which
2579 are represented as a buffer for the string data which is terminated
2580 with a NUL (0) byte.
2582 Although not a strict requirement, ASN.1 strings that are parsed using
2583 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2584 well as any string whose value has been set with the ASN1_STRING_set()
2585 function will additionally NUL terminate the byte array in the
2586 ASN1_STRING structure.
2588 However, it is possible for applications to directly construct valid
2589 ASN1_STRING structures which do not NUL terminate the byte array by
2590 directly setting the "data" and "length" fields in the ASN1_STRING
2591 array. This can also happen by using the ASN1_STRING_set0() function.
2593 Numerous OpenSSL functions that print ASN.1 data have been found to
2594 assume that the ASN1_STRING byte array will be NUL terminated, even
2595 though this is not guaranteed for strings that have been directly
2596 constructed. Where an application requests an ASN.1 structure to be
2597 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2598 been directly constructed by the application without NUL terminating
2599 the "data" field, then a read buffer overrun can occur.
2601 The same thing can also occur during name constraints processing
2602 of certificates (for example if a certificate has been directly
2603 constructed by the application instead of loading it via the OpenSSL
2604 parsing functions, and the certificate contains non NUL terminated
2605 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2606 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2608 If a malicious actor can cause an application to directly construct an
2609 ASN1_STRING and then process it through one of the affected OpenSSL
2610 functions then this issue could be hit. This might result in a crash
2611 (causing a Denial of Service attack). It could also result in the
2612 disclosure of private memory contents (such as private keys, or
2613 sensitive plaintext).
2618 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2620 * Fixed a problem with verifying a certificate chain when using the
2621 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2622 the certificates present in a certificate chain. It is not set by default.
2624 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2625 the chain that have explicitly encoded elliptic curve parameters was added
2626 as an additional strict check.
2628 An error in the implementation of this check meant that the result of a
2629 previous check to confirm that certificates in the chain are valid CA
2630 certificates was overwritten. This effectively bypasses the check
2631 that non-CA certificates must not be able to issue other certificates.
2633 If a "purpose" has been configured then there is a subsequent opportunity
2634 for checks that the certificate is a valid CA. All of the named "purpose"
2635 values implemented in libcrypto perform this check. Therefore, where
2636 a purpose is set the certificate chain will still be rejected even when the
2637 strict flag has been used. A purpose is set by default in libssl client and
2638 server certificate verification routines, but it can be overridden or
2639 removed by an application.
2641 In order to be affected, an application must explicitly set the
2642 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2643 for the certificate verification or, in the case of TLS client or server
2644 applications, override the default purpose.
2649 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2650 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2651 renegotiation ClientHello omits the signature_algorithms extension (where it
2652 was present in the initial ClientHello), but includes a
2653 signature_algorithms_cert extension then a NULL pointer dereference will
2654 result, leading to a crash and a denial of service attack.
2656 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2657 (which is the default configuration). OpenSSL TLS clients are not impacted by
2661 *Peter Kästle and Samuel Sapalski*
2663 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2665 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2666 create a unique hash value based on the issuer and serial number data
2667 contained within an X509 certificate. However, it was failing to correctly
2668 handle any errors that may occur while parsing the issuer field (which might
2669 occur if the issuer field is maliciously constructed). This may subsequently
2670 result in a NULL pointer deref and a crash leading to a potential denial of
2676 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2677 padding mode to correctly check for rollback attacks. This is considered a
2678 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2683 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2684 functions. Previously they could overflow the output length argument in some
2685 cases where the input length is close to the maximum permissible length for
2686 an integer on the platform. In such cases the return value from the function
2687 call would be 1 (indicating success), but the output length value would be
2688 negative. This could cause applications to behave incorrectly or crash.
2693 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2694 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2695 could be exploited in a side channel attack to recover the password. Since
2696 the attack is local host only this is outside of the current OpenSSL
2697 threat model and therefore no CVE is assigned.
2699 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2704 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2706 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2707 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2708 If an attacker can control both items being compared then this could lead
2709 to a possible denial of service attack. OpenSSL itself uses the
2710 GENERAL_NAME_cmp function for two purposes:
2711 1) Comparing CRL distribution point names between an available CRL and a
2712 CRL distribution point embedded in an X509 certificate
2713 2) When verifying that a timestamp response token signer matches the
2714 timestamp authority name (exposed via the API functions
2715 TS_RESP_verify_response and TS_RESP_verify_token)
2720 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2722 * Certificates with explicit curve parameters are now disallowed in
2723 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2727 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2728 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2729 conversely, silently ignore DTLS protocol version bounds when configuring
2730 TLS-based contexts. The commands can be repeated to set bounds of both
2731 types. The same applies with the corresponding "min_protocol" and
2732 "max_protocol" command-line switches, in case some application uses both TLS
2735 SSL_CTX instances that are created for a fixed protocol version (e.g.
2736 TLSv1_server_method()) also silently ignore version bounds. Previously
2737 attempts to apply bounds to these protocol versions would result in an
2738 error. Now only the "version-flexible" SSL_CTX instances are subject to
2739 limits in configuration files in command-line options.
2743 * Handshake now fails if Extended Master Secret extension is dropped
2748 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2750 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2752 * Fixed segmentation fault in SSL_check_chain()
2753 Server or client applications that call the SSL_check_chain() function
2754 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2755 dereference as a result of incorrect handling of the
2756 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2757 or unrecognised signature algorithm is received from the peer. This could
2758 be exploited by a malicious peer in a Denial of Service attack.
2763 * Added AES consttime code for no-asm configurations
2764 an optional constant time support for AES was added
2765 when building openssl for no-asm.
2766 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2767 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2768 At this time this feature is by default disabled.
2769 It will be enabled by default in 3.0.
2773 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2775 * Revert the change of EOF detection while reading in libssl to avoid
2776 regressions in applications depending on the current way of reporting
2777 the EOF. As the existing method is not fully accurate the change to
2778 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2779 branch and will be present in the 3.0 release.
2783 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2784 when primes for RSA keys are computed.
2785 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2786 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2787 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2788 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2789 This avoids possible fingerprinting of newly generated RSA modules.
2793 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2795 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2796 while reading in libssl then we would report an error back to the
2797 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2798 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2799 therefore give a hint as to what went wrong.
2803 * Check that ed25519 and ed448 are allowed by the security level. Previously
2804 signature algorithms not using an MD were not being checked that they were
2805 allowed by the security level.
2809 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2810 was not quite right. The behaviour was not consistent between resumption
2811 and normal handshakes, and also not quite consistent with historical
2812 behaviour. The behaviour in various scenarios has been clarified and
2813 it has been updated to make it match historical behaviour as closely as
2818 * *[VMS only]* The header files that the VMS compilers include automatically,
2819 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2820 that the C++ compiler doesn't understand. This is a shortcoming in the
2821 compiler, but can be worked around with `__cplusplus` guards.
2823 C++ applications that use OpenSSL libraries must be compiled using the
2824 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2825 functions. Otherwise, only functions with symbols of less than 31
2826 characters can be used, as the linker will not be able to successfully
2827 resolve symbols with longer names.
2831 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2832 The presence of this system service is determined at run-time.
2836 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2841 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2843 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2844 number generator (RNG). This was intended to include protection in the
2845 event of a fork() system call in order to ensure that the parent and child
2846 processes did not share the same RNG state. However, this protection was not
2847 being used in the default case.
2849 A partial mitigation for this issue is that the output from a high
2850 precision timer is mixed into the RNG state so the likelihood of a parent
2851 and child process sharing state is significantly reduced.
2853 If an application already calls OPENSSL_init_crypto() explicitly using
2854 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2857 *Matthias St. Pierre*
2859 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2860 used even when parsing explicit parameters, when loading a encoded key
2861 or calling `EC_GROUP_new_from_ecpkparameters()`/
2862 `EC_GROUP_new_from_ecparameters()`.
2863 This prevents bypass of security hardening and performance gains,
2864 especially for curves with specialized EC_METHODs.
2865 By default, if a key encoded with explicit parameters is loaded and later
2866 encoded, the output is still encoded with explicit parameters, even if
2867 internally a "named" EC_GROUP is used for computation.
2871 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2872 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2873 NULL. After this change, only the cofactor parameter can be NULL. It also
2874 does some minimal sanity checks on the passed order.
2879 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2880 An attack is simple, if the first CMS_recipientInfo is valid but the
2881 second CMS_recipientInfo is chosen ciphertext. If the second
2882 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2883 encryption key will be replaced by garbage, and the message cannot be
2884 decoded, but if the RSA decryption fails, the correct encryption key is
2885 used and the recipient will not notice the attack.
2886 As a work around for this potential attack the length of the decrypted
2887 key must be equal to the cipher default key length, in case the
2888 certificate is not given and all recipientInfo are tried out.
2889 The old behaviour can be re-enabled in the CMS code by setting the
2890 CMS_DEBUG_DECRYPT flag.
2895 * Early start up entropy quality from the DEVRANDOM seed source has been
2896 improved for older Linux systems. The RAND subsystem will wait for
2897 /dev/random to be producing output before seeding from /dev/urandom.
2898 The seeded state is stored for future library initialisations using
2899 a system global shared memory segment. The shared memory identifier
2900 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2901 the desired value. The default identifier is 114.
2905 * Correct the extended master secret constant on EBCDIC systems. Without this
2906 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2907 negotiate EMS will fail. Unfortunately this also means that TLS connections
2908 between EBCDIC systems with this fix, and EBCDIC systems without this
2909 fix will fail if they negotiate EMS.
2913 * Use Windows installation paths in the mingw builds
2915 Mingw isn't a POSIX environment per se, which means that Windows
2916 paths should be used for installation.
2921 * Changed DH_check to accept parameters with order q and 2q subgroups.
2922 With order 2q subgroups the bit 0 of the private key is not secret
2923 but DH_generate_key works around that by clearing bit 0 of the
2924 private key for those. This avoids leaking bit 0 of the private key.
2928 * Significantly reduce secure memory usage by the randomness pools.
2932 * Revert the DEVRANDOM_WAIT feature for Linux systems
2934 The DEVRANDOM_WAIT feature added a select() call to wait for the
2935 /dev/random device to become readable before reading from the
2936 /dev/urandom device.
2938 It turned out that this change had negative side effects on
2939 performance which were not acceptable. After some discussion it
2940 was decided to revert this feature and leave it up to the OS
2941 resp. the platform maintainer to ensure a proper initialization
2942 during early boot time.
2944 *Matthias St. Pierre*
2946 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2948 * Add build tests for C++. These are generated files that only do one
2949 thing, to include one public OpenSSL head file each. This tests that
2950 the public header files can be usefully included in a C++ application.
2952 This test isn't enabled by default. It can be enabled with the option
2953 'enable-buildtest-c++'.
2957 * Enable SHA3 pre-hashing for ECDSA and DSA.
2961 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2962 This changes the size when using the `genpkey` command when no size is given.
2963 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2964 generation commands to use 2048 bits by default.
2968 * Reorganize the manual pages to consistently have RETURN VALUES,
2969 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2970 util/fix-doc-nits accordingly.
2972 *Paul Yang, Joshua Lock*
2974 * Add the missing accessor EVP_PKEY_get0_engine()
2978 * Have commands like `s_client` and `s_server` output the signature scheme
2979 along with other cipher suite parameters when debugging.
2983 * Make OPENSSL_config() error agnostic again.
2987 * Do the error handling in RSA decryption constant time.
2991 * Prevent over long nonces in ChaCha20-Poly1305.
2993 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2994 for every encryption operation. RFC 7539 specifies that the nonce value
2995 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2996 and front pads the nonce with 0 bytes if it is less than 12
2997 bytes. However it also incorrectly allows a nonce to be set of up to 16
2998 bytes. In this case only the last 12 bytes are significant and any
2999 additional leading bytes are ignored.
3001 It is a requirement of using this cipher that nonce values are
3002 unique. Messages encrypted using a reused nonce value are susceptible to
3003 serious confidentiality and integrity attacks. If an application changes
3004 the default nonce length to be longer than 12 bytes and then makes a
3005 change to the leading bytes of the nonce expecting the new value to be a
3006 new unique nonce then such an application could inadvertently encrypt
3007 messages with a reused nonce.
3009 Additionally the ignored bytes in a long nonce are not covered by the
3010 integrity guarantee of this cipher. Any application that relies on the
3011 integrity of these ignored leading bytes of a long nonce may be further
3012 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3013 is safe because no such use sets such a long nonce value. However user
3014 applications that use this cipher directly and set a non-default nonce
3015 length to be longer than 12 bytes may be vulnerable.
3017 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3023 * Add DEVRANDOM_WAIT feature for Linux systems
3025 On older Linux systems where the getrandom() system call is not available,
3026 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
3027 Contrary to getrandom(), the /dev/urandom device will not block during
3028 early boot when the kernel CSPRNG has not been seeded yet.
3030 To mitigate this known weakness, use select() to wait for /dev/random to
3031 become readable before reading from /dev/urandom.
3033 * Ensure that SM2 only uses SM3 as digest algorithm
3037 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
3039 * Change the info callback signals for the start and end of a post-handshake
3040 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
3041 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
3042 confused by this and assume that a TLSv1.2 renegotiation has started. This
3043 can break KeyUpdate handling. Instead we no longer signal the start and end
3044 of a post handshake message exchange (although the messages themselves are
3045 still signalled). This could break some applications that were expecting
3046 the old signals. However without this KeyUpdate is not usable for many
3051 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
3053 * Timing vulnerability in DSA signature generation
3055 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3056 timing side channel attack. An attacker could use variations in the signing
3057 algorithm to recover the private key.
3059 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3064 * Timing vulnerability in ECDSA signature generation
3066 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3067 timing side channel attack. An attacker could use variations in the signing
3068 algorithm to recover the private key.
3070 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3075 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
3076 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
3077 of two gigabytes and the error handling improved.
3079 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
3080 categorized as a normal bug, not a security issue, because the DRBG reseeds
3081 automatically and is fully functional even without additional randomness
3082 provided by the application.
3084 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
3086 * Add a new ClientHello callback. Provides a callback interface that gives
3087 the application the ability to adjust the nascent SSL object at the
3088 earliest stage of ClientHello processing, immediately after extensions have
3089 been collected but before they have been processed. In particular, this
3090 callback can adjust the supported TLS versions in response to the contents
3095 * Add SM2 base algorithm support.
3099 * s390x assembly pack: add (improved) hardware-support for the following
3100 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
3101 aes-cfb/cfb8, aes-ecb.
3105 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3106 parameter is no longer accepted, as it leads to a corrupt table. NULL
3107 pem_str is reserved for alias entries only.
3111 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3112 step for prime curves. The new implementation is based on formulae from
3113 differential addition-and-doubling in homogeneous projective coordinates
3114 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
3115 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
3116 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
3117 to work in projective coordinates.
3119 *Billy Bob Brumley, Nicola Tuveri*
3121 * Change generating and checking of primes so that the error rate of not
3122 being prime depends on the intended use based on the size of the input.
3123 For larger primes this will result in more rounds of Miller-Rabin.
3124 The maximal error rate for primes with more than 1080 bits is lowered
3127 *Kurt Roeckx, Annie Yousar*
3129 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3133 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
3134 moving between systems, and to avoid confusion when a Windows build is
3135 done with mingw vs with MSVC. For POSIX installs, there's still a
3136 symlink or copy named 'tsget' to avoid that confusion as well.
3140 * Revert blinding in ECDSA sign and instead make problematic addition
3141 length-invariant. Switch even to fixed-length Montgomery multiplication.
3145 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3146 step for binary curves. The new implementation is based on formulae from
3147 differential addition-and-doubling in mixed Lopez-Dahab projective
3148 coordinates, modified to independently blind the operands.
3150 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3152 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3153 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3154 EC_METHODs to implement their own specialized "ladder step", to take
3155 advantage of more favorable coordinate systems or more efficient
3156 differential addition-and-doubling algorithms.
3158 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3160 * Modified the random device based seed sources to keep the relevant
3161 file descriptors open rather than reopening them on each access.
3162 This allows such sources to operate in a chroot() jail without
3163 the associated device nodes being available. This behaviour can be
3164 controlled using RAND_keep_random_devices_open().
3168 * Numerous side-channel attack mitigations have been applied. This may have
3169 performance impacts for some algorithms for the benefit of improved
3170 security. Specific changes are noted in this change log by their respective
3175 * AIX shared library support overhaul. Switch to AIX "natural" way of
3176 handling shared libraries, which means collecting shared objects of
3177 different versions and bitnesses in one common archive. This allows to
3178 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3179 doesn't affect the way 3rd party applications are linked, only how
3180 multi-version installation is managed.
3184 * Make ec_group_do_inverse_ord() more robust and available to other
3185 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3186 mitigations are applied to the fallback BN_mod_inverse().
3187 When using this function rather than BN_mod_inverse() directly, new
3188 EC cryptosystem implementations are then safer-by-default.
3192 * Add coordinate blinding for EC_POINT and implement projective
3193 coordinate blinding for generic prime curves as a countermeasure to
3194 chosen point SCA attacks.
3196 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3198 * Add blinding to ECDSA and DSA signatures to protect against side channel
3199 attacks discovered by Keegan Ryan (NCC Group).
3203 * Enforce checking in the `pkeyutl` command to ensure that the input
3204 length does not exceed the maximum supported digest length when performing
3205 a sign, verify or verifyrecover operation.
3209 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3210 I/O in combination with something like select() or poll() will hang. This
3211 can be turned off again using SSL_CTX_clear_mode().
3212 Many applications do not properly handle non-application data records, and
3213 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3214 around the problems in those applications, but can also break some.
3215 It's recommended to read the manpages about SSL_read(), SSL_write(),
3216 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3217 SSL_CTX_set_read_ahead() again.
3221 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3222 now allow empty (zero character) pass phrases.
3226 * Apply blinding to binary field modular inversion and remove patent
3227 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3231 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3232 binary and prime elliptic curves.
3236 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3237 constant time fixed point multiplication.
3241 * Revise elliptic curve scalar multiplication with timing attack
3242 defenses: ec_wNAF_mul redirects to a constant time implementation
3243 when computing fixed point and variable point multiplication (which
3244 in OpenSSL are mostly used with secret scalars in keygen, sign,
3245 ECDH derive operations).
3246 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3249 * Updated CONTRIBUTING
3253 * Updated DRBG / RAND to request nonce and additional low entropy
3254 randomness from the system.
3256 *Matthias St. Pierre*
3258 * Updated 'openssl rehash' to use OpenSSL consistent default.
3262 * Moved the load of the ssl_conf module to libcrypto, which helps
3263 loading engines that libssl uses before libssl is initialised.
3267 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3271 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3273 *Ingo Schwarze, Rich Salz*
3275 * Added output of accepting IP address and port for 'openssl s_server'
3279 * Added a new API for TLSv1.3 ciphersuites:
3280 SSL_CTX_set_ciphersuites()
3281 SSL_set_ciphersuites()
3285 * Memory allocation failures consistently add an error to the error
3290 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3291 in libcrypto when run as setuid/setgid.
3295 * Load any config file by default when libssl is used.
3299 * Added new public header file <openssl/rand_drbg.h> and documentation
3300 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3302 *Matthias St. Pierre*
3304 * QNX support removed (cannot find contributors to get their approval
3305 for the license change).
3309 * TLSv1.3 replay protection for early data has been implemented. See the
3310 SSL_read_early_data() man page for further details.
3314 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3315 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3316 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3317 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3318 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3319 configuration has been separated out. See the ciphers man page or the
3320 SSL_CTX_set_ciphersuites() man page for more information.
3324 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3325 in responder mode now supports the new "-multi" option, which
3326 spawns the specified number of child processes to handle OCSP
3327 requests. The "-timeout" option now also limits the OCSP
3328 responder's patience to wait to receive the full client request
3329 on a newly accepted connection. Child processes are respawned
3330 as needed, and the CA index file is automatically reloaded
3331 when changed. This makes it possible to run the "ocsp" responder
3332 as a long-running service, making the OpenSSL CA somewhat more
3333 feature-complete. In this mode, most diagnostic messages logged
3334 after entering the event loop are logged via syslog(3) rather than
3339 * Added support for X448 and Ed448. Heavily based on original work by
3344 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3345 objects loaded. This adds the functions OSSL_STORE_expect() and
3346 OSSL_STORE_find() as well as needed tools to construct searches and
3347 get the search data out of them.
3351 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3352 version of OpenSSL should review their configuration settings to ensure
3353 that they are still appropriate for TLSv1.3. For further information see:
3354 <https://wiki.openssl.org/index.php/TLS1.3>
3358 * Grand redesign of the OpenSSL random generator
3360 The default RAND method now utilizes an AES-CTR DRBG according to
3361 NIST standard SP 800-90Ar1. The new random generator is essentially
3362 a port of the default random generator from the OpenSSL FIPS 2.0
3363 object module. It is a hybrid deterministic random bit generator
3364 using an AES-CTR bit stream and which seeds and reseeds itself
3365 automatically using trusted system entropy sources.
3367 Some of its new features are:
3368 - Support for multiple DRBG instances with seed chaining.
3369 - The default RAND method makes use of a DRBG.
3370 - There is a public and private DRBG instance.
3371 - The DRBG instances are fork-safe.
3372 - Keep all global DRBG instances on the secure heap if it is enabled.
3373 - The public and private DRBG instance are per thread for lock free
3376 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3378 * Changed Configure so it only says what it does and doesn't dump
3379 so much data. Instead, ./configdata.pm should be used as a script
3380 to display all sorts of configuration data.
3384 * Added processing of "make variables" to Configure.
3388 * Added SHA512/224 and SHA512/256 algorithm support.
3392 * The last traces of Netware support, first removed in 1.1.0, have
3397 * Get rid of Makefile.shared, and in the process, make the processing
3398 of certain files (rc.obj, or the .def/.map/.opt files produced from
3399 the ordinal files) more visible and hopefully easier to trace and
3400 debug (or make silent).
3404 * Make it possible to have environment variable assignments as
3405 arguments to config / Configure.
3409 * Add multi-prime RSA (RFC 8017) support.
3413 * Add SM3 implemented according to GB/T 32905-2016
3414 *Jack Lloyd <jack.lloyd@ribose.com>,*
3415 *Ronald Tse <ronald.tse@ribose.com>,*
3416 *Erick Borsboom <erick.borsboom@ribose.com>*
3418 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3419 as documented in RFC6066.
3420 Based on a patch from Tomasz Moń
3422 *Filipe Raimundo da Silva*
3424 * Add SM4 implemented according to GB/T 32907-2016.
3425 *Jack Lloyd <jack.lloyd@ribose.com>,*
3426 *Ronald Tse <ronald.tse@ribose.com>,*
3427 *Erick Borsboom <erick.borsboom@ribose.com>*
3429 * Reimplement -newreq-nodes and ERR_error_string_n; the
3430 original author does not agree with the license change.
3434 * Add ARIA AEAD TLS support.
3438 * Some macro definitions to support VS6 have been removed. Visual
3439 Studio 6 has not worked since 1.1.0
3443 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3444 without clearing the errors.
3448 * Add "atfork" functions. If building on a system that without
3449 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3450 requirements. The RAND facility now uses/requires this.
3458 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3459 not possible to disable entirely. However, it's still possible to
3460 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3463 To disable, configure with 'no-ui-console'. 'no-ui' is still
3464 possible to use as an alias. Check at compile time with the
3465 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3466 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3470 * Add a STORE module, which implements a uniform and URI based reader of
3471 stores that can contain keys, certificates, CRLs and numerous other
3472 objects. The main API is loosely based on a few stdio functions,
3473 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3474 OSSL_STORE_error and OSSL_STORE_close.
3475 The implementation uses backends called "loaders" to implement arbitrary
3476 URI schemes. There is one built in "loader" for the 'file' scheme.
3480 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3481 then adjusted to work on FreeBSD 8.4 as well.
3482 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3483 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3487 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3488 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3489 error code calls like this:
3491 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3493 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3494 that can be encoded in C. For the foreseeable future, this will only
3497 *Richard Levitte and Tim Hudson*
3499 * Removed BSD cryptodev engine.
3503 * Add a build target 'build_all_generated', to build all generated files
3504 and only that. This can be used to prepare everything that requires
3505 things like perl for a system that lacks perl and then move everything
3506 to that system and do the rest of the build there.
3510 * In the UI interface, make it possible to duplicate the user data. This
3511 can be used by engines that need to retain the data for a longer time
3512 than just the call where this user data is passed.
3516 * Ignore the '-named_curve auto' value for compatibility of applications
3519 *Tomáš Mráz <tmraz@fedoraproject.org>*
3521 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3522 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3523 alerts across multiple records (some of which could be empty). In practice
3524 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3525 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3526 support this at all. Supporting it adds significant complexity to the
3527 record layer, and its removal is unlikely to cause interoperability
3532 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3533 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3534 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3539 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3540 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3542 *Richard Levitte, Andy Polyakov*
3544 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3549 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3550 platform rather than 'mingw'.
3554 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3555 success if they are asked to add an object which already exists
3556 in the store. This change cascades to other functions which load
3557 certificates and CRLs.
3561 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3562 facilitate stack unwinding even from assembly subroutines.
3566 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3567 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3571 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3572 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3573 which is the minimum version we support.
3577 * Certificate time validation (X509_cmp_time) enforces stricter
3578 compliance with RFC 5280. Fractional seconds and timezone offsets
3579 are no longer allowed.
3583 * Add support for ARIA
3587 * s_client will now send the Server Name Indication (SNI) extension by
3588 default unless the new "-noservername" option is used. The server name is
3589 based on the host provided to the "-connect" option unless overridden by
3590 using "-servername".
3594 * Add support for SipHash
3598 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3599 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3600 prevent issues where no progress is being made and the peer continually
3601 sends unrecognised record types, using up resources processing them.
3605 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3606 using the algorithm defined in
3607 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3611 * Heartbeat support has been removed; the ABI is changed for now.
3613 *Richard Levitte, Rich Salz*
3615 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3619 * The RSA "null" method, which was partially supported to avoid patent
3620 issues, has been replaced to always returns NULL.
3627 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3629 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3630 used even when parsing explicit parameters, when loading a encoded key
3631 or calling `EC_GROUP_new_from_ecpkparameters()`/
3632 `EC_GROUP_new_from_ecparameters()`.
3633 This prevents bypass of security hardening and performance gains,
3634 especially for curves with specialized EC_METHODs.
3635 By default, if a key encoded with explicit parameters is loaded and later
3636 encoded, the output is still encoded with explicit parameters, even if
3637 internally a "named" EC_GROUP is used for computation.
3641 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3642 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3643 NULL. After this change, only the cofactor parameter can be NULL. It also
3644 does some minimal sanity checks on the passed order.
3649 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3650 An attack is simple, if the first CMS_recipientInfo is valid but the
3651 second CMS_recipientInfo is chosen ciphertext. If the second
3652 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3653 encryption key will be replaced by garbage, and the message cannot be
3654 decoded, but if the RSA decryption fails, the correct encryption key is
3655 used and the recipient will not notice the attack.
3656 As a work around for this potential attack the length of the decrypted
3657 key must be equal to the cipher default key length, in case the
3658 certificate is not given and all recipientInfo are tried out.
3659 The old behaviour can be re-enabled in the CMS code by setting the
3660 CMS_DEBUG_DECRYPT flag.
3665 * Use Windows installation paths in the mingw builds
3667 Mingw isn't a POSIX environment per se, which means that Windows
3668 paths should be used for installation.
3673 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3675 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3676 This changes the size when using the `genpkey` command when no size is given.
3677 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3678 generation commands to use 2048 bits by default.
3682 * Prevent over long nonces in ChaCha20-Poly1305.
3684 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3685 for every encryption operation. RFC 7539 specifies that the nonce value
3686 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3687 and front pads the nonce with 0 bytes if it is less than 12
3688 bytes. However it also incorrectly allows a nonce to be set of up to 16
3689 bytes. In this case only the last 12 bytes are significant and any
3690 additional leading bytes are ignored.
3692 It is a requirement of using this cipher that nonce values are
3693 unique. Messages encrypted using a reused nonce value are susceptible to
3694 serious confidentiality and integrity attacks. If an application changes
3695 the default nonce length to be longer than 12 bytes and then makes a
3696 change to the leading bytes of the nonce expecting the new value to be a
3697 new unique nonce then such an application could inadvertently encrypt
3698 messages with a reused nonce.
3700 Additionally the ignored bytes in a long nonce are not covered by the
3701 integrity guarantee of this cipher. Any application that relies on the
3702 integrity of these ignored leading bytes of a long nonce may be further
3703 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3704 is safe because no such use sets such a long nonce value. However user
3705 applications that use this cipher directly and set a non-default nonce
3706 length to be longer than 12 bytes may be vulnerable.
3708 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3714 * Added SCA hardening for modular field inversion in EC_GROUP through
3715 a new dedicated field_inv() pointer in EC_METHOD.
3716 This also addresses a leakage affecting conversions from projective
3717 to affine coordinates.
3719 *Billy Bob Brumley, Nicola Tuveri*
3721 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3722 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3726 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3730 * Remove the 'dist' target and add a tarball building script. The
3731 'dist' target has fallen out of use, and it shouldn't be
3732 necessary to configure just to create a source distribution.
3736 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3738 * Timing vulnerability in DSA signature generation
3740 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3741 timing side channel attack. An attacker could use variations in the signing
3742 algorithm to recover the private key.
3744 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3749 * Timing vulnerability in ECDSA signature generation
3751 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3752 timing side channel attack. An attacker could use variations in the signing
3753 algorithm to recover the private key.
3755 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3760 * Add coordinate blinding for EC_POINT and implement projective
3761 coordinate blinding for generic prime curves as a countermeasure to
3762 chosen point SCA attacks.
3764 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3766 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3768 * Client DoS due to large DH parameter
3770 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3771 malicious server can send a very large prime value to the client. This will
3772 cause the client to spend an unreasonably long period of time generating a
3773 key for this prime resulting in a hang until the client has finished. This
3774 could be exploited in a Denial Of Service attack.
3776 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3781 * Cache timing vulnerability in RSA Key Generation
3783 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3784 a cache timing side channel attack. An attacker with sufficient access to
3785 mount cache timing attacks during the RSA key generation process could
3786 recover the private key.
3788 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3789 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3794 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3795 parameter is no longer accepted, as it leads to a corrupt table. NULL
3796 pem_str is reserved for alias entries only.
3800 * Revert blinding in ECDSA sign and instead make problematic addition
3801 length-invariant. Switch even to fixed-length Montgomery multiplication.
3805 * Change generating and checking of primes so that the error rate of not
3806 being prime depends on the intended use based on the size of the input.
3807 For larger primes this will result in more rounds of Miller-Rabin.
3808 The maximal error rate for primes with more than 1080 bits is lowered
3811 *Kurt Roeckx, Annie Yousar*
3813 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3817 * Add blinding to ECDSA and DSA signatures to protect against side channel
3818 attacks discovered by Keegan Ryan (NCC Group).
3822 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3823 now allow empty (zero character) pass phrases.
3827 * Certificate time validation (X509_cmp_time) enforces stricter
3828 compliance with RFC 5280. Fractional seconds and timezone offsets
3829 are no longer allowed.
3833 * Fixed a text canonicalisation bug in CMS
3835 Where a CMS detached signature is used with text content the text goes
3836 through a canonicalisation process first prior to signing or verifying a
3837 signature. This process strips trailing space at the end of lines, converts
3838 line terminators to CRLF and removes additional trailing line terminators
3839 at the end of a file. A bug in the canonicalisation process meant that
3840 some characters, such as form-feed, were incorrectly treated as whitespace
3841 and removed. This is contrary to the specification (RFC5485). This fix
3842 could mean that detached text data signed with an earlier version of
3843 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3844 signed with a fixed OpenSSL may fail to verify with an earlier version of
3845 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3846 and use the "-binary" flag (for the "cms" command line application) or set
3847 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3851 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3853 * Constructed ASN.1 types with a recursive definition could exceed the stack
3855 Constructed ASN.1 types with a recursive definition (such as can be found
3856 in PKCS7) could eventually exceed the stack given malicious input with
3857 excessive recursion. This could result in a Denial Of Service attack. There
3858 are no such structures used within SSL/TLS that come from untrusted sources
3859 so this is considered safe.
3861 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3867 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3869 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3870 effectively reduced to only comparing the least significant bit of each
3871 byte. This allows an attacker to forge messages that would be considered as
3872 authenticated in an amount of tries lower than that guaranteed by the
3873 security claims of the scheme. The module can only be compiled by the
3874 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3876 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3882 * Add a build target 'build_all_generated', to build all generated files
3883 and only that. This can be used to prepare everything that requires
3884 things like perl for a system that lacks perl and then move everything
3885 to that system and do the rest of the build there.
3889 * Backport SSL_OP_NO_RENGOTIATION
3891 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3892 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3893 changes this is no longer possible in 1.1.0. Therefore, the new
3894 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3895 1.1.0 to provide equivalent functionality.
3897 Note that if an application built against 1.1.0h headers (or above) is run
3898 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3899 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3903 * Removed the OS390-Unix config target. It relied on a script that doesn't
3908 * rsaz_1024_mul_avx2 overflow bug on x86_64
3910 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3911 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3912 Analysis suggests that attacks against RSA and DSA as a result of this
3913 defect would be very difficult to perform and are not believed likely.
3914 Attacks against DH1024 are considered just feasible, because most of the
3915 work necessary to deduce information about a private key may be performed
3916 offline. The amount of resources required for such an attack would be
3917 significant. However, for an attack on TLS to be meaningful, the server
3918 would have to share the DH1024 private key among multiple clients, which is
3919 no longer an option since CVE-2016-0701.
3921 This only affects processors that support the AVX2 but not ADX extensions
3922 like Intel Haswell (4th generation).
3924 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3925 was originally found via the OSS-Fuzz project.
3930 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3932 * bn_sqrx8x_internal carry bug on x86_64
3934 There is a carry propagating bug in the x86_64 Montgomery squaring
3935 procedure. No EC algorithms are affected. Analysis suggests that attacks
3936 against RSA and DSA as a result of this defect would be very difficult to
3937 perform and are not believed likely. Attacks against DH are considered just
3938 feasible (although very difficult) because most of the work necessary to
3939 deduce information about a private key may be performed offline. The amount
3940 of resources required for such an attack would be very significant and
3941 likely only accessible to a limited number of attackers. An attacker would
3942 additionally need online access to an unpatched system using the target
3943 private key in a scenario with persistent DH parameters and a private
3944 key that is shared between multiple clients.
3946 This only affects processors that support the BMI1, BMI2 and ADX extensions
3947 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3949 This issue was reported to OpenSSL by the OSS-Fuzz project.
3954 * Malformed X.509 IPAddressFamily could cause OOB read
3956 If an X.509 certificate has a malformed IPAddressFamily extension,
3957 OpenSSL could do a one-byte buffer overread. The most likely result
3958 would be an erroneous display of the certificate in text format.
3960 This issue was reported to OpenSSL by the OSS-Fuzz project.
3965 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3967 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3968 platform rather than 'mingw'.
3972 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3973 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3974 which is the minimum version we support.
3978 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3980 * Encrypt-Then-Mac renegotiation crash
3982 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3983 negotiated where it was not in the original handshake (or vice-versa) then
3984 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3985 and servers are affected.
3987 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3992 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3994 * Truncated packet could crash via OOB read
3996 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3997 cipher is being used, then a truncated packet can cause that host to
3998 perform an out-of-bounds read, usually resulting in a crash.
4000 This issue was reported to OpenSSL by Robert Święcki of Google.
4005 * Bad (EC)DHE parameters cause a client crash
4007 If a malicious server supplies bad parameters for a DHE or ECDHE key
4008 exchange then this can result in the client attempting to dereference a
4009 NULL pointer leading to a client crash. This could be exploited in a Denial
4012 This issue was reported to OpenSSL by Guido Vranken.
4017 * BN_mod_exp may produce incorrect results on x86_64
4019 There is a carry propagating bug in the x86_64 Montgomery squaring
4020 procedure. No EC algorithms are affected. Analysis suggests that attacks
4021 against RSA and DSA as a result of this defect would be very difficult to
4022 perform and are not believed likely. Attacks against DH are considered just
4023 feasible (although very difficult) because most of the work necessary to
4024 deduce information about a private key may be performed offline. The amount
4025 of resources required for such an attack would be very significant and
4026 likely only accessible to a limited number of attackers. An attacker would
4027 additionally need online access to an unpatched system using the target
4028 private key in a scenario with persistent DH parameters and a private
4029 key that is shared between multiple clients. For example this can occur by
4030 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4031 similar to CVE-2015-3193 but must be treated as a separate problem.
4033 This issue was reported to OpenSSL by the OSS-Fuzz project.
4038 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
4040 * ChaCha20/Poly1305 heap-buffer-overflow
4042 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
4043 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
4044 crash. This issue is not considered to be exploitable beyond a DoS.
4046 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
4051 * CMS Null dereference
4053 Applications parsing invalid CMS structures can crash with a NULL pointer
4054 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
4055 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
4056 structure callback if an attempt is made to free certain invalid encodings.
4057 Only CHOICE structures using a callback which do not handle NULL value are
4060 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
4065 * Montgomery multiplication may produce incorrect results
4067 There is a carry propagating bug in the Broadwell-specific Montgomery
4068 multiplication procedure that handles input lengths divisible by, but
4069 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4070 and DH private keys are impossible. This is because the subroutine in
4071 question is not used in operations with the private key itself and an input
4072 of the attacker's direct choice. Otherwise the bug can manifest itself as
4073 transient authentication and key negotiation failures or reproducible
4074 erroneous outcome of public-key operations with specially crafted input.
4075 Among EC algorithms only Brainpool P-512 curves are affected and one
4076 presumably can attack ECDH key negotiation. Impact was not analyzed in
4077 detail, because pre-requisites for attack are considered unlikely. Namely
4078 multiple clients have to choose the curve in question and the server has to
4079 share the private key among them, neither of which is default behaviour.
4080 Even then only clients that chose the curve will be affected.
4082 This issue was publicly reported as transient failures and was not
4083 initially recognized as a security issue. Thanks to Richard Morgan for
4084 providing reproducible case.
4089 * Removed automatic addition of RPATH in shared libraries and executables,
4090 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
4094 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
4096 * Fix Use After Free for large message sizes
4098 The patch applied to address CVE-2016-6307 resulted in an issue where if a
4099 message larger than approx 16k is received then the underlying buffer to
4100 store the incoming message is reallocated and moved. Unfortunately a
4101 dangling pointer to the old location is left which results in an attempt to
4102 write to the previously freed location. This is likely to result in a
4103 crash, however it could potentially lead to execution of arbitrary code.
4105 This issue only affects OpenSSL 1.1.0a.
4107 This issue was reported to OpenSSL by Robert Święcki.
4112 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
4114 * OCSP Status Request extension unbounded memory growth
4116 A malicious client can send an excessively large OCSP Status Request
4117 extension. If that client continually requests renegotiation, sending a
4118 large OCSP Status Request extension each time, then there will be unbounded
4119 memory growth on the server. This will eventually lead to a Denial Of
4120 Service attack through memory exhaustion. Servers with a default
4121 configuration are vulnerable even if they do not support OCSP. Builds using
4122 the "no-ocsp" build time option are not affected.
4124 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4129 * SSL_peek() hang on empty record
4131 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
4132 sends an empty record. This could be exploited by a malicious peer in a
4133 Denial Of Service attack.
4135 This issue was reported to OpenSSL by Alex Gaynor.
4140 * Excessive allocation of memory in tls_get_message_header() and
4141 dtls1_preprocess_fragment()
4143 A (D)TLS message includes 3 bytes for its length in the header for the
4144 message. This would allow for messages up to 16Mb in length. Messages of
4145 this length are excessive and OpenSSL includes a check to ensure that a
4146 peer is sending reasonably sized messages in order to avoid too much memory
4147 being consumed to service a connection. A flaw in the logic of version
4148 1.1.0 means that memory for the message is allocated too early, prior to
4149 the excessive message length check. Due to way memory is allocated in
4150 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4151 to service a connection. This could lead to a Denial of Service through
4152 memory exhaustion. However, the excessive message length check still takes
4153 place, and this would cause the connection to immediately fail. Assuming
4154 that the application calls SSL_free() on the failed connection in a timely
4155 manner then the 21Mb of allocated memory will then be immediately freed
4156 again. Therefore, the excessive memory allocation will be transitory in
4157 nature. This then means that there is only a security impact if:
4159 1) The application does not call SSL_free() in a timely manner in the event
4160 that the connection fails
4162 2) The application is working in a constrained environment where there is
4163 very little free memory
4165 3) The attacker initiates multiple connection attempts such that there are
4166 multiple connections in a state where memory has been allocated for the
4167 connection; SSL_free() has not yet been called; and there is insufficient
4168 memory to service the multiple requests.
4170 Except in the instance of (1) above any Denial Of Service is likely to be
4171 transitory because as soon as the connection fails the memory is
4172 subsequently freed again in the SSL_free() call. However there is an
4173 increased risk during this period of application crashes due to the lack of
4174 memory - which would then mean a more serious Denial of Service.
4176 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4177 (CVE-2016-6307 and CVE-2016-6308)
4181 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4182 had to be removed. Primary reason is that vendor assembler can't
4183 assemble our modules with -KPIC flag. As result it, assembly
4184 support, was not even available as option. But its lack means
4185 lack of side-channel resistant code, which is incompatible with
4186 security by todays standards. Fortunately gcc is readily available
4187 prepackaged option, which we firmly point at...
4191 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4193 * Windows command-line tool supports UTF-8 opt-in option for arguments
4194 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4195 (to any value) allows Windows user to access PKCS#12 file generated
4196 with Windows CryptoAPI and protected with non-ASCII password, as well
4197 as files generated under UTF-8 locale on Linux also protected with
4202 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4203 have been disabled by default and removed from DEFAULT, just like RC4.
4204 See the RC4 item below to re-enable both.
4208 * The method for finding the storage location for the Windows RAND seed file
4209 has changed. First we check %RANDFILE%. If that is not set then we check
4210 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4211 all else fails we fall back to C:\.
4215 * The EVP_EncryptUpdate() function has had its return type changed from void
4216 to int. A return of 0 indicates and error while a return of 1 indicates
4221 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4222 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4223 off the constant time implementation for RSA, DSA and DH have been made
4224 no-ops and deprecated.
4228 * Windows RAND implementation was simplified to only get entropy by
4229 calling CryptGenRandom(). Various other RAND-related tickets
4232 *Joseph Wylie Yandle, Rich Salz*
4234 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4235 and `OPENSSL_LH_`, respectively. The old names are available
4236 with API compatibility. They new names are now completely documented.
4240 * Unify TYPE_up_ref(obj) methods signature.
4241 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4242 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4243 int (instead of void) like all others TYPE_up_ref() methods.
4244 So now these methods also check the return value of CRYPTO_atomic_add(),
4245 and the validity of object reference counter.
4247 *fdasilvayy@gmail.com*
4249 * With Windows Visual Studio builds, the .pdb files are installed
4250 alongside the installed libraries and executables. For a static
4251 library installation, ossl_static.pdb is the associate compiler
4252 generated .pdb file to be used when linking programs.
4256 * Remove openssl.spec. Packaging files belong with the packagers.
4260 * Automatic Darwin/OSX configuration has had a refresh, it will now
4261 recognise x86_64 architectures automatically. You can still decide
4262 to build for a different bitness with the environment variable
4263 KERNEL_BITS (can be 32 or 64), for example:
4265 KERNEL_BITS=32 ./config
4269 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4270 256 bit AES and HMAC with SHA256.
4274 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4278 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4282 * To enable users to have their own config files and build file templates,
4283 Configure looks in the directory indicated by the environment variable
4284 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4285 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4286 name and is used as is.
4290 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4291 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4292 X509_CERT_FILE_CTX was removed.
4296 * "shared" builds are now the default. To create only static libraries use
4297 the "no-shared" Configure option.
4301 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4302 All of these option have not worked for some while and are fundamental
4307 * Make various cleanup routines no-ops and mark them as deprecated. Most
4308 global cleanup functions are no longer required because they are handled
4309 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4310 Explicitly de-initing can cause problems (e.g. where a library that uses
4311 OpenSSL de-inits, but an application is still using it). The affected
4312 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4313 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4314 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4315 COMP_zlib_cleanup().
4319 * --strict-warnings no longer enables runtime debugging options
4320 such as REF_DEBUG. Instead, debug options are automatically
4321 enabled with '--debug' builds.
4323 *Andy Polyakov, Emilia Käsper*
4325 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4326 have been moved out of the public header files. New functions for managing
4327 these have been added.
4331 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4332 objects have been moved out of the public header files. New
4333 functions for managing these have been added.
4337 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4338 have been moved out of the public header files. New functions for managing
4339 these have been added.
4343 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4344 moved out of the public header files. New functions for managing these
4349 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4353 * Removed the mk1mf build scripts.
4357 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4358 it is always safe to #include a header now.
4362 * Removed the aged BC-32 config and all its supporting scripts
4366 * Removed support for Ultrix, Netware, and OS/2.
4370 * Add support for HKDF.
4372 *Alessandro Ghedini*
4374 * Add support for blake2b and blake2s
4378 * Added support for "pipelining". Ciphers that have the
4379 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4380 encryptions/decryptions simultaneously. There are currently no built-in
4381 ciphers with this property but the expectation is that engines will be able
4382 to offer it to significantly improve throughput. Support has been extended
4383 into libssl so that multiple records for a single connection can be
4384 processed in one go (for >=TLS 1.1).
4388 * Added the AFALG engine. This is an async capable engine which is able to
4389 offload work to the Linux kernel. In this initial version it only supports
4390 AES128-CBC. The kernel must be version 4.1.0 or greater.
4394 * OpenSSL now uses a new threading API. It is no longer necessary to
4395 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4396 are two supported threading models: pthreads and windows threads. It is
4397 also possible to configure OpenSSL at compile time for "no-threads". The
4398 old threading API should no longer be used. The functions have been
4399 replaced with "no-op" compatibility macros.
4401 *Alessandro Ghedini, Matt Caswell*
4403 * Modify behavior of ALPN to invoke callback after SNI/servername
4404 callback, such that updates to the SSL_CTX affect ALPN.
4408 * Add SSL_CIPHER queries for authentication and key-exchange.
4412 * Changes to the DEFAULT cipherlist:
4413 - Prefer (EC)DHE handshakes over plain RSA.
4414 - Prefer AEAD ciphers over legacy ciphers.
4415 - Prefer ECDSA over RSA when both certificates are available.
4416 - Prefer TLSv1.2 ciphers/PRF.
4417 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4422 * Change the ECC default curve list to be this, in order: x25519,
4423 secp256r1, secp521r1, secp384r1.
4427 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4428 disabled by default. They can be re-enabled using the
4429 enable-weak-ssl-ciphers option to Configure.
4433 * If the server has ALPN configured, but supports no protocols that the
4434 client advertises, send a fatal "no_application_protocol" alert.
4435 This behaviour is SHALL in RFC 7301, though it isn't universally
4436 implemented by other servers.
4440 * Add X25519 support.
4441 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4442 for public and private key encoding using the format documented in
4443 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4444 key generation and key derivation.
4446 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4451 * Deprecate SRP_VBASE_get_by_user.
4452 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4453 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4454 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4455 seed, even if the seed is configured.
4457 Users should use SRP_VBASE_get1_by_user instead. Note that in
4458 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4459 also that even though configuring the SRP seed attempts to hide
4460 invalid usernames by continuing the handshake with fake
4461 credentials, this behaviour is not constant time and no strong
4462 guarantees are made that the handshake is indistinguishable from
4463 that of a valid user.
4467 * Configuration change; it's now possible to build dynamic engines
4468 without having to build shared libraries and vice versa. This
4469 only applies to the engines in `engines/`, those in `crypto/engine/`
4470 will always be built into libcrypto (i.e. "static").
4472 Building dynamic engines is enabled by default; to disable, use
4473 the configuration option "disable-dynamic-engine".
4475 The only requirements for building dynamic engines are the
4476 presence of the DSO module and building with position independent
4477 code, so they will also automatically be disabled if configuring
4478 with "disable-dso" or "disable-pic".
4480 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4481 are also taken away from openssl/opensslconf.h, as they are
4486 * Configuration change; if there is a known flag to compile
4487 position independent code, it will always be applied on the
4488 libcrypto and libssl object files, and never on the application
4489 object files. This means other libraries that use routines from
4490 libcrypto / libssl can be made into shared libraries regardless
4491 of how OpenSSL was configured.
4493 If this isn't desirable, the configuration options "disable-pic"
4494 or "no-pic" can be used to disable the use of PIC. This will
4495 also disable building shared libraries and dynamic engines.
4499 * Removed JPAKE code. It was experimental and has no wide use.
4503 * The INSTALL_PREFIX Makefile variable has been renamed to
4504 DESTDIR. That makes for less confusion on what this variable
4505 is for. Also, the configuration option --install_prefix is
4510 * Heartbeat for TLS has been removed and is disabled by default
4511 for DTLS; configure with enable-heartbeats. Code that uses the
4512 old #define's might need to be updated.
4514 *Emilia Käsper, Rich Salz*
4516 * Rename REF_CHECK to REF_DEBUG.
4520 * New "unified" build system
4522 The "unified" build system is aimed to be a common system for all
4523 platforms we support. With it comes new support for VMS.
4525 This system builds supports building in a different directory tree
4526 than the source tree. It produces one Makefile (for unix family
4527 or lookalikes), or one descrip.mms (for VMS).
4529 The source of information to make the Makefile / descrip.mms is
4530 small files called 'build.info', holding the necessary
4531 information for each directory with source to compile, and a
4532 template in Configurations, like unix-Makefile.tmpl or
4535 With this change, the library names were also renamed on Windows
4536 and on VMS. They now have names that are closer to the standard
4537 on Unix, and include the major version number, and in certain
4538 cases, the architecture they are built for. See "Notes on shared
4539 libraries" in INSTALL.
4541 We rely heavily on the perl module Text::Template.
4545 * Added support for auto-initialisation and de-initialisation of the library.
4546 OpenSSL no longer requires explicit init or deinit routines to be called,
4547 except in certain circumstances. See the OPENSSL_init_crypto() and
4548 OPENSSL_init_ssl() man pages for further information.
4552 * The arguments to the DTLSv1_listen function have changed. Specifically the
4553 "peer" argument is now expected to be a BIO_ADDR object.
4555 * Rewrite of BIO networking library. The BIO library lacked consistent
4556 support of IPv6, and adding it required some more extensive
4557 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4558 which hold all types of addresses and chains of address information.
4559 It also introduces a new API, with functions like BIO_socket,
4560 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4561 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4562 have been adapted accordingly.
4566 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4571 * CRIME protection: disable compression by default, even if OpenSSL is
4572 compiled with zlib enabled. Applications can still enable compression
4573 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4574 using the SSL_CONF library to configure compression.
4578 * The signature of the session callback configured with
4579 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4580 was explicitly marked as `const unsigned char*` instead of
4585 * Always DPURIFY. Remove the use of uninitialized memory in the
4586 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4590 * Removed many obsolete configuration items, including
4591 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4592 MD2_CHAR, MD2_INT, MD2_LONG
4594 IDEA_SHORT, IDEA_LONG
4595 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4597 *Rich Salz, with advice from Andy Polyakov*
4599 * Many BN internals have been moved to an internal header file.
4601 *Rich Salz with help from Andy Polyakov*
4603 * Configuration and writing out the results from it has changed.
4604 Files such as Makefile include/openssl/opensslconf.h and are now
4605 produced through general templates, such as Makefile.in and
4606 crypto/opensslconf.h.in and some help from the perl module
4609 Also, the center of configuration information is no longer
4610 Makefile. Instead, Configure produces a perl module in
4611 configdata.pm which holds most of the config data (in the hash
4612 table %config), the target data that comes from the target
4613 configuration in one of the `Configurations/*.conf` files (in
4618 * To clarify their intended purposes, the Configure options
4619 --prefix and --openssldir change their semantics, and become more
4620 straightforward and less interdependent.
4622 --prefix shall be used exclusively to give the location INSTALLTOP
4623 where programs, scripts, libraries, include files and manuals are
4624 going to be installed. The default is now /usr/local.
4626 --openssldir shall be used exclusively to give the default
4627 location OPENSSLDIR where certificates, private keys, CRLs are
4628 managed. This is also where the default openssl.cnf gets
4630 If the directory given with this option is a relative path, the
4631 values of both the --prefix value and the --openssldir value will
4632 be combined to become OPENSSLDIR.
4633 The default for --openssldir is INSTALLTOP/ssl.
4635 Anyone who uses --openssldir to specify where OpenSSL is to be
4636 installed MUST change to use --prefix instead.
4640 * The GOST engine was out of date and therefore it has been removed. An up
4641 to date GOST engine is now being maintained in an external repository.
4642 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4643 support for GOST ciphersuites (these are only activated if a GOST engine
4648 * EGD is no longer supported by default; use enable-egd when
4651 *Ben Kaduk and Rich Salz*
4653 * The distribution now has Makefile.in files, which are used to
4654 create Makefile's when Configure is run. *Configure must be run
4655 before trying to build now.*
4659 * The return value for SSL_CIPHER_description() for error conditions
4664 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4666 Obtaining and performing DNSSEC validation of TLSA records is
4667 the application's responsibility. The application provides
4668 the TLSA records of its choice to OpenSSL, and these are then
4669 used to authenticate the peer.
4671 The TLSA records need not even come from DNS. They can, for
4672 example, be used to implement local end-entity certificate or
4673 trust-anchor "pinning", where the "pin" data takes the form
4674 of TLSA records, which can augment or replace verification
4675 based on the usual WebPKI public certification authorities.
4679 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4680 continues to support deprecated interfaces in default builds.
4681 However, applications are strongly advised to compile their
4682 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4683 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4684 or the 1.1.0 releases.
4686 In environments in which all applications have been ported to
4687 not use any deprecated interfaces OpenSSL's Configure script
4688 should be used with the --api=1.1.0 option to entirely remove
4689 support for the deprecated features from the library and
4690 unconditionally disable them in the installed headers.
4691 Essentially the same effect can be achieved with the "no-deprecated"
4692 argument to Configure, except that this will always restrict
4693 the build to just the latest API, rather than a fixed API
4696 As applications are ported to future revisions of the API,
4697 they should update their compile-time OPENSSL_API_COMPAT define
4698 accordingly, but in most cases should be able to continue to
4699 compile with later releases.
4701 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4702 0x10000000L and 0x00908000L, respectively. However those
4703 versions did not support the OPENSSL_API_COMPAT feature, and
4704 so applications are not typically tested for explicit support
4705 of just the undeprecated features of either release.
4709 * Add support for setting the minimum and maximum supported protocol.
4710 It can bet set via the SSL_set_min_proto_version() and
4711 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4712 MaxProtocol. It's recommended to use the new APIs to disable
4713 protocols instead of disabling individual protocols using
4714 SSL_set_options() or SSL_CONF's Protocol. This change also
4715 removes support for disabling TLS 1.2 in the OpenSSL TLS
4716 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4720 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4724 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4725 and integrates ECDSA and ECDH functionality into EC. Implementations can
4726 now redirect key generation and no longer need to convert to or from
4729 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4730 include the ec.h header file instead.
4734 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4735 ciphers who are no longer supported and drops support the ephemeral RSA key
4736 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4740 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4741 opaque. For HMAC_CTX, the following constructors and destructors
4744 HMAC_CTX *HMAC_CTX_new(void);
4745 void HMAC_CTX_free(HMAC_CTX *ctx);
4747 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4748 destroy such methods has been added. See EVP_MD_meth_new(3) and
4749 EVP_CIPHER_meth_new(3) for documentation.
4752 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4753 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4754 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4755 an already created structure.
4756 2) For consistency with the majority of our object creators and
4757 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4758 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4759 for deprecated builds.
4763 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4764 cryptographic operations to be performed asynchronously as long as an
4765 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4766 further details. Libssl has also had this capability integrated with the
4767 introduction of the new mode SSL_MODE_ASYNC and associated error
4768 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4769 pages. This work was developed in partnership with Intel Corp.
4773 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4774 always enabled now. If you want to disable the support you should
4775 exclude it using the list of supported ciphers. This also means that the
4776 "-no_ecdhe" option has been removed from s_server.
4780 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4781 SSL_{CTX_}set1_curves() which can set a list.
4785 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4786 curve you want to support using SSL_{CTX_}set1_curves().
4790 * State machine rewrite. The state machine code has been significantly
4791 refactored in order to remove much duplication of code and solve issues
4792 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4793 further details). This change does have some associated API changes.
4794 Notably the SSL_state() function has been removed and replaced by
4795 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4796 SSL_set_state() has been removed altogether. The previous handshake states
4797 defined in ssl.h and ssl3.h have also been removed.
4801 * All instances of the string "ssleay" in the public API were replaced
4802 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4803 Some error codes related to internal RSA_eay API's were renamed.
4807 * The demo files in crypto/threads were moved to demo/threads.
4811 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4814 *Matt Caswell, Rich Salz*
4816 * New ASN.1 embed macro.
4818 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4819 structure is not allocated: it is part of the parent. That is instead of
4827 This reduces memory fragmentation and make it impossible to accidentally
4828 set a mandatory field to NULL.
4830 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4831 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4832 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4837 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4841 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4842 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4843 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4844 DES and RC4 ciphersuites.
4848 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4849 This changes the decoding behaviour for some invalid messages,
4850 though the change is mostly in the more lenient direction, and
4851 legacy behaviour is preserved as much as possible.
4855 * Fix no-stdio build.
4856 *David Woodhouse <David.Woodhouse@intel.com> and also*
4857 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4859 * New testing framework
4860 The testing framework has been largely rewritten and is now using
4861 perl and the perl modules Test::Harness and an extended variant of
4862 Test::More called OpenSSL::Test to do its work. All test scripts in
4863 test/ have been rewritten into test recipes, and all direct calls to
4864 executables in test/Makefile have become individual recipes using the
4865 simplified testing OpenSSL::Test::Simple.
4867 For documentation on our testing modules, do:
4869 perldoc test/testlib/OpenSSL/Test/Simple.pm
4870 perldoc test/testlib/OpenSSL/Test.pm
4874 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4875 are used; the latter aborts on memory leaks (usually checked on exit).
4876 Some undocumented "set malloc, etc., hooks" functions were removed
4877 and others were changed. All are now documented.
4881 * In DSA_generate_parameters_ex, if the provided seed is too short,
4884 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4886 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4887 from RFC4279, RFC4785, RFC5487, RFC5489.
4889 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4890 original RSA_PSK patch.
4894 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4895 era flag was never set throughout the codebase (only read). Also removed
4896 SSL3_FLAGS_POP_BUFFER which was only used if
4897 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4901 * Changed the default name options in the "ca", "crl", "req" and "x509"
4902 to be "oneline" instead of "compat".
4906 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4907 not aware of clients that still exhibit this bug, and the workaround
4908 hasn't been working properly for a while.
4912 * The return type of BIO_number_read() and BIO_number_written() as well as
4913 the corresponding num_read and num_write members in the BIO structure has
4914 changed from unsigned long to uint64_t. On platforms where an unsigned
4915 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4920 * Given the pervasive nature of TLS extensions it is inadvisable to run
4921 OpenSSL without support for them. It also means that maintaining
4922 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4923 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4927 * Removed support for the two export grade static DH ciphersuites
4928 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4929 were newly added (along with a number of other static DH ciphersuites) to
4930 1.0.2. However the two export ones have *never* worked since they were
4931 introduced. It seems strange in any case to be adding new export
4932 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4936 * Version negotiation has been rewritten. In particular SSLv23_method(),
4937 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4938 and turned into macros which simply call the new preferred function names
4939 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4940 should use the new names instead. Also as part of this change the ssl23.h
4941 header file has been removed.
4945 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4946 code and the associated standard is no longer considered fit-for-purpose.
4950 * RT2547 was closed. When generating a private key, try to make the
4951 output file readable only by the owner. This behavior change might
4952 be noticeable when interacting with other software.
4954 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4959 * Added HTTP GET support to the ocsp command.
4963 * Changed default digest for the dgst and enc commands from MD5 to
4968 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4972 * Added support for TLS extended master secret from
4973 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4974 initial patch which was a great help during development.
4978 * All libssl internal structures have been removed from the public header
4979 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4980 now redundant). Users should not attempt to access internal structures
4981 directly. Instead they should use the provided API functions.
4985 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4986 Access to deprecated functions can be re-enabled by running config with
4987 "enable-deprecated". In addition applications wishing to use deprecated
4988 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4989 will, by default, disable some transitive includes that previously existed
4990 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4994 * Added support for OCB mode. OpenSSL has been granted a patent license
4995 compatible with the OpenSSL license for use of OCB. Details are available
4996 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4997 for OCB can be removed by calling config with no-ocb.
5001 * SSLv2 support has been removed. It still supports receiving an SSLv2
5002 compatible client hello.
5006 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
5007 done while fixing the error code for the key-too-small case.
5009 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
5011 * CA.sh has been removed; use CA.pl instead.
5015 * Removed old DES API.
5019 * Remove various unsupported platforms:
5025 Sinix/ReliantUNIX RM400
5030 16-bit platforms such as WIN16
5034 * Clean up OPENSSL_NO_xxx #define's
5035 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
5036 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
5037 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
5038 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
5039 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
5040 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
5041 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
5042 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
5043 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
5044 - Remove MS_STATIC; it's a relic from platforms <32 bits.
5048 * Cleaned up dead code
5049 Remove all but one '#ifdef undef' which is to be looked at.
5053 * Clean up calling of xxx_free routines.
5054 Just like free(), fix most of the xxx_free routines to accept
5055 NULL. Remove the non-null checks from callers. Save much code.
5059 * Add secure heap for storage of private keys (when possible).
5060 Add BIO_s_secmem(), CBIGNUM, etc.
5061 Contributed by Akamai Technologies under our Corporate CLA.
5065 * Experimental support for a new, fast, unbiased prime candidate generator,
5066 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
5068 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
5070 * New output format NSS in the sess_id command line tool. This allows
5071 exporting the session id and the master key in NSS keylog format.
5073 *Martin Kaiser <martin@kaiser.cx>*
5075 * Harmonize version and its documentation. -f flag is used to display
5078 *mancha <mancha1@zoho.com>*
5080 * Fix eckey_priv_encode so it immediately returns an error upon a failure
5081 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
5083 *mancha <mancha1@zoho.com>*
5085 * Fix some double frees. These are not thought to be exploitable.
5087 *mancha <mancha1@zoho.com>*
5089 * A missing bounds check in the handling of the TLS heartbeat extension
5090 can be used to reveal up to 64k of memory to a connected client or
5093 Thanks for Neel Mehta of Google Security for discovering this bug and to
5094 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
5095 preparing the fix ([CVE-2014-0160])
5097 *Adam Langley, Bodo Moeller*
5099 * Fix for the attack described in the paper "Recovering OpenSSL
5100 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
5101 by Yuval Yarom and Naomi Benger. Details can be obtained from:
5102 <http://eprint.iacr.org/2014/140>
5104 Thanks to Yuval Yarom and Naomi Benger for discovering this
5105 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5107 *Yuval Yarom and Naomi Benger*
5109 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5110 this fixes a limitation in previous versions of OpenSSL.
5114 * Experimental encrypt-then-mac support.
5116 Experimental support for encrypt then mac from
5117 draft-gutmann-tls-encrypt-then-mac-02.txt
5119 To enable it set the appropriate extension number (0x42 for the test
5120 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
5122 For non-compliant peers (i.e. just about everything) this should have no
5125 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
5129 * Add EVP support for key wrapping algorithms, to avoid problems with
5130 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5131 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5132 algorithms and include tests cases.
5136 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
5141 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5142 MGF1 digest and OAEP label.
5146 * Make openssl verify return errors.
5148 *Chris Palmer <palmer@google.com> and Ben Laurie*
5150 * New function ASN1_TIME_diff to calculate the difference between two
5151 ASN1_TIME structures or one structure and the current time.
5155 * Update fips_test_suite to support multiple command line options. New
5156 test to induce all self test errors in sequence and check expected
5161 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5162 sign or verify all in one operation.
5166 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5167 test programs and fips_test_suite. Includes functionality to parse
5168 the minimal script output of fipsalgest.pl directly.
5172 * Add authorisation parameter to FIPS_module_mode_set().
5176 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5180 * Use separate DRBG fields for internal and external flags. New function
5181 FIPS_drbg_health_check() to perform on demand health checking. Add
5182 generation tests to fips_test_suite with reduced health check interval to
5183 demonstrate periodic health checking. Add "nodh" option to
5184 fips_test_suite to skip very slow DH test.
5188 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5193 * More extensive health check for DRBG checking many more failure modes.
5194 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5195 combination: call this in fips_test_suite.
5199 * Add support for canonical generation of DSA parameter 'g'. See
5202 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5203 POST to handle HMAC cases.
5207 * Add functions FIPS_module_version() and FIPS_module_version_text()
5208 to return numerical and string versions of the FIPS module number.
5212 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5213 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5214 outside the validated module in the FIPS capable OpenSSL.
5218 * Minor change to DRBG entropy callback semantics. In some cases
5219 there is no multiple of the block length between min_len and
5220 max_len. Allow the callback to return more than max_len bytes
5221 of entropy but discard any extra: it is the callback's responsibility
5222 to ensure that the extra data discarded does not impact the
5223 requested amount of entropy.
5227 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5228 information in FIPS186-3, SP800-57 and SP800-131A.
5232 * CCM support via EVP. Interface is very similar to GCM case except we
5233 must supply all data in one chunk (i.e. no update, final) and the
5234 message length must be supplied if AAD is used. Add algorithm test
5239 * Initial version of POST overhaul. Add POST callback to allow the status
5240 of POST to be monitored and/or failures induced. Modify fips_test_suite
5241 to use callback. Always run all selftests even if one fails.
5245 * XTS support including algorithm test driver in the fips_gcmtest program.
5246 Note: this does increase the maximum key length from 32 to 64 bytes but
5247 there should be no binary compatibility issues as existing applications
5248 will never use XTS mode.
5252 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5253 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5254 performs algorithm blocking for unapproved PRNG types. Also do not
5255 set PRNG type in FIPS_mode_set(): leave this to the application.
5256 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5257 the standard OpenSSL PRNG: set additional data to a date time vector.
5261 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5262 This shouldn't present any incompatibility problems because applications
5263 shouldn't be using these directly and any that are will need to rethink
5264 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5268 * Extensive self tests and health checking required by SP800-90 DRBG.
5269 Remove strength parameter from FIPS_drbg_instantiate and always
5270 instantiate at maximum supported strength.
5274 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5278 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5282 * New function DH_compute_key_padded() to compute a DH key and pad with
5283 leading zeroes if needed: this complies with SP800-56A et al.
5287 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5288 anything, incomplete, subject to change and largely untested at present.
5292 * Modify fipscanisteronly build option to only build the necessary object
5293 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5297 * Add experimental option FIPSSYMS to give all symbols in
5298 fipscanister.o and FIPS or fips prefix. This will avoid
5299 conflicts with future versions of OpenSSL. Add perl script
5300 util/fipsas.pl to preprocess assembly language source files
5301 and rename any affected symbols.
5305 * Add selftest checks and algorithm block of non-fips algorithms in
5306 FIPS mode. Remove DES2 from selftests.
5310 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5311 return internal method without any ENGINE dependencies. Add new
5312 tiny fips sign and verify functions.
5316 * New build option no-ec2m to disable characteristic 2 code.
5320 * New build option "fipscanisteronly". This only builds fipscanister.o
5321 and (currently) associated fips utilities. Uses the file Makefile.fips
5322 instead of Makefile.org as the prototype.
5326 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5327 Update fips_gcmtest to use IV generator.
5331 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5332 setting output buffer to NULL. The `*Final` function must be
5333 called although it will not retrieve any additional data. The tag
5334 can be set or retrieved with a ctrl. The IV length is by default 12
5335 bytes (96 bits) but can be set to an alternative value. If the IV
5336 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5341 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5342 underlying do_cipher function handles all cipher semantics itself
5343 including padding and finalisation. This is useful if (for example)
5344 an ENGINE cipher handles block padding itself. The behaviour of
5345 do_cipher is subtly changed if this flag is set: the return value
5346 is the number of characters written to the output buffer (zero is
5347 no longer an error code) or a negative error code. Also if the
5348 input buffer is NULL and length 0 finalisation should be performed.
5352 * If a candidate issuer certificate is already part of the constructed
5353 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5357 * Improve forward-security support: add functions
5359 void SSL_CTX_set_not_resumable_session_callback(
5360 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5361 void SSL_set_not_resumable_session_callback(
5362 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5364 for use by SSL/TLS servers; the callback function will be called whenever a
5365 new session is created, and gets to decide whether the session may be
5366 cached to make it resumable (return 0) or not (return 1). (As by the
5367 SSL/TLS protocol specifications, the session_id sent by the server will be
5368 empty to indicate that the session is not resumable; also, the server will
5369 not generate RFC 4507 (RFC 5077) session tickets.)
5371 A simple reasonable callback implementation is to return is_forward_secure.
5372 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5373 by the SSL/TLS server library, indicating whether it can provide forward
5376 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5378 * New -verify_name option in command line utilities to set verification
5383 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5384 Add CMAC pkey methods.
5388 * Experimental renegotiation in s_server -www mode. If the client
5389 browses /reneg connection is renegotiated. If /renegcert it is
5390 renegotiated requesting a certificate.
5394 * Add an "external" session cache for debugging purposes to s_server. This
5395 should help trace issues which normally are only apparent in deployed
5396 multi-process servers.
5400 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5401 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5402 BIO_set_cipher() and some obscure PEM functions were changed so they
5403 can now return an error. The RAND changes required a change to the
5404 RAND_METHOD structure.
5408 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5409 a gcc attribute to warn if the result of a function is ignored. This
5410 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5411 whose return value is often ignored.
5415 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5416 These allow SCTs (signed certificate timestamps) to be requested and
5417 validated when establishing a connection.
5419 *Rob Percival <robpercival@google.com>*
5424 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5426 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5427 used even when parsing explicit parameters, when loading a encoded key
5428 or calling `EC_GROUP_new_from_ecpkparameters()`/
5429 `EC_GROUP_new_from_ecparameters()`.
5430 This prevents bypass of security hardening and performance gains,
5431 especially for curves with specialized EC_METHODs.
5432 By default, if a key encoded with explicit parameters is loaded and later
5433 encoded, the output is still encoded with explicit parameters, even if
5434 internally a "named" EC_GROUP is used for computation.
5438 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5439 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5440 NULL. After this change, only the cofactor parameter can be NULL. It also
5441 does some minimal sanity checks on the passed order.
5446 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5447 An attack is simple, if the first CMS_recipientInfo is valid but the
5448 second CMS_recipientInfo is chosen ciphertext. If the second
5449 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5450 encryption key will be replaced by garbage, and the message cannot be
5451 decoded, but if the RSA decryption fails, the correct encryption key is
5452 used and the recipient will not notice the attack.
5453 As a work around for this potential attack the length of the decrypted
5454 key must be equal to the cipher default key length, in case the
5455 certificate is not given and all recipientInfo are tried out.
5456 The old behaviour can be re-enabled in the CMS code by setting the
5457 CMS_DEBUG_DECRYPT flag.
5462 * Document issue with installation paths in diverse Windows builds
5464 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5465 binaries and run-time config file.
5470 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5472 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5473 This changes the size when using the `genpkey` command when no size is given.
5474 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5475 generation commands to use 2048 bits by default.
5479 * Add FIPS support for Android Arm 64-bit
5481 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5482 Module in Version 2.0.10. For some reason, the corresponding target
5483 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5484 built with FIPS support on Android Arm 64-bit. This omission has been
5487 *Matthias St. Pierre*
5489 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5491 * 0-byte record padding oracle
5493 If an application encounters a fatal protocol error and then calls
5494 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5495 then OpenSSL can respond differently to the calling application if a 0 byte
5496 record is received with invalid padding compared to if a 0 byte record is
5497 received with an invalid MAC. If the application then behaves differently
5498 based on that in a way that is detectable to the remote peer, then this
5499 amounts to a padding oracle that could be used to decrypt data.
5501 In order for this to be exploitable "non-stitched" ciphersuites must be in
5502 use. Stitched ciphersuites are optimised implementations of certain
5503 commonly used ciphersuites. Also the application must call SSL_shutdown()
5504 twice even if a protocol error has occurred (applications should not do
5505 this but some do anyway).
5507 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5508 Aviram, with additional investigation by Steven Collison and Andrew
5509 Hourselt. It was reported to OpenSSL on 10th December 2018.
5514 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5518 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5520 * Microarchitecture timing vulnerability in ECC scalar multiplication
5522 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5523 shown to be vulnerable to a microarchitecture timing side channel attack.
5524 An attacker with sufficient access to mount local timing attacks during
5525 ECDSA signature generation could recover the private key.
5527 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5528 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5534 * Timing vulnerability in DSA signature generation
5536 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5537 timing side channel attack. An attacker could use variations in the signing
5538 algorithm to recover the private key.
5540 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5545 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5546 Module, accidentally introduced while backporting security fixes from the
5547 development branch and hindering the use of ECC in FIPS mode.
5551 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5553 * Client DoS due to large DH parameter
5555 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5556 malicious server can send a very large prime value to the client. This will
5557 cause the client to spend an unreasonably long period of time generating a
5558 key for this prime resulting in a hang until the client has finished. This
5559 could be exploited in a Denial Of Service attack.
5561 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5566 * Cache timing vulnerability in RSA Key Generation
5568 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5569 a cache timing side channel attack. An attacker with sufficient access to
5570 mount cache timing attacks during the RSA key generation process could
5571 recover the private key.
5573 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5574 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5579 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5580 parameter is no longer accepted, as it leads to a corrupt table. NULL
5581 pem_str is reserved for alias entries only.
5585 * Revert blinding in ECDSA sign and instead make problematic addition
5586 length-invariant. Switch even to fixed-length Montgomery multiplication.
5590 * Change generating and checking of primes so that the error rate of not
5591 being prime depends on the intended use based on the size of the input.
5592 For larger primes this will result in more rounds of Miller-Rabin.
5593 The maximal error rate for primes with more than 1080 bits is lowered
5596 *Kurt Roeckx, Annie Yousar*
5598 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5602 * Add blinding to ECDSA and DSA signatures to protect against side channel
5603 attacks discovered by Keegan Ryan (NCC Group).
5607 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5608 now allow empty (zero character) pass phrases.
5612 * Certificate time validation (X509_cmp_time) enforces stricter
5613 compliance with RFC 5280. Fractional seconds and timezone offsets
5614 are no longer allowed.
5618 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5620 * Constructed ASN.1 types with a recursive definition could exceed the stack
5622 Constructed ASN.1 types with a recursive definition (such as can be found
5623 in PKCS7) could eventually exceed the stack given malicious input with
5624 excessive recursion. This could result in a Denial Of Service attack. There
5625 are no such structures used within SSL/TLS that come from untrusted sources
5626 so this is considered safe.
5628 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5634 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5636 * Read/write after SSL object in error state
5638 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5639 mechanism. The intent was that if a fatal error occurred during a handshake
5640 then OpenSSL would move into the error state and would immediately fail if
5641 you attempted to continue the handshake. This works as designed for the
5642 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5643 SSL_connect()), however due to a bug it does not work correctly if
5644 SSL_read() or SSL_write() is called directly. In that scenario, if the
5645 handshake fails then a fatal error will be returned in the initial function
5646 call. If SSL_read()/SSL_write() is subsequently called by the application
5647 for the same SSL object then it will succeed and the data is passed without
5648 being decrypted/encrypted directly from the SSL/TLS record layer.
5650 In order to exploit this issue an application bug would have to be present
5651 that resulted in a call to SSL_read()/SSL_write() being issued after having
5652 already received a fatal error.
5654 This issue was reported to OpenSSL by David Benjamin (Google).
5659 * rsaz_1024_mul_avx2 overflow bug on x86_64
5661 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5662 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5663 Analysis suggests that attacks against RSA and DSA as a result of this
5664 defect would be very difficult to perform and are not believed likely.
5665 Attacks against DH1024 are considered just feasible, because most of the
5666 work necessary to deduce information about a private key may be performed
5667 offline. The amount of resources required for such an attack would be
5668 significant. However, for an attack on TLS to be meaningful, the server
5669 would have to share the DH1024 private key among multiple clients, which is
5670 no longer an option since CVE-2016-0701.
5672 This only affects processors that support the AVX2 but not ADX extensions
5673 like Intel Haswell (4th generation).
5675 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5676 was originally found via the OSS-Fuzz project.
5681 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5683 * bn_sqrx8x_internal carry bug on x86_64
5685 There is a carry propagating bug in the x86_64 Montgomery squaring
5686 procedure. No EC algorithms are affected. Analysis suggests that attacks
5687 against RSA and DSA as a result of this defect would be very difficult to
5688 perform and are not believed likely. Attacks against DH are considered just
5689 feasible (although very difficult) because most of the work necessary to
5690 deduce information about a private key may be performed offline. The amount
5691 of resources required for such an attack would be very significant and
5692 likely only accessible to a limited number of attackers. An attacker would
5693 additionally need online access to an unpatched system using the target
5694 private key in a scenario with persistent DH parameters and a private
5695 key that is shared between multiple clients.
5697 This only affects processors that support the BMI1, BMI2 and ADX extensions
5698 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5700 This issue was reported to OpenSSL by the OSS-Fuzz project.
5705 * Malformed X.509 IPAddressFamily could cause OOB read
5707 If an X.509 certificate has a malformed IPAddressFamily extension,
5708 OpenSSL could do a one-byte buffer overread. The most likely result
5709 would be an erroneous display of the certificate in text format.
5711 This issue was reported to OpenSSL by the OSS-Fuzz project.
5715 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5717 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5718 platform rather than 'mingw'.
5722 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5724 * Truncated packet could crash via OOB read
5726 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5727 cipher is being used, then a truncated packet can cause that host to
5728 perform an out-of-bounds read, usually resulting in a crash.
5730 This issue was reported to OpenSSL by Robert Święcki of Google.
5735 * BN_mod_exp may produce incorrect results on x86_64
5737 There is a carry propagating bug in the x86_64 Montgomery squaring
5738 procedure. No EC algorithms are affected. Analysis suggests that attacks
5739 against RSA and DSA as a result of this defect would be very difficult to
5740 perform and are not believed likely. Attacks against DH are considered just
5741 feasible (although very difficult) because most of the work necessary to
5742 deduce information about a private key may be performed offline. The amount
5743 of resources required for such an attack would be very significant and
5744 likely only accessible to a limited number of attackers. An attacker would
5745 additionally need online access to an unpatched system using the target
5746 private key in a scenario with persistent DH parameters and a private
5747 key that is shared between multiple clients. For example this can occur by
5748 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5749 similar to CVE-2015-3193 but must be treated as a separate problem.
5751 This issue was reported to OpenSSL by the OSS-Fuzz project.
5756 * Montgomery multiplication may produce incorrect results
5758 There is a carry propagating bug in the Broadwell-specific Montgomery
5759 multiplication procedure that handles input lengths divisible by, but
5760 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5761 and DH private keys are impossible. This is because the subroutine in
5762 question is not used in operations with the private key itself and an input
5763 of the attacker's direct choice. Otherwise the bug can manifest itself as
5764 transient authentication and key negotiation failures or reproducible
5765 erroneous outcome of public-key operations with specially crafted input.
5766 Among EC algorithms only Brainpool P-512 curves are affected and one
5767 presumably can attack ECDH key negotiation. Impact was not analyzed in
5768 detail, because pre-requisites for attack are considered unlikely. Namely
5769 multiple clients have to choose the curve in question and the server has to
5770 share the private key among them, neither of which is default behaviour.
5771 Even then only clients that chose the curve will be affected.
5773 This issue was publicly reported as transient failures and was not
5774 initially recognized as a security issue. Thanks to Richard Morgan for
5775 providing reproducible case.
5780 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5781 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5782 prevent issues where no progress is being made and the peer continually
5783 sends unrecognised record types, using up resources processing them.
5787 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5789 * Missing CRL sanity check
5791 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5792 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5793 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5795 This issue only affects the OpenSSL 1.0.2i
5800 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5802 * OCSP Status Request extension unbounded memory growth
5804 A malicious client can send an excessively large OCSP Status Request
5805 extension. If that client continually requests renegotiation, sending a
5806 large OCSP Status Request extension each time, then there will be unbounded
5807 memory growth on the server. This will eventually lead to a Denial Of
5808 Service attack through memory exhaustion. Servers with a default
5809 configuration are vulnerable even if they do not support OCSP. Builds using
5810 the "no-ocsp" build time option are not affected.
5812 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5817 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5820 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5826 * OOB write in MDC2_Update()
5828 An overflow can occur in MDC2_Update() either if called directly or
5829 through the EVP_DigestUpdate() function using MDC2. If an attacker
5830 is able to supply very large amounts of input data after a previous
5831 call to EVP_EncryptUpdate() with a partial block then a length check
5832 can overflow resulting in a heap corruption.
5834 The amount of data needed is comparable to SIZE_MAX which is impractical
5837 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5842 * Malformed SHA512 ticket DoS
5844 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5845 DoS attack where a malformed ticket will result in an OOB read which will
5848 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5849 a custom server callback and ticket lookup mechanism.
5851 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5856 * OOB write in BN_bn2dec()
5858 The function BN_bn2dec() does not check the return value of BN_div_word().
5859 This can cause an OOB write if an application uses this function with an
5860 overly large BIGNUM. This could be a problem if an overly large certificate
5861 or CRL is printed out from an untrusted source. TLS is not affected because
5862 record limits will reject an oversized certificate before it is parsed.
5864 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5869 * OOB read in TS_OBJ_print_bio()
5871 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5872 the total length the OID text representation would use and not the amount
5873 of data written. This will result in OOB reads when large OIDs are
5876 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5881 * Pointer arithmetic undefined behaviour
5883 Avoid some undefined pointer arithmetic
5885 A common idiom in the codebase is to check limits in the following manner:
5888 Where "p" points to some malloc'd data of SIZE bytes and
5891 "len" here could be from some externally supplied data (e.g. from a TLS
5894 The rules of C pointer arithmetic are such that "p + len" is only well
5895 defined where len <= SIZE. Therefore the above idiom is actually
5896 undefined behaviour.
5898 For example this could cause problems if some malloc implementation
5899 provides an address for "p" such that "p + len" actually overflows for
5900 values of len that are too big and therefore p + len < limit.
5902 This issue was reported to OpenSSL by Guido Vranken
5907 * Constant time flag not preserved in DSA signing
5909 Operations in the DSA signing algorithm should run in constant time in
5910 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5911 implementation means that a non-constant time codepath is followed for
5912 certain operations. This has been demonstrated through a cache-timing
5913 attack to be sufficient for an attacker to recover the private DSA key.
5915 This issue was reported by César Pereida (Aalto University), Billy Brumley
5916 (Tampere University of Technology), and Yuval Yarom (The University of
5917 Adelaide and NICTA).
5922 * DTLS buffered message DoS
5924 In a DTLS connection where handshake messages are delivered out-of-order
5925 those messages that OpenSSL is not yet ready to process will be buffered
5926 for later use. Under certain circumstances, a flaw in the logic means that
5927 those messages do not get removed from the buffer even though the handshake
5928 has been completed. An attacker could force up to approx. 15 messages to
5929 remain in the buffer when they are no longer required. These messages will
5930 be cleared when the DTLS connection is closed. The default maximum size for
5931 a message is 100k. Therefore, the attacker could force an additional 1500k
5932 to be consumed per connection. By opening many simultaneous connections an
5933 attacker could cause a DoS attack through memory exhaustion.
5935 This issue was reported to OpenSSL by Quan Luo.
5940 * DTLS replay protection DoS
5942 A flaw in the DTLS replay attack protection mechanism means that records
5943 that arrive for future epochs update the replay protection "window" before
5944 the MAC for the record has been validated. This could be exploited by an
5945 attacker by sending a record for the next epoch (which does not have to
5946 decrypt or have a valid MAC), with a very large sequence number. This means
5947 that all subsequent legitimate packets are dropped causing a denial of
5948 service for a specific DTLS connection.
5950 This issue was reported to OpenSSL by the OCAP audit team.
5955 * Certificate message OOB reads
5957 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5958 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5959 theoretical DoS risk but this has not been observed in practice on common
5962 The messages affected are client certificate, client certificate request
5963 and server certificate. As a result the attack can only be performed
5964 against a client or a server which enables client authentication.
5966 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5971 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5973 * Prevent padding oracle in AES-NI CBC MAC check
5975 A MITM attacker can use a padding oracle attack to decrypt traffic
5976 when the connection uses an AES CBC cipher and the server support
5979 This issue was introduced as part of the fix for Lucky 13 padding
5980 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5981 constant time by making sure that always the same bytes are read and
5982 compared against either the MAC or padding bytes. But it no longer
5983 checked that there was enough data to have both the MAC and padding
5986 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5990 * Fix EVP_EncodeUpdate overflow
5992 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5993 Base64 encoding of binary data. If an attacker is able to supply very large
5994 amounts of input data then a length check can overflow resulting in a heap
5997 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5998 the `PEM_write_bio*` family of functions. These are mainly used within the
5999 OpenSSL command line applications, so any application which processes data
6000 from an untrusted source and outputs it as a PEM file should be considered
6001 vulnerable to this issue. User applications that call these APIs directly
6002 with large amounts of untrusted data may also be vulnerable.
6004 This issue was reported by Guido Vranken.
6009 * Fix EVP_EncryptUpdate overflow
6011 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6012 is able to supply very large amounts of input data after a previous call to
6013 EVP_EncryptUpdate() with a partial block then a length check can overflow
6014 resulting in a heap corruption. Following an analysis of all OpenSSL
6015 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6016 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6017 the first called function after an EVP_EncryptInit(), and therefore that
6018 specific call must be safe. The second form is where the length passed to
6019 EVP_EncryptUpdate() can be seen from the code to be some small value and
6020 therefore there is no possibility of an overflow. Since all instances are
6021 one of these two forms, it is believed that there can be no overflows in
6022 internal code due to this problem. It should be noted that
6023 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6024 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6025 of these calls have also been analysed too and it is believed there are no
6026 instances in internal usage where an overflow could occur.
6028 This issue was reported by Guido Vranken.
6033 * Prevent ASN.1 BIO excessive memory allocation
6035 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6036 a short invalid encoding can cause allocation of large amounts of memory
6037 potentially consuming excessive resources or exhausting memory.
6039 Any application parsing untrusted data through d2i BIO functions is
6040 affected. The memory based functions such as d2i_X509() are *not* affected.
6041 Since the memory based functions are used by the TLS library, TLS
6042 applications are not affected.
6044 This issue was reported by Brian Carpenter.
6051 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6052 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6053 in arbitrary stack data being returned in the buffer.
6055 This issue was reported by Guido Vranken.
6060 * Modify behavior of ALPN to invoke callback after SNI/servername
6061 callback, such that updates to the SSL_CTX affect ALPN.
6065 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6070 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6071 methods are enabled and ssl2 is disabled the methods return NULL.
6075 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
6077 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6078 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6079 provide any "EXPORT" or "LOW" strength ciphers.
6083 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6084 is by default disabled at build-time. Builds that are not configured with
6085 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6086 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6087 will need to explicitly call either of:
6089 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6091 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6093 as appropriate. Even if either of those is used, or the application
6094 explicitly uses the version-specific SSLv2_method() or its client and
6095 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6096 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6097 ciphers, and SSLv2 56-bit DES are no longer available.
6102 * Fix a double-free in DSA code
6104 A double free bug was discovered when OpenSSL parses malformed DSA private
6105 keys and could lead to a DoS attack or memory corruption for applications
6106 that receive DSA private keys from untrusted sources. This scenario is
6109 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6115 * Disable SRP fake user seed to address a server memory leak.
6117 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6119 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6120 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6121 was changed to ignore the "fake user" SRP seed, even if the seed
6124 Users should use SRP_VBASE_get1_by_user instead. Note that in
6125 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6126 also that even though configuring the SRP seed attempts to hide
6127 invalid usernames by continuing the handshake with fake
6128 credentials, this behaviour is not constant time and no strong
6129 guarantees are made that the handshake is indistinguishable from
6130 that of a valid user.
6135 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6137 In the BN_hex2bn function the number of hex digits is calculated using an
6138 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6139 large values of `i` this can result in `bn_expand` not allocating any
6140 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6141 field as NULL leading to a subsequent NULL ptr deref. For very large values
6142 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6143 In this case memory is allocated to the internal BIGNUM data field, but it
6144 is insufficiently sized leading to heap corruption. A similar issue exists
6145 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6146 is ever called by user applications with very large untrusted hex/dec data.
6147 This is anticipated to be a rare occurrence.
6149 All OpenSSL internal usage of these functions use data that is not expected
6150 to be untrusted, e.g. config file data or application command line
6151 arguments. If user developed applications generate config file data based
6152 on untrusted data then it is possible that this could also lead to security
6153 consequences. This is also anticipated to be rare.
6155 This issue was reported to OpenSSL by Guido Vranken.
6160 * Fix memory issues in `BIO_*printf` functions
6162 The internal `fmtstr` function used in processing a "%s" format string in
6163 the `BIO_*printf` functions could overflow while calculating the length of a
6164 string and cause an OOB read when printing very long strings.
6166 Additionally the internal `doapr_outch` function can attempt to write to an
6167 OOB memory location (at an offset from the NULL pointer) in the event of a
6168 memory allocation failure. In 1.0.2 and below this could be caused where
6169 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6170 could be in processing a very long "%s" format string. Memory leaks can
6173 The first issue may mask the second issue dependent on compiler behaviour.
6174 These problems could enable attacks where large amounts of untrusted data
6175 is passed to the `BIO_*printf` functions. If applications use these functions
6176 in this way then they could be vulnerable. OpenSSL itself uses these
6177 functions when printing out human-readable dumps of ASN.1 data. Therefore
6178 applications that print this data could be vulnerable if the data is from
6179 untrusted sources. OpenSSL command line applications could also be
6180 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6181 as command line arguments.
6183 Libssl is not considered directly vulnerable. Additionally certificates etc
6184 received via remote connections via libssl are also unlikely to be able to
6185 trigger these issues because of message size limits enforced within libssl.
6187 This issue was reported to OpenSSL Guido Vranken.
6192 * Side channel attack on modular exponentiation
6194 A side-channel attack was found which makes use of cache-bank conflicts on
6195 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6196 of RSA keys. The ability to exploit this issue is limited as it relies on
6197 an attacker who has control of code in a thread running on the same
6198 hyper-threaded core as the victim thread which is performing decryptions.
6200 This issue was reported to OpenSSL by Yuval Yarom, The University of
6201 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6202 Nadia Heninger, University of Pennsylvania with more information at
6203 <http://cachebleed.info>.
6208 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6209 if no keysize is specified with default_bits. This fixes an
6210 omission in an earlier change that changed all RSA/DSA key generation
6211 commands to use 2048 bits by default.
6215 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6217 * DH small subgroups
6219 Historically OpenSSL only ever generated DH parameters based on "safe"
6220 primes. More recently (in version 1.0.2) support was provided for
6221 generating X9.42 style parameter files such as those required for RFC 5114
6222 support. The primes used in such files may not be "safe". Where an
6223 application is using DH configured with parameters based on primes that are
6224 not "safe" then an attacker could use this fact to find a peer's private
6225 DH exponent. This attack requires that the attacker complete multiple
6226 handshakes in which the peer uses the same private DH exponent. For example
6227 this could be used to discover a TLS server's private DH exponent if it's
6228 reusing the private DH exponent or it's using a static DH ciphersuite.
6230 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6231 TLS. It is not on by default. If the option is not set then the server
6232 reuses the same private DH exponent for the life of the server process and
6233 would be vulnerable to this attack. It is believed that many popular
6234 applications do set this option and would therefore not be at risk.
6236 The fix for this issue adds an additional check where a "q" parameter is
6237 available (as is the case in X9.42 based parameters). This detects the
6238 only known attack, and is the only possible defense for static DH
6239 ciphersuites. This could have some performance impact.
6241 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6242 default and cannot be disabled. This could have some performance impact.
6244 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6249 * SSLv2 doesn't block disabled ciphers
6251 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6252 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6253 been disabled, provided that the SSLv2 protocol was not also disabled via
6256 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6257 and Sebastian Schinzel.
6262 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6264 * BN_mod_exp may produce incorrect results on x86_64
6266 There is a carry propagating bug in the x86_64 Montgomery squaring
6267 procedure. No EC algorithms are affected. Analysis suggests that attacks
6268 against RSA and DSA as a result of this defect would be very difficult to
6269 perform and are not believed likely. Attacks against DH are considered just
6270 feasible (although very difficult) because most of the work necessary to
6271 deduce information about a private key may be performed offline. The amount
6272 of resources required for such an attack would be very significant and
6273 likely only accessible to a limited number of attackers. An attacker would
6274 additionally need online access to an unpatched system using the target
6275 private key in a scenario with persistent DH parameters and a private
6276 key that is shared between multiple clients. For example this can occur by
6277 default in OpenSSL DHE based SSL/TLS ciphersuites.
6279 This issue was reported to OpenSSL by Hanno Böck.
6284 * Certificate verify crash with missing PSS parameter
6286 The signature verification routines will crash with a NULL pointer
6287 dereference if presented with an ASN.1 signature using the RSA PSS
6288 algorithm and absent mask generation function parameter. Since these
6289 routines are used to verify certificate signature algorithms this can be
6290 used to crash any certificate verification operation and exploited in a
6291 DoS attack. Any application which performs certificate verification is
6292 vulnerable including OpenSSL clients and servers which enable client
6295 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6300 * X509_ATTRIBUTE memory leak
6302 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6303 memory. This structure is used by the PKCS#7 and CMS routines so any
6304 application which reads PKCS#7 or CMS data from untrusted sources is
6305 affected. SSL/TLS is not affected.
6307 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6313 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6314 This changes the decoding behaviour for some invalid messages,
6315 though the change is mostly in the more lenient direction, and
6316 legacy behaviour is preserved as much as possible.
6320 * In DSA_generate_parameters_ex, if the provided seed is too short,
6323 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6325 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6327 * Alternate chains certificate forgery
6329 During certificate verification, OpenSSL will attempt to find an
6330 alternative certificate chain if the first attempt to build such a chain
6331 fails. An error in the implementation of this logic can mean that an
6332 attacker could cause certain checks on untrusted certificates to be
6333 bypassed, such as the CA flag, enabling them to use a valid leaf
6334 certificate to act as a CA and "issue" an invalid certificate.
6336 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6341 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6343 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6344 incompatibility in the handling of HMAC. The previous ABI has now been
6349 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6351 * Malformed ECParameters causes infinite loop
6353 When processing an ECParameters structure OpenSSL enters an infinite loop
6354 if the curve specified is over a specially malformed binary polynomial
6357 This can be used to perform denial of service against any
6358 system which processes public keys, certificate requests or
6359 certificates. This includes TLS clients and TLS servers with
6360 client authentication enabled.
6362 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6367 * Exploitable out-of-bounds read in X509_cmp_time
6369 X509_cmp_time does not properly check the length of the ASN1_TIME
6370 string and can read a few bytes out of bounds. In addition,
6371 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6374 An attacker can use this to craft malformed certificates and CRLs of
6375 various sizes and potentially cause a segmentation fault, resulting in
6376 a DoS on applications that verify certificates or CRLs. TLS clients
6377 that verify CRLs are affected. TLS clients and servers with client
6378 authentication enabled may be affected if they use custom verification
6381 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6382 independently by Hanno Böck.
6387 * PKCS7 crash with missing EnvelopedContent
6389 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6390 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6391 with missing content and trigger a NULL pointer dereference on parsing.
6393 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6394 structures from untrusted sources are affected. OpenSSL clients and
6395 servers are not affected.
6397 This issue was reported to OpenSSL by Michal Zalewski (Google).
6402 * CMS verify infinite loop with unknown hash function
6404 When verifying a signedData message the CMS code can enter an infinite loop
6405 if presented with an unknown hash function OID. This can be used to perform
6406 denial of service against any system which verifies signedData messages using
6408 This issue was reported to OpenSSL by Johannes Bauer.
6413 * Race condition handling NewSessionTicket
6415 If a NewSessionTicket is received by a multi-threaded client when attempting to
6416 reuse a previous ticket then a race condition can occur potentially leading to
6417 a double free of the ticket data.
6422 * Only support 256-bit or stronger elliptic curves with the
6423 'ecdh_auto' setting (server) or by default (client). Of supported
6424 curves, prefer P-256 (both).
6428 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6430 * ClientHello sigalgs DoS fix
6432 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6433 invalid signature algorithms extension a NULL pointer dereference will
6434 occur. This can be exploited in a DoS attack against the server.
6436 This issue was was reported to OpenSSL by David Ramos of Stanford
6440 *Stephen Henson and Matt Caswell*
6442 * Multiblock corrupted pointer fix
6444 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6445 feature only applies on 64 bit x86 architecture platforms that support AES
6446 NI instructions. A defect in the implementation of "multiblock" can cause
6447 OpenSSL's internal write buffer to become incorrectly set to NULL when
6448 using non-blocking IO. Typically, when the user application is using a
6449 socket BIO for writing, this will only result in a failed connection.
6450 However if some other BIO is used then it is likely that a segmentation
6451 fault will be triggered, thus enabling a potential DoS attack.
6453 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6458 * Segmentation fault in DTLSv1_listen fix
6460 The DTLSv1_listen function is intended to be stateless and processes the
6461 initial ClientHello from many peers. It is common for user code to loop
6462 over the call to DTLSv1_listen until a valid ClientHello is received with
6463 an associated cookie. A defect in the implementation of DTLSv1_listen means
6464 that state is preserved in the SSL object from one invocation to the next
6465 that can lead to a segmentation fault. Errors processing the initial
6466 ClientHello can trigger this scenario. An example of such an error could be
6467 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6470 This issue was reported to OpenSSL by Per Allansson.
6475 * Segmentation fault in ASN1_TYPE_cmp fix
6477 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6478 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6479 certificate signature algorithm consistency this can be used to crash any
6480 certificate verification operation and exploited in a DoS attack. Any
6481 application which performs certificate verification is vulnerable including
6482 OpenSSL clients and servers which enable client authentication.
6487 * Segmentation fault for invalid PSS parameters fix
6489 The signature verification routines will crash with a NULL pointer
6490 dereference if presented with an ASN.1 signature using the RSA PSS
6491 algorithm and invalid parameters. Since these routines are used to verify
6492 certificate signature algorithms this can be used to crash any
6493 certificate verification operation and exploited in a DoS attack. Any
6494 application which performs certificate verification is vulnerable including
6495 OpenSSL clients and servers which enable client authentication.
6497 This issue was was reported to OpenSSL by Brian Carpenter.
6502 * ASN.1 structure reuse memory corruption fix
6504 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6505 memory corruption via an invalid write. Such reuse is and has been
6506 strongly discouraged and is believed to be rare.
6508 Applications that parse structures containing CHOICE or ANY DEFINED BY
6509 components may be affected. Certificate parsing (d2i_X509 and related
6510 functions) are however not affected. OpenSSL clients and servers are
6516 * PKCS7 NULL pointer dereferences fix
6518 The PKCS#7 parsing code does not handle missing outer ContentInfo
6519 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6520 missing content and trigger a NULL pointer dereference on parsing.
6522 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6523 otherwise parse PKCS#7 structures from untrusted sources are
6524 affected. OpenSSL clients and servers are not affected.
6526 This issue was reported to OpenSSL by Michal Zalewski (Google).
6531 * DoS via reachable assert in SSLv2 servers fix
6533 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6534 servers that both support SSLv2 and enable export cipher suites by sending
6535 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6537 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6538 (OpenSSL development team).
6543 * Empty CKE with client auth and DHE fix
6545 If client auth is used then a server can seg fault in the event of a DHE
6546 ciphersuite being selected and a zero length ClientKeyExchange message
6547 being sent by the client. This could be exploited in a DoS attack.
6552 * Handshake with unseeded PRNG fix
6554 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6555 with an unseeded PRNG. The conditions are:
6556 - The client is on a platform where the PRNG has not been seeded
6557 automatically, and the user has not seeded manually
6558 - A protocol specific client method version has been used (i.e. not
6559 SSL_client_methodv23)
6560 - A ciphersuite is used that does not require additional random data from
6561 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6563 If the handshake succeeds then the client random that has been used will
6564 have been generated from a PRNG with insufficient entropy and therefore the
6565 output may be predictable.
6567 For example using the following command with an unseeded openssl will
6568 succeed on an unpatched platform:
6570 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6575 * Use After Free following d2i_ECPrivatekey error fix
6577 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6578 could cause a use after free condition. This, in turn, could cause a double
6579 free in several private key parsing functions (such as d2i_PrivateKey
6580 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6581 for applications that receive EC private keys from untrusted
6582 sources. This scenario is considered rare.
6584 This issue was discovered by the BoringSSL project and fixed in their
6590 * X509_to_X509_REQ NULL pointer deref fix
6592 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6593 the certificate key is invalid. This function is rarely used in practice.
6595 This issue was discovered by Brian Carpenter.
6600 * Removed the export ciphers from the DEFAULT ciphers
6604 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6606 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6607 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6608 So far those who have to target multiple platforms would compromise
6609 and argue that binary targeting say ARMv5 would still execute on
6610 ARMv8. "Universal" build resolves this compromise by providing
6611 near-optimal performance even on newer platforms.
6615 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6616 (other platforms pending).
6618 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6620 * Add support for the SignedCertificateTimestampList certificate and
6621 OCSP response extensions from RFC6962.
6625 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6626 for corner cases. (Certain input points at infinity could lead to
6627 bogus results, with non-infinity inputs mapped to infinity too.)
6631 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6632 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6633 common cases are optimized and there still is room for further
6634 improvements. Vector Permutation AES for Altivec is also added.
6638 * Add support for little-endian ppc64 Linux target.
6640 *Marcelo Cerri (IBM)*
6642 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6643 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6644 are optimized and there still is room for further improvements.
6645 Both 32- and 64-bit modes are supported.
6647 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6649 * Improved ARMv7 NEON support.
6653 * Support for SPARC Architecture 2011 crypto extensions, first
6654 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6655 SHA256/512, MD5, GHASH and modular exponentiation.
6657 *Andy Polyakov, David Miller*
6659 * Accelerated modular exponentiation for Intel processors, a.k.a.
6662 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6664 * Support for new and upcoming Intel processors, including AVX2,
6665 BMI and SHA ISA extensions. This includes additional "stitched"
6666 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6669 This work was sponsored by Intel Corp.
6673 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6674 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6675 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6679 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6680 this fixes a limitation in previous versions of OpenSSL.
6684 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6685 MGF1 digest and OAEP label.
6689 * Add EVP support for key wrapping algorithms, to avoid problems with
6690 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6691 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6692 algorithms and include tests cases.
6696 * Add functions to allocate and set the fields of an ECDSA_METHOD
6699 *Douglas E. Engert, Steve Henson*
6701 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6702 difference in days and seconds between two tm or ASN1_TIME structures.
6706 * Add -rev test option to s_server to just reverse order of characters
6707 received by client and send back to server. Also prints an abbreviated
6708 summary of the connection parameters.
6712 * New option -brief for s_client and s_server to print out a brief summary
6713 of connection parameters.
6717 * Add callbacks for arbitrary TLS extensions.
6719 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6721 * New option -crl_download in several openssl utilities to download CRLs
6722 from CRLDP extension in certificates.
6726 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6730 * New function X509_CRL_diff to generate a delta CRL from the difference
6731 of two full CRLs. Add support to "crl" utility.
6735 * New functions to set lookup_crls function and to retrieve
6736 X509_STORE from X509_STORE_CTX.
6740 * Print out deprecated issuer and subject unique ID fields in
6745 * Extend OCSP I/O functions so they can be used for simple general purpose
6746 HTTP as well as OCSP. New wrapper function which can be used to download
6747 CRLs using the OCSP API.
6751 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6755 * `SSL_CONF*` functions. These provide a common framework for application
6756 configuration using configuration files or command lines.
6760 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6761 message callback and prints the results. Needs compile time option
6762 "enable-ssl-trace". New options to s_client and s_server to enable
6767 * New ctrl and macro to retrieve supported points extensions.
6768 Print out extension in s_server and s_client.
6772 * New functions to retrieve certificate signature and signature
6777 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6782 * New Suite B modes for TLS code. These use and enforce the requirements
6783 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6784 only use Suite B curves. The Suite B modes can be set by using the
6785 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6789 * New chain verification flags for Suite B levels of security. Check
6790 algorithms are acceptable when flags are set in X509_verify_cert.
6794 * Make tls1_check_chain return a set of flags indicating checks passed
6795 by a certificate chain. Add additional tests to handle client
6796 certificates: checks for matching certificate type and issuer name
6801 * If an attempt is made to use a signature algorithm not in the peer
6802 preference list abort the handshake. If client has no suitable
6803 signature algorithms in response to a certificate request do not
6804 use the certificate.
6808 * If server EC tmp key is not in client preference list abort handshake.
6812 * Add support for certificate stores in CERT structure. This makes it
6813 possible to have different stores per SSL structure or one store in
6814 the parent SSL_CTX. Include distinct stores for certificate chain
6815 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6816 to build and store a certificate chain in CERT structure: returning
6817 an error if the chain cannot be built: this will allow applications
6818 to test if a chain is correctly configured.
6820 Note: if the CERT based stores are not set then the parent SSL_CTX
6821 store is used to retain compatibility with existing behaviour.
6825 * New function ssl_set_client_disabled to set a ciphersuite disabled
6826 mask based on the current session, check mask when sending client
6827 hello and checking the requested ciphersuite.
6831 * New ctrls to retrieve and set certificate types in a certificate
6832 request message. Print out received values in s_client. If certificate
6833 types is not set with custom values set sensible values based on
6834 supported signature algorithms.
6838 * Support for distinct client and server supported signature algorithms.
6842 * Add certificate callback. If set this is called whenever a certificate
6843 is required by client or server. An application can decide which
6844 certificate chain to present based on arbitrary criteria: for example
6845 supported signature algorithms. Add very simple example to s_server.
6846 This fixes many of the problems and restrictions of the existing client
6847 certificate callback: for example you can now clear an existing
6848 certificate and specify the whole chain.
6852 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6853 the certificate can be used for (if anything). Set valid_flags field
6854 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6855 to have similar checks in it.
6857 Add new "cert_flags" field to CERT structure and include a "strict mode".
6858 This enforces some TLS certificate requirements (such as only permitting
6859 certificate signature algorithms contained in the supported algorithms
6860 extension) which some implementations ignore: this option should be used
6861 with caution as it could cause interoperability issues.
6865 * Update and tidy signature algorithm extension processing. Work out
6866 shared signature algorithms based on preferences and peer algorithms
6867 and print them out in s_client and s_server. Abort handshake if no
6868 shared signature algorithms.
6872 * Add new functions to allow customised supported signature algorithms
6873 for SSL and SSL_CTX structures. Add options to s_client and s_server
6878 * New function SSL_certs_clear() to delete all references to certificates
6879 from an SSL structure. Before this once a certificate had been added
6880 it couldn't be removed.
6884 * Integrate hostname, email address and IP address checking with certificate
6885 verification. New verify options supporting checking in openssl utility.
6889 * Fixes and wildcard matching support to hostname and email checking
6890 functions. Add manual page.
6892 *Florian Weimer (Red Hat Product Security Team)*
6894 * New functions to check a hostname email or IP address against a
6895 certificate. Add options x509 utility to print results of checks against
6900 * Fix OCSP checking.
6902 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6904 * Initial experimental support for explicitly trusted non-root CAs.
6905 OpenSSL still tries to build a complete chain to a root but if an
6906 intermediate CA has a trust setting included that is used. The first
6907 setting is used: whether to trust (e.g., -addtrust option to the x509
6912 * Add -trusted_first option which attempts to find certificates in the
6913 trusted store even if an untrusted chain is also supplied.
6917 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6918 platform support for Linux and Android.
6922 * Support for linux-x32, ILP32 environment in x86_64 framework.
6926 * Experimental multi-implementation support for FIPS capable OpenSSL.
6927 When in FIPS mode the approved implementations are used as normal,
6928 when not in FIPS mode the internal unapproved versions are used instead.
6929 This means that the FIPS capable OpenSSL isn't forced to use the
6930 (often lower performance) FIPS implementations outside FIPS mode.
6934 * Transparently support X9.42 DH parameters when calling
6935 PEM_read_bio_DHparameters. This means existing applications can handle
6936 the new parameter format automatically.
6940 * Initial experimental support for X9.42 DH parameter format: mainly
6941 to support use of 'q' parameter for RFC5114 parameters.
6945 * Add DH parameters from RFC5114 including test data to dhtest.
6949 * Support for automatic EC temporary key parameter selection. If enabled
6950 the most preferred EC parameters are automatically used instead of
6951 hardcoded fixed parameters. Now a server just has to call:
6952 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6953 support ECDH and use the most appropriate parameters.
6957 * Enhance and tidy EC curve and point format TLS extension code. Use
6958 static structures instead of allocation if default values are used.
6959 New ctrls to set curves we wish to support and to retrieve shared curves.
6960 Print out shared curves in s_server. New options to s_server and s_client
6961 to set list of supported curves.
6965 * New ctrls to retrieve supported signature algorithms and
6966 supported curve values as an array of NIDs. Extend openssl utility
6967 to print out received values.
6971 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6972 between NIDs and the more common NIST names such as "P-256". Enhance
6973 ecparam utility and ECC method to recognise the NIST names for curves.
6977 * Enhance SSL/TLS certificate chain handling to support different
6978 chains for each certificate instead of one chain in the parent SSL_CTX.
6982 * Support for fixed DH ciphersuite client authentication: where both
6983 server and client use DH certificates with common parameters.
6987 * Support for fixed DH ciphersuites: those requiring DH server
6992 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6994 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6995 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6996 X509_CINF_get_signature were reverted post internal team review.
7001 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
7003 * OCSP Status Request extension unbounded memory growth
7005 A malicious client can send an excessively large OCSP Status Request
7006 extension. If that client continually requests renegotiation, sending a
7007 large OCSP Status Request extension each time, then there will be unbounded
7008 memory growth on the server. This will eventually lead to a Denial Of
7009 Service attack through memory exhaustion. Servers with a default
7010 configuration are vulnerable even if they do not support OCSP. Builds using
7011 the "no-ocsp" build time option are not affected.
7013 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7018 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
7021 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
7027 * OOB write in MDC2_Update()
7029 An overflow can occur in MDC2_Update() either if called directly or
7030 through the EVP_DigestUpdate() function using MDC2. If an attacker
7031 is able to supply very large amounts of input data after a previous
7032 call to EVP_EncryptUpdate() with a partial block then a length check
7033 can overflow resulting in a heap corruption.
7035 The amount of data needed is comparable to SIZE_MAX which is impractical
7038 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7043 * Malformed SHA512 ticket DoS
7045 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
7046 DoS attack where a malformed ticket will result in an OOB read which will
7049 The use of SHA512 in TLS session tickets is comparatively rare as it requires
7050 a custom server callback and ticket lookup mechanism.
7052 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7057 * OOB write in BN_bn2dec()
7059 The function BN_bn2dec() does not check the return value of BN_div_word().
7060 This can cause an OOB write if an application uses this function with an
7061 overly large BIGNUM. This could be a problem if an overly large certificate
7062 or CRL is printed out from an untrusted source. TLS is not affected because
7063 record limits will reject an oversized certificate before it is parsed.
7065 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7070 * OOB read in TS_OBJ_print_bio()
7072 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
7073 the total length the OID text representation would use and not the amount
7074 of data written. This will result in OOB reads when large OIDs are
7077 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7082 * Pointer arithmetic undefined behaviour
7084 Avoid some undefined pointer arithmetic
7086 A common idiom in the codebase is to check limits in the following manner:
7089 Where "p" points to some malloc'd data of SIZE bytes and
7092 "len" here could be from some externally supplied data (e.g. from a TLS
7095 The rules of C pointer arithmetic are such that "p + len" is only well
7096 defined where len <= SIZE. Therefore, the above idiom is actually
7097 undefined behaviour.
7099 For example this could cause problems if some malloc implementation
7100 provides an address for "p" such that "p + len" actually overflows for
7101 values of len that are too big and therefore p + len < limit.
7103 This issue was reported to OpenSSL by Guido Vranken
7108 * Constant time flag not preserved in DSA signing
7110 Operations in the DSA signing algorithm should run in constant time in
7111 order to avoid side channel attacks. A flaw in the OpenSSL DSA
7112 implementation means that a non-constant time codepath is followed for
7113 certain operations. This has been demonstrated through a cache-timing
7114 attack to be sufficient for an attacker to recover the private DSA key.
7116 This issue was reported by César Pereida (Aalto University), Billy Brumley
7117 (Tampere University of Technology), and Yuval Yarom (The University of
7118 Adelaide and NICTA).
7123 * DTLS buffered message DoS
7125 In a DTLS connection where handshake messages are delivered out-of-order
7126 those messages that OpenSSL is not yet ready to process will be buffered
7127 for later use. Under certain circumstances, a flaw in the logic means that
7128 those messages do not get removed from the buffer even though the handshake
7129 has been completed. An attacker could force up to approx. 15 messages to
7130 remain in the buffer when they are no longer required. These messages will
7131 be cleared when the DTLS connection is closed. The default maximum size for
7132 a message is 100k. Therefore, the attacker could force an additional 1500k
7133 to be consumed per connection. By opening many simultaneous connections an
7134 attacker could cause a DoS attack through memory exhaustion.
7136 This issue was reported to OpenSSL by Quan Luo.
7141 * DTLS replay protection DoS
7143 A flaw in the DTLS replay attack protection mechanism means that records
7144 that arrive for future epochs update the replay protection "window" before
7145 the MAC for the record has been validated. This could be exploited by an
7146 attacker by sending a record for the next epoch (which does not have to
7147 decrypt or have a valid MAC), with a very large sequence number. This means
7148 that all subsequent legitimate packets are dropped causing a denial of
7149 service for a specific DTLS connection.
7151 This issue was reported to OpenSSL by the OCAP audit team.
7156 * Certificate message OOB reads
7158 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7159 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7160 theoretical DoS risk but this has not been observed in practice on common
7163 The messages affected are client certificate, client certificate request
7164 and server certificate. As a result the attack can only be performed
7165 against a client or a server which enables client authentication.
7167 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7172 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7174 * Prevent padding oracle in AES-NI CBC MAC check
7176 A MITM attacker can use a padding oracle attack to decrypt traffic
7177 when the connection uses an AES CBC cipher and the server support
7180 This issue was introduced as part of the fix for Lucky 13 padding
7181 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7182 constant time by making sure that always the same bytes are read and
7183 compared against either the MAC or padding bytes. But it no longer
7184 checked that there was enough data to have both the MAC and padding
7187 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7192 * Fix EVP_EncodeUpdate overflow
7194 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7195 Base64 encoding of binary data. If an attacker is able to supply very large
7196 amounts of input data then a length check can overflow resulting in a heap
7199 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7200 the `PEM_write_bio*` family of functions. These are mainly used within the
7201 OpenSSL command line applications, so any application which processes data
7202 from an untrusted source and outputs it as a PEM file should be considered
7203 vulnerable to this issue. User applications that call these APIs directly
7204 with large amounts of untrusted data may also be vulnerable.
7206 This issue was reported by Guido Vranken.
7211 * Fix EVP_EncryptUpdate overflow
7213 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7214 is able to supply very large amounts of input data after a previous call to
7215 EVP_EncryptUpdate() with a partial block then a length check can overflow
7216 resulting in a heap corruption. Following an analysis of all OpenSSL
7217 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7218 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7219 the first called function after an EVP_EncryptInit(), and therefore that
7220 specific call must be safe. The second form is where the length passed to
7221 EVP_EncryptUpdate() can be seen from the code to be some small value and
7222 therefore there is no possibility of an overflow. Since all instances are
7223 one of these two forms, it is believed that there can be no overflows in
7224 internal code due to this problem. It should be noted that
7225 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7226 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7227 of these calls have also been analysed too and it is believed there are no
7228 instances in internal usage where an overflow could occur.
7230 This issue was reported by Guido Vranken.
7235 * Prevent ASN.1 BIO excessive memory allocation
7237 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7238 a short invalid encoding can casuse allocation of large amounts of memory
7239 potentially consuming excessive resources or exhausting memory.
7241 Any application parsing untrusted data through d2i BIO functions is
7242 affected. The memory based functions such as d2i_X509() are *not* affected.
7243 Since the memory based functions are used by the TLS library, TLS
7244 applications are not affected.
7246 This issue was reported by Brian Carpenter.
7253 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7254 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7255 in arbitrary stack data being returned in the buffer.
7257 This issue was reported by Guido Vranken.
7262 * Modify behavior of ALPN to invoke callback after SNI/servername
7263 callback, such that updates to the SSL_CTX affect ALPN.
7267 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7272 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7273 methods are enabled and ssl2 is disabled the methods return NULL.
7277 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7279 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7280 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7281 provide any "EXPORT" or "LOW" strength ciphers.
7285 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7286 is by default disabled at build-time. Builds that are not configured with
7287 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7288 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7289 will need to explicitly call either of:
7291 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7293 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7295 as appropriate. Even if either of those is used, or the application
7296 explicitly uses the version-specific SSLv2_method() or its client and
7297 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7298 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7299 ciphers, and SSLv2 56-bit DES are no longer available.
7304 * Fix a double-free in DSA code
7306 A double free bug was discovered when OpenSSL parses malformed DSA private
7307 keys and could lead to a DoS attack or memory corruption for applications
7308 that receive DSA private keys from untrusted sources. This scenario is
7311 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7317 * Disable SRP fake user seed to address a server memory leak.
7319 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7321 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7322 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7323 was changed to ignore the "fake user" SRP seed, even if the seed
7326 Users should use SRP_VBASE_get1_by_user instead. Note that in
7327 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7328 also that even though configuring the SRP seed attempts to hide
7329 invalid usernames by continuing the handshake with fake
7330 credentials, this behaviour is not constant time and no strong
7331 guarantees are made that the handshake is indistinguishable from
7332 that of a valid user.
7337 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7339 In the BN_hex2bn function the number of hex digits is calculated using an
7340 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7341 large values of `i` this can result in `bn_expand` not allocating any
7342 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7343 field as NULL leading to a subsequent NULL ptr deref. For very large values
7344 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7345 In this case memory is allocated to the internal BIGNUM data field, but it
7346 is insufficiently sized leading to heap corruption. A similar issue exists
7347 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7348 is ever called by user applications with very large untrusted hex/dec data.
7349 This is anticipated to be a rare occurrence.
7351 All OpenSSL internal usage of these functions use data that is not expected
7352 to be untrusted, e.g. config file data or application command line
7353 arguments. If user developed applications generate config file data based
7354 on untrusted data then it is possible that this could also lead to security
7355 consequences. This is also anticipated to be rare.
7357 This issue was reported to OpenSSL by Guido Vranken.
7362 * Fix memory issues in `BIO_*printf` functions
7364 The internal `fmtstr` function used in processing a "%s" format string in
7365 the `BIO_*printf` functions could overflow while calculating the length of a
7366 string and cause an OOB read when printing very long strings.
7368 Additionally the internal `doapr_outch` function can attempt to write to an
7369 OOB memory location (at an offset from the NULL pointer) in the event of a
7370 memory allocation failure. In 1.0.2 and below this could be caused where
7371 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7372 could be in processing a very long "%s" format string. Memory leaks can
7375 The first issue may mask the second issue dependent on compiler behaviour.
7376 These problems could enable attacks where large amounts of untrusted data
7377 is passed to the `BIO_*printf` functions. If applications use these functions
7378 in this way then they could be vulnerable. OpenSSL itself uses these
7379 functions when printing out human-readable dumps of ASN.1 data. Therefore
7380 applications that print this data could be vulnerable if the data is from
7381 untrusted sources. OpenSSL command line applications could also be
7382 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7383 as command line arguments.
7385 Libssl is not considered directly vulnerable. Additionally certificates etc
7386 received via remote connections via libssl are also unlikely to be able to
7387 trigger these issues because of message size limits enforced within libssl.
7389 This issue was reported to OpenSSL Guido Vranken.
7394 * Side channel attack on modular exponentiation
7396 A side-channel attack was found which makes use of cache-bank conflicts on
7397 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7398 of RSA keys. The ability to exploit this issue is limited as it relies on
7399 an attacker who has control of code in a thread running on the same
7400 hyper-threaded core as the victim thread which is performing decryptions.
7402 This issue was reported to OpenSSL by Yuval Yarom, The University of
7403 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7404 Nadia Heninger, University of Pennsylvania with more information at
7405 <http://cachebleed.info>.
7410 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7411 if no keysize is specified with default_bits. This fixes an
7412 omission in an earlier change that changed all RSA/DSA key generation
7413 commands to use 2048 bits by default.
7417 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7419 * Protection for DH small subgroup attacks
7421 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7422 switched on by default and cannot be disabled. This could have some
7427 * SSLv2 doesn't block disabled ciphers
7429 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7430 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7431 been disabled, provided that the SSLv2 protocol was not also disabled via
7434 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7435 and Sebastian Schinzel.
7440 * Reject DH handshakes with parameters shorter than 1024 bits.
7444 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7446 * Certificate verify crash with missing PSS parameter
7448 The signature verification routines will crash with a NULL pointer
7449 dereference if presented with an ASN.1 signature using the RSA PSS
7450 algorithm and absent mask generation function parameter. Since these
7451 routines are used to verify certificate signature algorithms this can be
7452 used to crash any certificate verification operation and exploited in a
7453 DoS attack. Any application which performs certificate verification is
7454 vulnerable including OpenSSL clients and servers which enable client
7457 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7462 * X509_ATTRIBUTE memory leak
7464 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7465 memory. This structure is used by the PKCS#7 and CMS routines so any
7466 application which reads PKCS#7 or CMS data from untrusted sources is
7467 affected. SSL/TLS is not affected.
7469 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7475 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7476 This changes the decoding behaviour for some invalid messages,
7477 though the change is mostly in the more lenient direction, and
7478 legacy behaviour is preserved as much as possible.
7482 * In DSA_generate_parameters_ex, if the provided seed is too short,
7483 use a random seed, as already documented.
7485 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7487 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7489 * Alternate chains certificate forgery
7491 During certificate verification, OpenSSL will attempt to find an
7492 alternative certificate chain if the first attempt to build such a chain
7493 fails. An error in the implementation of this logic can mean that an
7494 attacker could cause certain checks on untrusted certificates to be
7495 bypassed, such as the CA flag, enabling them to use a valid leaf
7496 certificate to act as a CA and "issue" an invalid certificate.
7498 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7504 * Race condition handling PSK identify hint
7506 If PSK identity hints are received by a multi-threaded client then
7507 the values are wrongly updated in the parent SSL_CTX structure. This can
7508 result in a race condition potentially leading to a double free of the
7514 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7516 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7517 incompatibility in the handling of HMAC. The previous ABI has now been
7520 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7522 * Malformed ECParameters causes infinite loop
7524 When processing an ECParameters structure OpenSSL enters an infinite loop
7525 if the curve specified is over a specially malformed binary polynomial
7528 This can be used to perform denial of service against any
7529 system which processes public keys, certificate requests or
7530 certificates. This includes TLS clients and TLS servers with
7531 client authentication enabled.
7533 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7538 * Exploitable out-of-bounds read in X509_cmp_time
7540 X509_cmp_time does not properly check the length of the ASN1_TIME
7541 string and can read a few bytes out of bounds. In addition,
7542 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7545 An attacker can use this to craft malformed certificates and CRLs of
7546 various sizes and potentially cause a segmentation fault, resulting in
7547 a DoS on applications that verify certificates or CRLs. TLS clients
7548 that verify CRLs are affected. TLS clients and servers with client
7549 authentication enabled may be affected if they use custom verification
7552 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7553 independently by Hanno Böck.
7558 * PKCS7 crash with missing EnvelopedContent
7560 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7561 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7562 with missing content and trigger a NULL pointer dereference on parsing.
7564 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7565 structures from untrusted sources are affected. OpenSSL clients and
7566 servers are not affected.
7568 This issue was reported to OpenSSL by Michal Zalewski (Google).
7573 * CMS verify infinite loop with unknown hash function
7575 When verifying a signedData message the CMS code can enter an infinite loop
7576 if presented with an unknown hash function OID. This can be used to perform
7577 denial of service against any system which verifies signedData messages using
7579 This issue was reported to OpenSSL by Johannes Bauer.
7584 * Race condition handling NewSessionTicket
7586 If a NewSessionTicket is received by a multi-threaded client when attempting to
7587 reuse a previous ticket then a race condition can occur potentially leading to
7588 a double free of the ticket data.
7593 * Reject DH handshakes with parameters shorter than 768 bits.
7595 *Kurt Roeckx and Emilia Kasper*
7597 * dhparam: generate 2048-bit parameters by default.
7599 *Kurt Roeckx and Emilia Kasper*
7601 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7603 * Segmentation fault in ASN1_TYPE_cmp fix
7605 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7606 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7607 certificate signature algorithm consistency this can be used to crash any
7608 certificate verification operation and exploited in a DoS attack. Any
7609 application which performs certificate verification is vulnerable including
7610 OpenSSL clients and servers which enable client authentication.
7615 * ASN.1 structure reuse memory corruption fix
7617 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7618 memory corruption via an invalid write. Such reuse is and has been
7619 strongly discouraged and is believed to be rare.
7621 Applications that parse structures containing CHOICE or ANY DEFINED BY
7622 components may be affected. Certificate parsing (d2i_X509 and related
7623 functions) are however not affected. OpenSSL clients and servers are
7629 * PKCS7 NULL pointer dereferences fix
7631 The PKCS#7 parsing code does not handle missing outer ContentInfo
7632 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7633 missing content and trigger a NULL pointer dereference on parsing.
7635 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7636 otherwise parse PKCS#7 structures from untrusted sources are
7637 affected. OpenSSL clients and servers are not affected.
7639 This issue was reported to OpenSSL by Michal Zalewski (Google).
7644 * DoS via reachable assert in SSLv2 servers fix
7646 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7647 servers that both support SSLv2 and enable export cipher suites by sending
7648 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7650 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7651 (OpenSSL development team).
7656 * Use After Free following d2i_ECPrivatekey error fix
7658 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7659 could cause a use after free condition. This, in turn, could cause a double
7660 free in several private key parsing functions (such as d2i_PrivateKey
7661 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7662 for applications that receive EC private keys from untrusted
7663 sources. This scenario is considered rare.
7665 This issue was discovered by the BoringSSL project and fixed in their
7671 * X509_to_X509_REQ NULL pointer deref fix
7673 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7674 the certificate key is invalid. This function is rarely used in practice.
7676 This issue was discovered by Brian Carpenter.
7681 * Removed the export ciphers from the DEFAULT ciphers
7685 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7687 * Build fixes for the Windows and OpenVMS platforms
7689 *Matt Caswell and Richard Levitte*
7691 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7693 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7694 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7695 dereference. This could lead to a Denial Of Service attack. Thanks to
7696 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7701 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7702 dtls1_buffer_record function under certain conditions. In particular this
7703 could occur if an attacker sent repeated DTLS records with the same
7704 sequence number but for the next epoch. The memory leak could be exploited
7705 by an attacker in a Denial of Service attack through memory exhaustion.
7706 Thanks to Chris Mueller for reporting this issue.
7711 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7712 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7713 method would be set to NULL which could later result in a NULL pointer
7714 dereference. Thanks to Frank Schmirler for reporting this issue.
7719 * Abort handshake if server key exchange message is omitted for ephemeral
7722 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7723 reporting this issue.
7728 * Remove non-export ephemeral RSA code on client and server. This code
7729 violated the TLS standard by allowing the use of temporary RSA keys in
7730 non-export ciphersuites and could be used by a server to effectively
7731 downgrade the RSA key length used to a value smaller than the server
7732 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7733 INRIA or reporting this issue.
7738 * Fixed issue where DH client certificates are accepted without verification.
7739 An OpenSSL server will accept a DH certificate for client authentication
7740 without the certificate verify message. This effectively allows a client to
7741 authenticate without the use of a private key. This only affects servers
7742 which trust a client certificate authority which issues certificates
7743 containing DH keys: these are extremely rare and hardly ever encountered.
7744 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7750 * Ensure that the session ID context of an SSL is updated when its
7751 SSL_CTX is updated via SSL_set_SSL_CTX.
7753 The session ID context is typically set from the parent SSL_CTX,
7754 and can vary with the CTX.
7758 * Fix various certificate fingerprint issues.
7760 By using non-DER or invalid encodings outside the signed portion of a
7761 certificate the fingerprint can be changed without breaking the signature.
7762 Although no details of the signed portion of the certificate can be changed
7763 this can cause problems with some applications: e.g. those using the
7764 certificate fingerprint for blacklists.
7766 1. Reject signatures with non zero unused bits.
7768 If the BIT STRING containing the signature has non zero unused bits reject
7769 the signature. All current signature algorithms require zero unused bits.
7771 2. Check certificate algorithm consistency.
7773 Check the AlgorithmIdentifier inside TBS matches the one in the
7774 certificate signature. NB: this will result in signature failure
7775 errors for some broken certificates.
7777 Thanks to Konrad Kraszewski from Google for reporting this issue.
7779 3. Check DSA/ECDSA signatures use DER.
7781 Re-encode DSA/ECDSA signatures and compare with the original received
7782 signature. Return an error if there is a mismatch.
7784 This will reject various cases including garbage after signature
7785 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7786 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7787 (negative or with leading zeroes).
7789 Further analysis was conducted and fixes were developed by Stephen Henson
7790 of the OpenSSL core team.
7796 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7797 results on some platforms, including x86_64. This bug occurs at random
7798 with a very low probability, and is not known to be exploitable in any
7799 way, though its exact impact is difficult to determine. Thanks to Pieter
7800 Wuille (Blockstream) who reported this issue and also suggested an initial
7801 fix. Further analysis was conducted by the OpenSSL development team and
7802 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7803 the OpenSSL core team.
7808 * Do not resume sessions on the server if the negotiated protocol
7809 version does not match the session's version. Resuming with a different
7810 version, while not strictly forbidden by the RFC, is of questionable
7811 sanity and breaks all known clients.
7813 *David Benjamin, Emilia Käsper*
7815 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7816 early CCS messages during renegotiation. (Note that because
7817 renegotiation is encrypted, this early CCS was not exploitable.)
7821 * Tighten client-side session ticket handling during renegotiation:
7822 ensure that the client only accepts a session ticket if the server sends
7823 the extension anew in the ServerHello. Previously, a TLS client would
7824 reuse the old extension state and thus accept a session ticket if one was
7825 announced in the initial ServerHello.
7827 Similarly, ensure that the client requires a session ticket if one
7828 was advertised in the ServerHello. Previously, a TLS client would
7829 ignore a missing NewSessionTicket message.
7833 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7837 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7838 sends a carefully crafted handshake message, to cause OpenSSL to fail
7839 to free up to 64k of memory causing a memory leak. This could be
7840 exploited in a Denial Of Service attack. This issue affects OpenSSL
7841 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7842 whether SRTP is used or configured. Implementations of OpenSSL that
7843 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7845 The fix was developed by the OpenSSL team.
7850 * Session Ticket Memory Leak.
7852 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7853 integrity of that ticket is first verified. In the event of a session
7854 ticket integrity check failing, OpenSSL will fail to free memory
7855 causing a memory leak. By sending a large number of invalid session
7856 tickets an attacker could exploit this issue in a Denial Of Service
7862 * Build option no-ssl3 is incomplete.
7864 When OpenSSL is configured with "no-ssl3" as a build option, servers
7865 could accept and complete an SSL 3.0 handshake, and clients could be
7866 configured to send them.
7869 *Akamai and the OpenSSL team*
7871 * Add support for TLS_FALLBACK_SCSV.
7872 Client applications doing fallback retries should call
7873 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7876 *Adam Langley, Bodo Moeller*
7878 * Add additional DigestInfo checks.
7880 Re-encode DigestInto in DER and check against the original when
7881 verifying RSA signature: this will reject any improperly encoded
7882 DigestInfo structures.
7884 Note: this is a precautionary measure and no attacks are currently known.
7888 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7890 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7891 SRP code can be overrun an internal buffer. Add sanity check that
7892 g, A, B < N to SRP code.
7894 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7895 Group for discovering this issue.
7900 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7901 TLS 1.0 instead of higher protocol versions when the ClientHello message
7902 is badly fragmented. This allows a man-in-the-middle attacker to force a
7903 downgrade to TLS 1.0 even if both the server and the client support a
7904 higher protocol version, by modifying the client's TLS records.
7906 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7907 researching this issue.
7912 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7913 to a denial of service attack. A malicious server can crash the client
7914 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7915 ciphersuite and sending carefully crafted handshake messages.
7917 Thanks to Felix Gröbert (Google) for discovering and researching this
7923 * By sending carefully crafted DTLS packets an attacker could cause openssl
7924 to leak memory. This can be exploited through a Denial of Service attack.
7925 Thanks to Adam Langley for discovering and researching this issue.
7930 * An attacker can force openssl to consume large amounts of memory whilst
7931 processing DTLS handshake messages. This can be exploited through a
7932 Denial of Service attack.
7933 Thanks to Adam Langley for discovering and researching this issue.
7938 * An attacker can force an error condition which causes openssl to crash
7939 whilst processing DTLS packets due to memory being freed twice. This
7940 can be exploited through a Denial of Service attack.
7941 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7947 * If a multithreaded client connects to a malicious server using a resumed
7948 session and the server sends an ec point format extension it could write
7949 up to 255 bytes to freed memory.
7951 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7957 * A malicious server can crash an OpenSSL client with a null pointer
7958 dereference (read) by specifying an SRP ciphersuite even though it was not
7959 properly negotiated with the client. This can be exploited through a
7960 Denial of Service attack.
7962 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7963 discovering and researching this issue.
7968 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7969 X509_name_oneline, X509_name_print_ex et al. to leak some information
7970 from the stack. Applications may be affected if they echo pretty printing
7971 output to the attacker.
7973 Thanks to Ivan Fratric (Google) for discovering this issue.
7976 *Emilia Käsper, and Steve Henson*
7978 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7979 for corner cases. (Certain input points at infinity could lead to
7980 bogus results, with non-infinity inputs mapped to infinity too.)
7984 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7986 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7987 handshake can force the use of weak keying material in OpenSSL
7988 SSL/TLS clients and servers.
7990 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7991 researching this issue. ([CVE-2014-0224])
7993 *KIKUCHI Masashi, Steve Henson*
7995 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7996 OpenSSL DTLS client the code can be made to recurse eventually crashing
7999 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8002 *Imre Rad, Steve Henson*
8004 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8005 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8006 client or server. This is potentially exploitable to run arbitrary
8007 code on a vulnerable client or server.
8009 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8011 *Jüri Aedla, Steve Henson*
8013 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8014 are subject to a denial of service attack.
8016 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8017 this issue. ([CVE-2014-3470])
8019 *Felix Gröbert, Ivan Fratric, Steve Henson*
8021 * Harmonize version and its documentation. -f flag is used to display
8024 *mancha <mancha1@zoho.com>*
8026 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8027 in i2d_ECPrivateKey.
8029 *mancha <mancha1@zoho.com>*
8031 * Fix some double frees. These are not thought to be exploitable.
8033 *mancha <mancha1@zoho.com>*
8035 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
8037 * A missing bounds check in the handling of the TLS heartbeat extension
8038 can be used to reveal up to 64k of memory to a connected client or
8041 Thanks for Neel Mehta of Google Security for discovering this bug and to
8042 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
8043 preparing the fix ([CVE-2014-0160])
8045 *Adam Langley, Bodo Moeller*
8047 * Fix for the attack described in the paper "Recovering OpenSSL
8048 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8049 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8050 <http://eprint.iacr.org/2014/140>
8052 Thanks to Yuval Yarom and Naomi Benger for discovering this
8053 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8055 *Yuval Yarom and Naomi Benger*
8057 * TLS pad extension: draft-agl-tls-padding-03
8059 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
8060 TLS client Hello record length value would otherwise be > 255 and
8061 less that 512 pad with a dummy extension containing zeroes so it
8062 is at least 512 bytes long.
8064 *Adam Langley, Steve Henson*
8066 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
8068 * Fix for TLS record tampering bug. A carefully crafted invalid
8069 handshake could crash OpenSSL with a NULL pointer exception.
8070 Thanks to Anton Johansson for reporting this issues.
8073 * Keep original DTLS digest and encryption contexts in retransmission
8074 structures so we can use the previous session parameters if they need
8075 to be resent. ([CVE-2013-6450])
8079 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8080 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8081 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8082 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8083 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8084 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8086 *Rob Stradling, Adam Langley*
8088 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
8090 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
8091 supporting platforms or when small records were transferred.
8093 *Andy Polyakov, Steve Henson*
8095 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
8097 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8099 This addresses the flaw in CBC record processing discovered by
8100 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8101 at: <http://www.isg.rhul.ac.uk/tls/>
8103 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8104 Security Group at Royal Holloway, University of London
8105 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8106 Emilia Käsper for the initial patch.
8109 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8111 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
8112 ciphersuites which can be exploited in a denial of service attack.
8113 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
8114 and detecting this bug and to Wolfgang Ettlinger
8115 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
8120 * Return an error when checking OCSP signatures when key is NULL.
8121 This fixes a DoS attack. ([CVE-2013-0166])
8125 * Make openssl verify return errors.
8127 *Chris Palmer <palmer@google.com> and Ben Laurie*
8129 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8130 the right response is stapled. Also change SSL_get_certificate()
8131 so it returns the certificate actually sent.
8132 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8134 *Rob Stradling <rob.stradling@comodo.com>*
8136 * Fix possible deadlock when decoding public keys.
8140 * Don't use TLS 1.0 record version number in initial client hello
8145 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
8147 * Sanity check record length before skipping explicit IV in TLS
8148 1.2, 1.1 and DTLS to fix DoS attack.
8150 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8151 fuzzing as a service testing platform.
8156 * Initialise tkeylen properly when encrypting CMS messages.
8157 Thanks to Solar Designer of Openwall for reporting this issue.
8161 * In FIPS mode don't try to use composite ciphers as they are not
8166 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8168 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8169 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8170 mean any application compiled against OpenSSL 1.0.0 headers setting
8171 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8172 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8173 0x10000000L Any application which was previously compiled against
8174 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8175 will need to be recompiled as a result. Letting be results in
8176 inability to disable specifically TLS 1.1 and in client context,
8177 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8181 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8182 disable just protocol X, but all protocols above X *if* there are
8183 protocols *below* X still enabled. In more practical terms it means
8184 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8185 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8186 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8191 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8193 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8194 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8195 in CRYPTO_realloc_clean.
8197 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8198 issue and to Adam Langley <agl@chromium.org> for fixing it.
8201 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8203 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8207 * Workarounds for some broken servers that "hang" if a client hello
8208 record length exceeds 255 bytes.
8210 1. Do not use record version number > TLS 1.0 in initial client
8211 hello: some (but not all) hanging servers will now work.
8212 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8213 the number of ciphers sent in the client hello. This should be
8214 set to an even number, such as 50, for example by passing:
8215 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8216 Most broken servers should now work.
8217 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8218 TLS 1.2 client support entirely.
8222 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8226 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8228 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8229 STRING form instead of a DigestInfo.
8233 * The format used for MDC2 RSA signatures is inconsistent between EVP
8234 and the RSA_sign/RSA_verify functions. This was made more apparent when
8235 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8236 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8237 the correct format in RSA_verify so both forms transparently work.
8241 * Some servers which support TLS 1.0 can choke if we initially indicate
8242 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8243 encrypted premaster secret. As a workaround use the maximum permitted
8244 client version in client hello, this should keep such servers happy
8245 and still work with previous versions of OpenSSL.
8249 * Add support for TLS/DTLS heartbeats.
8251 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8253 * Add support for SCTP.
8255 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8257 * Improved PRNG seeding for VOS.
8259 *Paul Green <Paul.Green@stratus.com>*
8261 * Extensive assembler packs updates, most notably:
8263 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8264 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8265 - x86_64: bit-sliced AES implementation;
8266 - ARM: NEON support, contemporary platforms optimizations;
8267 - s390x: z196 support;
8268 - `*`: GHASH and GF(2^m) multiplication implementations;
8272 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8273 (removal of unnecessary code)
8275 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8277 * Add TLS key material exporter from RFC 5705.
8281 * Add DTLS-SRTP negotiation from RFC 5764.
8285 * Add Next Protocol Negotiation,
8286 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8287 disabled with a no-npn flag to config or Configure. Code donated
8290 *Adam Langley <agl@google.com> and Ben Laurie*
8292 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8293 NIST-P256, NIST-P521, with constant-time single point multiplication on
8294 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8295 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8296 Code made available under Apache License version 2.0.
8298 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8299 line to include this in your build of OpenSSL, and run "make depend" (or
8300 "make update"). This enables the following EC_METHODs:
8302 EC_GFp_nistp224_method()
8303 EC_GFp_nistp256_method()
8304 EC_GFp_nistp521_method()
8306 EC_GROUP_new_by_curve_name() will automatically use these (while
8307 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8310 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8312 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8313 all platforms. Move ssize_t definition from e_os.h to the public
8314 header file e_os2.h as it now appears in public header file cms.h
8318 * New -sigopt option to the ca, req and x509 utilities. Additional
8319 signature parameters can be passed using this option and in
8324 * Add RSA PSS signing function. This will generate and set the
8325 appropriate AlgorithmIdentifiers for PSS based on those in the
8326 corresponding EVP_MD_CTX structure. No application support yet.
8330 * Support for companion algorithm specific ASN1 signing routines.
8331 New function ASN1_item_sign_ctx() signs a pre-initialised
8332 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8333 the appropriate parameters.
8337 * Add new algorithm specific ASN1 verification initialisation function
8338 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8339 handling will be the same no matter what EVP_PKEY_METHOD is used.
8340 Add a PSS handler to support verification of PSS signatures: checked
8341 against a number of sample certificates.
8345 * Add signature printing for PSS. Add PSS OIDs.
8347 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8349 * Add algorithm specific signature printing. An individual ASN1 method
8350 can now print out signatures instead of the standard hex dump.
8352 More complex signatures (e.g. PSS) can print out more meaningful
8353 information. Include DSA version that prints out the signature
8358 * Password based recipient info support for CMS library: implementing
8363 * Split password based encryption into PBES2 and PBKDF2 functions. This
8364 neatly separates the code into cipher and PBE sections and is required
8365 for some algorithms that split PBES2 into separate pieces (such as
8366 password based CMS).
8370 * Session-handling fixes:
8371 - Fix handling of connections that are resuming with a session ID,
8372 but also support Session Tickets.
8373 - Fix a bug that suppressed issuing of a new ticket if the client
8374 presented a ticket with an expired session.
8375 - Try to set the ticket lifetime hint to something reasonable.
8376 - Make tickets shorter by excluding irrelevant information.
8377 - On the client side, don't ignore renewed tickets.
8379 *Adam Langley, Bodo Moeller (Google)*
8381 * Fix PSK session representation.
8385 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8387 This work was sponsored by Intel.
8391 * Add GCM support to TLS library. Some custom code is needed to split
8392 the IV between the fixed (from PRF) and explicit (from TLS record)
8393 portions. This adds all GCM ciphersuites supported by RFC5288 and
8394 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8395 add a special AESGCM string for GCM only.
8399 * Expand range of ctrls for AES GCM. Permit setting invocation
8400 field on decrypt and retrieval of invocation field only on encrypt.
8404 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8405 As required by RFC5289 these ciphersuites cannot be used if for
8406 versions of TLS earlier than 1.2.
8410 * For FIPS capable OpenSSL interpret a NULL default public key method
8411 as unset and return the appropriate default but do *not* set the default.
8412 This means we can return the appropriate method in applications that
8413 switch between FIPS and non-FIPS modes.
8417 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8418 ENGINE is used then we cannot handle that in the FIPS module so we
8419 keep original code iff non-FIPS operations are allowed.
8423 * Add -attime option to openssl utilities.
8425 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8427 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8431 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8432 FIPS EC methods unconditionally for now.
8436 * New build option no-ec2m to disable characteristic 2 code.
8440 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8441 all cases can be covered as some introduce binary incompatibilities.
8445 * Redirect RSA operations to FIPS module including keygen,
8446 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8450 * Add similar low-level API blocking to ciphers.
8454 * low-level digest APIs are not approved in FIPS mode: any attempt
8455 to use these will cause a fatal error. Applications that *really* want
8456 to use them can use the `private_*` version instead.
8460 * Redirect cipher operations to FIPS module for FIPS builds.
8464 * Redirect digest operations to FIPS module for FIPS builds.
8468 * Update build system to add "fips" flag which will link in fipscanister.o
8469 for static and shared library builds embedding a signature if needed.
8473 * Output TLS supported curves in preference order instead of numerical
8474 order. This is currently hardcoded for the highest order curves first.
8475 This should be configurable so applications can judge speed vs strength.
8479 * Add TLS v1.2 server support for client authentication.
8483 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8488 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8489 FIPS modules versions.
8493 * Add TLS v1.2 client side support for client authentication. Keep cache
8494 of handshake records longer as we don't know the hash algorithm to use
8495 until after the certificate request message is received.
8499 * Initial TLS v1.2 client support. Add a default signature algorithms
8500 extension including all the algorithms we support. Parse new signature
8501 format in client key exchange. Relax some ECC signing restrictions for
8502 TLS v1.2 as indicated in RFC5246.
8506 * Add server support for TLS v1.2 signature algorithms extension. Switch
8507 to new signature format when needed using client digest preference.
8508 All server ciphersuites should now work correctly in TLS v1.2. No client
8509 support yet and no support for client certificates.
8513 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8514 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8515 ciphersuites. At present only RSA key exchange ciphersuites work with
8516 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8517 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8518 and version checking.
8522 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8523 with this defined it will not be affected by any changes to ssl internal
8524 structures. Add several utility functions to allow openssl application
8525 to work with OPENSSL_NO_SSL_INTERN defined.
8529 * A long standing patch to add support for SRP from EdelWeb (Peter
8530 Sylvester and Christophe Renou) was integrated.
8531 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8532 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8535 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8539 * Permit abbreviated handshakes when renegotiating using the function
8540 SSL_renegotiate_abbreviated().
8542 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8544 * Add call to ENGINE_register_all_complete() to
8545 ENGINE_load_builtin_engines(), so some implementations get used
8546 automatically instead of needing explicit application support.
8550 * Add support for TLS key exporter as described in RFC5705.
8552 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8554 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8555 a few changes are required:
8557 Add SSL_OP_NO_TLSv1_1 flag.
8558 Add TLSv1_1 methods.
8559 Update version checking logic to handle version 1.1.
8560 Add explicit IV handling (ported from DTLS code).
8561 Add command line options to s_client/s_server.
8568 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8570 * X509_ATTRIBUTE memory leak
8572 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8573 memory. This structure is used by the PKCS#7 and CMS routines so any
8574 application which reads PKCS#7 or CMS data from untrusted sources is
8575 affected. SSL/TLS is not affected.
8577 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8583 * Race condition handling PSK identify hint
8585 If PSK identity hints are received by a multi-threaded client then
8586 the values are wrongly updated in the parent SSL_CTX structure. This can
8587 result in a race condition potentially leading to a double free of the
8593 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8595 * Malformed ECParameters causes infinite loop
8597 When processing an ECParameters structure OpenSSL enters an infinite loop
8598 if the curve specified is over a specially malformed binary polynomial
8601 This can be used to perform denial of service against any
8602 system which processes public keys, certificate requests or
8603 certificates. This includes TLS clients and TLS servers with
8604 client authentication enabled.
8606 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8611 * Exploitable out-of-bounds read in X509_cmp_time
8613 X509_cmp_time does not properly check the length of the ASN1_TIME
8614 string and can read a few bytes out of bounds. In addition,
8615 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8618 An attacker can use this to craft malformed certificates and CRLs of
8619 various sizes and potentially cause a segmentation fault, resulting in
8620 a DoS on applications that verify certificates or CRLs. TLS clients
8621 that verify CRLs are affected. TLS clients and servers with client
8622 authentication enabled may be affected if they use custom verification
8625 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8626 independently by Hanno Böck.
8631 * PKCS7 crash with missing EnvelopedContent
8633 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8634 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8635 with missing content and trigger a NULL pointer dereference on parsing.
8637 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8638 structures from untrusted sources are affected. OpenSSL clients and
8639 servers are not affected.
8641 This issue was reported to OpenSSL by Michal Zalewski (Google).
8646 * CMS verify infinite loop with unknown hash function
8648 When verifying a signedData message the CMS code can enter an infinite loop
8649 if presented with an unknown hash function OID. This can be used to perform
8650 denial of service against any system which verifies signedData messages using
8652 This issue was reported to OpenSSL by Johannes Bauer.
8657 * Race condition handling NewSessionTicket
8659 If a NewSessionTicket is received by a multi-threaded client when attempting to
8660 reuse a previous ticket then a race condition can occur potentially leading to
8661 a double free of the ticket data.
8666 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8668 * Segmentation fault in ASN1_TYPE_cmp fix
8670 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8671 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8672 certificate signature algorithm consistency this can be used to crash any
8673 certificate verification operation and exploited in a DoS attack. Any
8674 application which performs certificate verification is vulnerable including
8675 OpenSSL clients and servers which enable client authentication.
8680 * ASN.1 structure reuse memory corruption fix
8682 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8683 memory corruption via an invalid write. Such reuse is and has been
8684 strongly discouraged and is believed to be rare.
8686 Applications that parse structures containing CHOICE or ANY DEFINED BY
8687 components may be affected. Certificate parsing (d2i_X509 and related
8688 functions) are however not affected. OpenSSL clients and servers are
8694 * PKCS7 NULL pointer dereferences fix
8696 The PKCS#7 parsing code does not handle missing outer ContentInfo
8697 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8698 missing content and trigger a NULL pointer dereference on parsing.
8700 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8701 otherwise parse PKCS#7 structures from untrusted sources are
8702 affected. OpenSSL clients and servers are not affected.
8704 This issue was reported to OpenSSL by Michal Zalewski (Google).
8709 * DoS via reachable assert in SSLv2 servers fix
8711 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8712 servers that both support SSLv2 and enable export cipher suites by sending
8713 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8715 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8716 (OpenSSL development team).
8721 * Use After Free following d2i_ECPrivatekey error fix
8723 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8724 could cause a use after free condition. This, in turn, could cause a double
8725 free in several private key parsing functions (such as d2i_PrivateKey
8726 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8727 for applications that receive EC private keys from untrusted
8728 sources. This scenario is considered rare.
8730 This issue was discovered by the BoringSSL project and fixed in their
8736 * X509_to_X509_REQ NULL pointer deref fix
8738 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8739 the certificate key is invalid. This function is rarely used in practice.
8741 This issue was discovered by Brian Carpenter.
8746 * Removed the export ciphers from the DEFAULT ciphers
8750 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8752 * Build fixes for the Windows and OpenVMS platforms
8754 *Matt Caswell and Richard Levitte*
8756 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8758 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8759 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8760 dereference. This could lead to a Denial Of Service attack. Thanks to
8761 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8766 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8767 dtls1_buffer_record function under certain conditions. In particular this
8768 could occur if an attacker sent repeated DTLS records with the same
8769 sequence number but for the next epoch. The memory leak could be exploited
8770 by an attacker in a Denial of Service attack through memory exhaustion.
8771 Thanks to Chris Mueller for reporting this issue.
8776 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8777 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8778 method would be set to NULL which could later result in a NULL pointer
8779 dereference. Thanks to Frank Schmirler for reporting this issue.
8784 * Abort handshake if server key exchange message is omitted for ephemeral
8787 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8788 reporting this issue.
8793 * Remove non-export ephemeral RSA code on client and server. This code
8794 violated the TLS standard by allowing the use of temporary RSA keys in
8795 non-export ciphersuites and could be used by a server to effectively
8796 downgrade the RSA key length used to a value smaller than the server
8797 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8798 INRIA or reporting this issue.
8803 * Fixed issue where DH client certificates are accepted without verification.
8804 An OpenSSL server will accept a DH certificate for client authentication
8805 without the certificate verify message. This effectively allows a client to
8806 authenticate without the use of a private key. This only affects servers
8807 which trust a client certificate authority which issues certificates
8808 containing DH keys: these are extremely rare and hardly ever encountered.
8809 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8815 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8816 results on some platforms, including x86_64. This bug occurs at random
8817 with a very low probability, and is not known to be exploitable in any
8818 way, though its exact impact is difficult to determine. Thanks to Pieter
8819 Wuille (Blockstream) who reported this issue and also suggested an initial
8820 fix. Further analysis was conducted by the OpenSSL development team and
8821 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8822 the OpenSSL core team.
8827 * Fix various certificate fingerprint issues.
8829 By using non-DER or invalid encodings outside the signed portion of a
8830 certificate the fingerprint can be changed without breaking the signature.
8831 Although no details of the signed portion of the certificate can be changed
8832 this can cause problems with some applications: e.g. those using the
8833 certificate fingerprint for blacklists.
8835 1. Reject signatures with non zero unused bits.
8837 If the BIT STRING containing the signature has non zero unused bits reject
8838 the signature. All current signature algorithms require zero unused bits.
8840 2. Check certificate algorithm consistency.
8842 Check the AlgorithmIdentifier inside TBS matches the one in the
8843 certificate signature. NB: this will result in signature failure
8844 errors for some broken certificates.
8846 Thanks to Konrad Kraszewski from Google for reporting this issue.
8848 3. Check DSA/ECDSA signatures use DER.
8850 Re-encode DSA/ECDSA signatures and compare with the original received
8851 signature. Return an error if there is a mismatch.
8853 This will reject various cases including garbage after signature
8854 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8855 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8856 (negative or with leading zeroes).
8858 Further analysis was conducted and fixes were developed by Stephen Henson
8859 of the OpenSSL core team.
8865 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8867 * Session Ticket Memory Leak.
8869 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8870 integrity of that ticket is first verified. In the event of a session
8871 ticket integrity check failing, OpenSSL will fail to free memory
8872 causing a memory leak. By sending a large number of invalid session
8873 tickets an attacker could exploit this issue in a Denial Of Service
8879 * Build option no-ssl3 is incomplete.
8881 When OpenSSL is configured with "no-ssl3" as a build option, servers
8882 could accept and complete an SSL 3.0 handshake, and clients could be
8883 configured to send them.
8886 *Akamai and the OpenSSL team*
8888 * Add support for TLS_FALLBACK_SCSV.
8889 Client applications doing fallback retries should call
8890 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8893 *Adam Langley, Bodo Moeller*
8895 * Add additional DigestInfo checks.
8897 Re-encode DigestInto in DER and check against the original when
8898 verifying RSA signature: this will reject any improperly encoded
8899 DigestInfo structures.
8901 Note: this is a precautionary measure and no attacks are currently known.
8905 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8907 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8908 to a denial of service attack. A malicious server can crash the client
8909 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8910 ciphersuite and sending carefully crafted handshake messages.
8912 Thanks to Felix Gröbert (Google) for discovering and researching this
8918 * By sending carefully crafted DTLS packets an attacker could cause openssl
8919 to leak memory. This can be exploited through a Denial of Service attack.
8920 Thanks to Adam Langley for discovering and researching this issue.
8925 * An attacker can force openssl to consume large amounts of memory whilst
8926 processing DTLS handshake messages. This can be exploited through a
8927 Denial of Service attack.
8928 Thanks to Adam Langley for discovering and researching this issue.
8933 * An attacker can force an error condition which causes openssl to crash
8934 whilst processing DTLS packets due to memory being freed twice. This
8935 can be exploited through a Denial of Service attack.
8936 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8942 * If a multithreaded client connects to a malicious server using a resumed
8943 session and the server sends an ec point format extension it could write
8944 up to 255 bytes to freed memory.
8946 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8952 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8953 X509_name_oneline, X509_name_print_ex et al. to leak some information
8954 from the stack. Applications may be affected if they echo pretty printing
8955 output to the attacker.
8957 Thanks to Ivan Fratric (Google) for discovering this issue.
8960 *Emilia Käsper, and Steve Henson*
8962 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8963 for corner cases. (Certain input points at infinity could lead to
8964 bogus results, with non-infinity inputs mapped to infinity too.)
8968 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8970 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8971 handshake can force the use of weak keying material in OpenSSL
8972 SSL/TLS clients and servers.
8974 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8975 researching this issue. ([CVE-2014-0224])
8977 *KIKUCHI Masashi, Steve Henson*
8979 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8980 OpenSSL DTLS client the code can be made to recurse eventually crashing
8983 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8986 *Imre Rad, Steve Henson*
8988 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8989 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8990 client or server. This is potentially exploitable to run arbitrary
8991 code on a vulnerable client or server.
8993 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8995 *Jüri Aedla, Steve Henson*
8997 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8998 are subject to a denial of service attack.
9000 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
9001 this issue. ([CVE-2014-3470])
9003 *Felix Gröbert, Ivan Fratric, Steve Henson*
9005 * Harmonize version and its documentation. -f flag is used to display
9008 *mancha <mancha1@zoho.com>*
9010 * Fix eckey_priv_encode so it immediately returns an error upon a failure
9011 in i2d_ECPrivateKey.
9013 *mancha <mancha1@zoho.com>*
9015 * Fix some double frees. These are not thought to be exploitable.
9017 *mancha <mancha1@zoho.com>*
9019 * Fix for the attack described in the paper "Recovering OpenSSL
9020 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
9021 by Yuval Yarom and Naomi Benger. Details can be obtained from:
9022 <http://eprint.iacr.org/2014/140>
9024 Thanks to Yuval Yarom and Naomi Benger for discovering this
9025 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
9027 *Yuval Yarom and Naomi Benger*
9029 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
9031 * Keep original DTLS digest and encryption contexts in retransmission
9032 structures so we can use the previous session parameters if they need
9033 to be resent. ([CVE-2013-6450])
9037 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
9038 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
9039 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
9040 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
9041 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
9042 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
9044 *Rob Stradling, Adam Langley*
9046 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
9048 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
9050 This addresses the flaw in CBC record processing discovered by
9051 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
9052 at: <http://www.isg.rhul.ac.uk/tls/>
9054 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9055 Security Group at Royal Holloway, University of London
9056 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
9057 Emilia Käsper for the initial patch.
9060 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
9062 * Return an error when checking OCSP signatures when key is NULL.
9063 This fixes a DoS attack. ([CVE-2013-0166])
9067 * Call OCSP Stapling callback after ciphersuite has been chosen, so
9068 the right response is stapled. Also change SSL_get_certificate()
9069 so it returns the certificate actually sent.
9070 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
9071 (This is a backport)
9073 *Rob Stradling <rob.stradling@comodo.com>*
9075 * Fix possible deadlock when decoding public keys.
9079 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
9081 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
9084 * Sanity check record length before skipping explicit IV in DTLS
9087 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
9088 fuzzing as a service testing platform.
9093 * Initialise tkeylen properly when encrypting CMS messages.
9094 Thanks to Solar Designer of Openwall for reporting this issue.
9098 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
9100 * Check for potentially exploitable overflows in asn1_d2i_read_bio
9101 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
9102 in CRYPTO_realloc_clean.
9104 Thanks to Tavis Ormandy, Google Security Team, for discovering this
9105 issue and to Adam Langley <agl@chromium.org> for fixing it.
9108 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
9110 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
9112 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
9113 in CMS and PKCS7 code. When RSA decryption fails use a random key for
9114 content decryption and always return the same error. Note: this attack
9115 needs on average 2^20 messages so it only affects automated senders. The
9116 old behaviour can be re-enabled in the CMS code by setting the
9117 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
9118 an MMA defence is not necessary.
9119 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
9120 this issue. ([CVE-2012-0884])
9124 * Fix CVE-2011-4619: make sure we really are receiving a
9125 client hello before rejecting multiple SGC restarts. Thanks to
9126 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
9130 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
9132 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
9133 Thanks to Antonio Martin, Enterprise Secure Access Research and
9134 Development, Cisco Systems, Inc. for discovering this bug and
9135 preparing a fix. ([CVE-2012-0050])
9139 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
9141 * Nadhem Alfardan and Kenny Paterson have discovered an extension
9142 of the Vaudenay padding oracle attack on CBC mode encryption
9143 which enables an efficient plaintext recovery attack against
9144 the OpenSSL implementation of DTLS. Their attack exploits timing
9145 differences arising during decryption processing. A research
9146 paper describing this attack can be found at:
9147 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
9148 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9149 Security Group at Royal Holloway, University of London
9150 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9151 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9152 for preparing the fix. ([CVE-2011-4108])
9154 *Robin Seggelmann, Michael Tuexen*
9156 * Clear bytes used for block padding of SSL 3.0 records.
9159 *Adam Langley (Google)*
9161 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9162 Kadianakis <desnacked@gmail.com> for discovering this issue and
9163 Adam Langley for preparing the fix. ([CVE-2011-4619])
9165 *Adam Langley (Google)*
9167 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9169 *Andrey Kulikov <amdeich@gmail.com>*
9171 * Prevent malformed RFC3779 data triggering an assertion failure.
9172 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9173 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9175 *Rob Austein <sra@hactrn.net>*
9177 * Improved PRNG seeding for VOS.
9179 *Paul Green <Paul.Green@stratus.com>*
9181 * Fix ssl_ciph.c set-up race.
9183 *Adam Langley (Google)*
9185 * Fix spurious failures in ecdsatest.c.
9187 *Emilia Käsper (Google)*
9189 * Fix the BIO_f_buffer() implementation (which was mixing different
9190 interpretations of the `..._len` fields).
9192 *Adam Langley (Google)*
9194 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9195 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9196 threads won't reuse the same blinding coefficients.
9198 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9199 lock to call BN_BLINDING_invert_ex, and avoids one use of
9200 BN_BLINDING_update for each BN_BLINDING structure (previously,
9201 the last update always remained unused).
9203 *Emilia Käsper (Google)*
9205 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9207 *Bob Buckholz (Google)*
9209 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9211 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9212 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9214 *Kaspar Brand <ossl@velox.ch>*
9216 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9217 for multi-threaded use of ECDH. ([CVE-2011-3210])
9219 *Adam Langley (Google)*
9221 * Fix x509_name_ex_d2i memory leak on bad inputs.
9225 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9226 signature public key algorithm by using OID xref utilities instead.
9227 Before this you could only use some ECC ciphersuites with SHA1 only.
9231 * Add protection against ECDSA timing attacks as mentioned in the paper
9232 by Billy Bob Brumley and Nicola Tuveri, see:
9233 <http://eprint.iacr.org/2011/232.pdf>
9235 *Billy Bob Brumley and Nicola Tuveri*
9237 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9239 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9241 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9243 * Fix bug in string printing code: if *any* escaping is enabled we must
9244 escape the escape character (backslash) or the resulting string is
9249 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9251 * Disable code workaround for ancient and obsolete Netscape browsers
9252 and servers: an attacker can use it in a ciphersuite downgrade attack.
9253 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9257 * Fixed J-PAKE implementation error, originally discovered by
9258 Sebastien Martini, further info and confirmation from Stefan
9259 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9263 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9265 * Fix extension code to avoid race conditions which can result in a buffer
9266 overrun vulnerability: resumed sessions must not be modified as they can
9267 be shared by multiple threads. CVE-2010-3864
9271 * Fix WIN32 build system to correctly link an ENGINE directory into
9276 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9278 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9281 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9283 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9285 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9286 context. The operation can be customised via the ctrl mechanism in
9287 case ENGINEs want to include additional functionality.
9291 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9295 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9296 output hashes compatible with older versions of OpenSSL.
9298 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9300 * Fix compression algorithm handling: if resuming a session use the
9301 compression algorithm of the resumed session instead of determining
9302 it from client hello again. Don't allow server to change algorithm.
9306 * Add load_crls() function to commands tidying load_certs() too. Add option
9307 to verify utility to allow additional CRLs to be included.
9311 * Update OCSP request code to permit adding custom headers to the request:
9312 some responders need this.
9316 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9319 *Julia Lawall <julia@diku.dk>*
9321 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9322 needlessly dereferenced structures, used obsolete functions and
9323 didn't handle all updated verify codes correctly.
9327 * Disable MD2 in the default configuration.
9331 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9332 indicate the initial BIO being pushed or popped. This makes it possible
9333 to determine whether the BIO is the one explicitly called or as a result
9334 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9335 it handles reference counts correctly and doesn't zero out the I/O bio
9336 when it is not being explicitly popped. WARNING: applications which
9337 included workarounds for the old buggy behaviour will need to be modified
9338 or they could free up already freed BIOs.
9342 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9343 renaming to all platforms (within the 0.9.8 branch, this was
9344 done conditionally on Netware platforms to avoid a name clash).
9346 *Guenter <lists@gknw.net>*
9348 * Add ECDHE and PSK support to DTLS.
9350 *Michael Tuexen <tuexen@fh-muenster.de>*
9352 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9357 * Add "missing" function EVP_MD_flags() (without this the only way to
9358 retrieve a digest flags is by accessing the structure directly. Update
9359 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9360 or cipher is registered as in the "from" argument. Print out all
9361 registered digests in the dgst usage message instead of manually
9362 attempting to work them out.
9366 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9367 this allows the use of compression and extensions. Change default cipher
9368 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9369 by default unless an application cipher string requests it.
9373 * Alter match criteria in PKCS12_parse(). It used to try to use local
9374 key ids to find matching certificates and keys but some PKCS#12 files
9375 don't follow the (somewhat unwritten) rules and this strategy fails.
9376 Now just gather all certificates together and the first private key
9377 then look for the first certificate that matches the key.
9381 * Support use of registered digest and cipher names for dgst and cipher
9382 commands instead of having to add each one as a special case. So now
9389 openssl dgst -sha256 foo
9391 and this works for ENGINE based algorithms too.
9395 * Update Gost ENGINE to support parameter files.
9397 *Victor B. Wagner <vitus@cryptocom.ru>*
9399 * Support GeneralizedTime in ca utility.
9401 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9403 * Enhance the hash format used for certificate directory links. The new
9404 form uses the canonical encoding (meaning equivalent names will work
9405 even if they aren't identical) and uses SHA1 instead of MD5. This form
9406 is incompatible with the older format and as a result c_rehash should
9407 be used to rebuild symbolic links.
9411 * Make PKCS#8 the default write format for private keys, replacing the
9412 traditional format. This form is standardised, more secure and doesn't
9413 include an implicit MD5 dependency.
9417 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9418 committed to OpenSSL should pass this lot as a minimum.
9422 * Add session ticket override functionality for use by EAP-FAST.
9424 *Jouni Malinen <j@w1.fi>*
9426 * Modify HMAC functions to return a value. Since these can be implemented
9427 in an ENGINE errors can occur.
9431 * Type-checked OBJ_bsearch_ex.
9435 * Type-checked OBJ_bsearch. Also some constification necessitated
9436 by type-checking. Still to come: TXT_DB, bsearch(?),
9437 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9442 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9443 seconds to a tm structure directly, instead of going through OS
9444 specific date routines. This avoids any issues with OS routines such
9445 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9446 and X509_time_adj_ex() to cover the extended range. The existing
9447 X509_time_adj() is still usable and will no longer have any date issues.
9451 * Delta CRL support. New use deltas option which will attempt to locate
9452 and search any appropriate delta CRLs available.
9454 This work was sponsored by Google.
9458 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9459 code and add additional score elements. Validate alternate CRL paths
9460 as part of the CRL checking and indicate a new error "CRL path validation
9461 error" in this case. Applications wanting additional details can use
9462 the verify callback and check the new "parent" field. If this is not
9463 NULL CRL path validation is taking place. Existing applications won't
9464 see this because it requires extended CRL support which is off by
9467 This work was sponsored by Google.
9471 * Support for freshest CRL extension.
9473 This work was sponsored by Google.
9477 * Initial indirect CRL support. Currently only supported in the CRLs
9478 passed directly and not via lookup. Process certificate issuer
9479 CRL entry extension and lookup CRL entries by bother issuer name
9480 and serial number. Check and process CRL issuer entry in IDP extension.
9482 This work was sponsored by Google.
9486 * Add support for distinct certificate and CRL paths. The CRL issuer
9487 certificate is validated separately in this case. Only enabled if
9488 an extended CRL support flag is set: this flag will enable additional
9489 CRL functionality in future.
9491 This work was sponsored by Google.
9495 * Add support for policy mappings extension.
9497 This work was sponsored by Google.
9501 * Fixes to pathlength constraint, self issued certificate handling,
9502 policy processing to align with RFC3280 and PKITS tests.
9504 This work was sponsored by Google.
9508 * Support for name constraints certificate extension. DN, email, DNS
9509 and URI types are currently supported.
9511 This work was sponsored by Google.
9515 * To cater for systems that provide a pointer-based thread ID rather
9516 than numeric, deprecate the current numeric thread ID mechanism and
9517 replace it with a structure and associated callback type. This
9518 mechanism allows a numeric "hash" to be extracted from a thread ID in
9519 either case, and on platforms where pointers are larger than 'long',
9520 mixing is done to help ensure the numeric 'hash' is usable even if it
9521 can't be guaranteed unique. The default mechanism is to use "&errno"
9522 as a pointer-based thread ID to distinguish between threads.
9524 Applications that want to provide their own thread IDs should now use
9525 CRYPTO_THREADID_set_callback() to register a callback that will call
9526 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9528 Note that ERR_remove_state() is now deprecated, because it is tied
9529 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9530 to free the current thread's error state should be replaced by
9531 ERR_remove_thread_state(NULL).
9533 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9534 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9535 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9536 application was previously providing a numeric thread callback that
9537 was inappropriate for distinguishing threads, then uniqueness might
9538 have been obtained with &errno that happened immediately in the
9539 intermediate development versions of OpenSSL; this is no longer the
9540 case, the numeric thread callback will now override the automatic use
9543 *Geoff Thorpe, with help from Bodo Moeller*
9545 * Initial support for different CRL issuing certificates. This covers a
9546 simple case where the self issued certificates in the chain exist and
9547 the real CRL issuer is higher in the existing chain.
9549 This work was sponsored by Google.
9553 * Removed effectively defunct crypto/store from the build.
9557 * Revamp of STACK to provide stronger type-checking. Still to come:
9558 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9559 ASN1_STRING, CONF_VALUE.
9563 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9564 RAM on SSL connections. This option can save about 34k per idle SSL.
9568 * Revamp of LHASH to provide stronger type-checking. Still to come:
9569 STACK, TXT_DB, bsearch, qsort.
9573 * Initial support for Cryptographic Message Syntax (aka CMS) based
9574 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9575 support for data, signedData, compressedData, digestedData and
9576 encryptedData, envelopedData types included. Scripts to check against
9577 RFC4134 examples draft and interop and consistency checks of many
9578 content types and variants.
9582 * Add options to enc utility to support use of zlib compression BIO.
9586 * Extend mk1mf to support importing of options and assembly language
9587 files from Configure script, currently only included in VC-WIN32.
9588 The assembly language rules can now optionally generate the source
9589 files from the associated perl scripts.
9593 * Implement remaining functionality needed to support GOST ciphersuites.
9594 Interop testing has been performed using CryptoPro implementations.
9596 *Victor B. Wagner <vitus@cryptocom.ru>*
9598 * s390x assembler pack.
9602 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9607 * Implement Opaque PRF Input TLS extension as specified in
9608 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9609 official specification yet and no extension type assignment by
9610 IANA exists, this extension (for now) will have to be explicitly
9611 enabled when building OpenSSL by providing the extension number
9612 to use. For example, specify an option
9614 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9616 to the "config" or "Configure" script to enable the extension,
9617 assuming extension number 0x9527 (which is a completely arbitrary
9618 and unofficial assignment based on the MD5 hash of the Internet
9619 Draft). Note that by doing so, you potentially lose
9620 interoperability with other TLS implementations since these might
9621 be using the same extension number for other purposes.
9623 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9624 opaque PRF input value to use in the handshake. This will create
9625 an internal copy of the length-'len' string at 'src', and will
9626 return non-zero for success.
9628 To get more control and flexibility, provide a callback function
9631 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9632 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9636 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9639 Callback function 'cb' will be called in handshakes, and is
9640 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9641 Argument 'arg' is for application purposes (the value as given to
9642 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9643 be provided to the callback function). The callback function
9644 has to return non-zero to report success: usually 1 to use opaque
9645 PRF input just if possible, or 2 to enforce use of the opaque PRF
9646 input. In the latter case, the library will abort the handshake
9647 if opaque PRF input is not successfully negotiated.
9649 Arguments 'peerinput' and 'len' given to the callback function
9650 will always be NULL and 0 in the case of a client. A server will
9651 see the client's opaque PRF input through these variables if
9652 available (NULL and 0 otherwise). Note that if the server
9653 provides an opaque PRF input, the length must be the same as the
9654 length of the client's opaque PRF input.
9656 Note that the callback function will only be called when creating
9657 a new session (session resumption can resume whatever was
9658 previously negotiated), and will not be called in SSL 2.0
9659 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9660 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9661 for applications that need to enforce opaque PRF input.
9665 * Update ssl code to support digests other than SHA1+MD5 for handshake
9668 *Victor B. Wagner <vitus@cryptocom.ru>*
9670 * Add RFC4507 support to OpenSSL. This includes the corrections in
9671 RFC4507bis. The encrypted ticket format is an encrypted encoded
9672 SSL_SESSION structure, that way new session features are automatically
9675 If a client application caches session in an SSL_SESSION structure
9676 support is transparent because tickets are now stored in the encoded
9679 The SSL_CTX structure automatically generates keys for ticket
9680 protection in servers so again support should be possible
9681 with no application modification.
9683 If a client or server wishes to disable RFC4507 support then the option
9684 SSL_OP_NO_TICKET can be set.
9686 Add a TLS extension debugging callback to allow the contents of any client
9687 or server extensions to be examined.
9689 This work was sponsored by Google.
9693 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9694 OpenSSL should now compile cleanly on gcc 4.2
9696 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9698 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9699 support including streaming MAC support: this is required for GOST
9700 ciphersuite support.
9702 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9704 * Add option -stream to use PKCS#7 streaming in smime utility. New
9705 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9706 to output in BER and PEM format.
9710 * Experimental support for use of HMAC via EVP_PKEY interface. This
9711 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9712 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9713 ENGINE support for HMAC keys which are unextractable. New -mac and
9714 -macopt options to dgst utility.
9718 * New option -sigopt to dgst utility. Update dgst to use
9719 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9720 alternative signing parameters such as X9.31 or PSS in the dgst
9725 * Change ssl_cipher_apply_rule(), the internal function that does
9726 the work each time a ciphersuite string requests enabling
9727 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9728 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9729 the order of disabled ciphersuites such that those ciphersuites
9730 that most recently went from enabled to disabled not only stay
9731 in order with respect to each other, but also have higher priority
9732 than other disabled ciphersuites the next time ciphersuites are
9735 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9736 the same ciphersuites as with "HIGH" alone, but in a specific
9737 order where the PSK ciphersuites come first (since they are the
9738 most recently disabled ciphersuites when "HIGH" is parsed).
9740 Also, change ssl_create_cipher_list() (using this new
9741 functionality) such that between otherwise identical
9742 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9747 * Change ssl_create_cipher_list() so that it automatically
9748 arranges the ciphersuites in reasonable order before starting
9749 to process the rule string. Thus, the definition for "DEFAULT"
9750 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9751 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9752 This makes it much easier to arrive at a reasonable default order
9753 in applications for which anonymous ciphers are OK (meaning
9754 that you can't actually use DEFAULT).
9756 *Bodo Moeller; suggested by Victor Duchovni*
9758 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9759 processing) into multiple integers instead of setting
9760 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9761 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9762 (These masks as well as the individual bit definitions are hidden
9763 away into the non-exported interface ssl/ssl_locl.h, so this
9764 change to the definition of the SSL_CIPHER structure shouldn't
9765 affect applications.) This give us more bits for each of these
9766 categories, so there is no longer a need to coagulate AES128 and
9767 AES256 into a single algorithm bit, and to coagulate Camellia128
9768 and Camellia256 into a single algorithm bit, which has led to all
9771 Thus, among other things, the kludge introduced in 0.9.7m and
9772 0.9.8e for masking out AES256 independently of AES128 or masking
9773 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9775 With the change, we also introduce new ciphersuite aliases that
9776 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9781 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9782 Use the leftmost N bytes of the signature input if the input is
9783 larger than the prime q (with N being the size in bytes of q).
9787 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9788 it yet and it is largely untested.
9792 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9796 * Initial incomplete changes to avoid need for function casts in OpenSSL
9797 some compilers (gcc 4.2 and later) reject their use. Safestack is
9798 reimplemented. Update ASN1 to avoid use of legacy functions.
9802 * Win32/64 targets are linked with Winsock2.
9806 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9807 to external functions. This can be used to increase CRL handling
9808 efficiency especially when CRLs are very large by (for example) storing
9809 the CRL revoked certificates in a database.
9813 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9814 new CRLs added to a directory can be used. New command line option
9815 -verify_return_error to s_client and s_server. This causes real errors
9816 to be returned by the verify callback instead of carrying on no matter
9817 what. This reflects the way a "real world" verify callback would behave.
9821 * GOST engine, supporting several GOST algorithms and public key formats.
9822 Kindly donated by Cryptocom.
9826 * Partial support for Issuing Distribution Point CRL extension. CRLs
9827 partitioned by DP are handled but no indirect CRL or reason partitioning
9828 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9829 selected via a scoring technique which handles IDP and AKID in CRLs.
9833 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9834 will ultimately be used for all verify operations: this will remove the
9835 X509_STORE dependency on certificate verification and allow alternative
9836 lookup methods. X509_STORE based implementations of these two callbacks.
9840 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9841 Modify get_crl() to find a valid (unexpired) CRL if possible.
9845 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9846 this would be called X509_CRL_cmp() but that name is already used by
9847 a function that just compares CRL issuer names. Cache several CRL
9848 extensions in X509_CRL structure and cache CRLDP in X509.
9852 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9853 this maps equivalent X509_NAME structures into a consistent structure.
9854 Name comparison can then be performed rapidly using memcmp().
9858 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9863 * Allow digests to supply their own micalg string for S/MIME type using
9864 the ctrl EVP_MD_CTRL_MICALG.
9868 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9869 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9870 ctrl. It can then customise the structure before and/or after signing
9875 * New function OBJ_add_sigid() to allow application defined signature OIDs
9876 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9877 to free up any added signature OIDs.
9881 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9882 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9883 digest and cipher tables. New options added to openssl utility:
9884 list-message-digest-algorithms and list-cipher-algorithms.
9888 * Change the array representation of binary polynomials: the list
9889 of degrees of non-zero coefficients is now terminated with -1.
9890 Previously it was terminated with 0, which was also part of the
9891 value; thus, the array representation was not applicable to
9892 polynomials where t^0 has coefficient zero. This change makes
9893 the array representation useful in a more general context.
9897 * Various modifications and fixes to SSL/TLS cipher string
9898 handling. For ECC, the code now distinguishes between fixed ECDH
9899 with RSA certificates on the one hand and with ECDSA certificates
9900 on the other hand, since these are separate ciphersuites. The
9901 unused code for Fortezza ciphersuites has been removed.
9903 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9904 (not "ECDHE"). For consistency with the code for DH
9905 certificates, use of ECDH certificates is now considered ECDH
9906 authentication, not RSA or ECDSA authentication (the latter is
9907 merely the CA's signing algorithm and not actively used in the
9910 The temporary ciphersuite alias "ECCdraft" is no longer
9911 available, and ECC ciphersuites are no longer excluded from "ALL"
9912 and "DEFAULT". The following aliases now exist for RFC 4492
9913 ciphersuites, most of these by analogy with the DH case:
9915 kECDHr - ECDH cert, signed with RSA
9916 kECDHe - ECDH cert, signed with ECDSA
9917 kECDH - ECDH cert (signed with either RSA or ECDSA)
9918 kEECDH - ephemeral ECDH
9919 ECDH - ECDH cert or ephemeral ECDH
9925 AECDH - anonymous ECDH
9926 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9930 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9931 Use correct micalg parameters depending on digest(s) in signed message.
9935 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9936 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9940 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9941 an engine to register a method. Add ENGINE lookups for methods and
9942 functional reference processing.
9946 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9947 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9952 * New -resign option to smime utility. This adds one or more signers
9953 to an existing PKCS#7 signedData structure. Also -md option to use an
9954 alternative message digest algorithm for signing.
9958 * Tidy up PKCS#7 routines and add new functions to make it easier to
9959 create PKCS7 structures containing multiple signers. Update smime
9960 application to support multiple signers.
9964 * New -macalg option to pkcs12 utility to allow setting of an alternative
9969 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9970 Reorganize PBE internals to lookup from a static table using NIDs,
9971 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9972 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9973 PRF which will be automatically used with PBES2.
9977 * Replace the algorithm specific calls to generate keys in "req" with the
9982 * Update PKCS#7 enveloped data routines to use new API. This is now
9983 supported by any public key method supporting the encrypt operation. A
9984 ctrl is added to allow the public key algorithm to examine or modify
9985 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9990 * Add a ctrl to asn1 method to allow a public key algorithm to express
9991 a default digest type to use. In most cases this will be SHA1 but some
9992 algorithms (such as GOST) need to specify an alternative digest. The
9993 return value indicates how strong the preference is 1 means optional and
9994 2 is mandatory (that is it is the only supported type). Modify
9995 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9996 use the default md. Update openssl utilities to use the default digest
9997 type for signing if it is not explicitly indicated.
10001 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
10002 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
10003 signing method from the key type. This effectively removes the link
10004 between digests and public key types.
10008 * Add an OID cross reference table and utility functions. Its purpose is to
10009 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
10010 rsaEncryption. This will allow some of the algorithm specific hackery
10011 needed to use the correct OID to be removed.
10015 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
10016 structures for PKCS7_sign(). They are now set up by the relevant public
10021 * Add provisional EC pkey method with support for ECDSA and ECDH.
10025 * Add support for key derivation (agreement) in the API, DH method and
10030 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
10031 public and private key formats. As a side effect these add additional
10032 command line functionality not previously available: DSA signatures can be
10033 generated and verified using pkeyutl and DH key support and generation in
10040 *Oliver Tappe <zooey@hirschkaefer.de>*
10042 * New make target "install_html_docs" installs HTML renditions of the
10045 *Oliver Tappe <zooey@hirschkaefer.de>*
10047 * New utility "genpkey" this is analogous to "genrsa" etc except it can
10048 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
10049 support key and parameter generation and add initial key generation
10050 functionality for RSA.
10054 * Add functions for main EVP_PKEY_method operations. The undocumented
10055 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
10056 `EVP_PKEY_{encrypt,decrypt}_old`.
10060 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
10061 key API, doesn't do much yet.
10065 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
10066 public key algorithms. New option to openssl utility:
10067 "list-public-key-algorithms" to print out info.
10071 * Implement the Supported Elliptic Curves Extension for
10072 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10076 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
10077 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
10081 * New utilities pkey and pkeyparam. These are similar to algorithm specific
10082 utilities such as rsa, dsa, dsaparam etc except they process any key
10087 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
10088 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
10089 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
10094 * Initial support for pluggable public key ASN1.
10095 De-spaghettify the public key ASN1 handling. Move public and private
10096 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
10097 algorithm specific handling to a single module within the relevant
10098 algorithm directory. Add functions to allow (near) opaque processing
10099 of public and private key structures.
10103 * Implement the Supported Point Formats Extension for
10104 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10108 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
10109 for the psk identity [hint] and the psk callback functions to the
10110 SSL_SESSION, SSL and SSL_CTX structure.
10113 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
10117 SSL_CTX_use_psk_identity_hint
10118 SSL_get_psk_identity_hint
10119 SSL_get_psk_identity
10120 SSL_use_psk_identity_hint
10122 *Mika Kousa and Pasi Eronen of Nokia Corporation*
10124 * Add RFC 3161 compliant time stamp request creation, response generation
10125 and response verification functionality.
10127 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
10129 * Add initial support for TLS extensions, specifically for the server_name
10130 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10131 have new members for a hostname. The SSL data structure has an
10132 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10133 stored in that context to allow for session resumption, even after the
10134 SSL has been switched to a new SSL_CTX in reaction to a client's
10135 server_name extension.
10137 New functions (subject to change):
10139 SSL_get_servername()
10140 SSL_get_servername_type()
10143 New CTRL codes and macros (subject to change):
10145 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10146 - SSL_CTX_set_tlsext_servername_callback()
10147 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10148 - SSL_CTX_set_tlsext_servername_arg()
10149 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10151 openssl s_client has a new '-servername ...' option.
10153 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10154 '-key2 ...', '-servername_fatal' (subject to change). This allows
10155 testing the HostName extension for a specific single hostname ('-cert'
10156 and '-key' remain fallbacks for handshakes without HostName
10157 negotiation). If the unrecognized_name alert has to be sent, this by
10158 default is a warning; it becomes fatal with the '-servername_fatal'
10161 *Peter Sylvester, Remy Allais, Christophe Renou*
10163 * Whirlpool hash implementation is added.
10167 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10168 bn(64,32). Because of instruction set limitations it doesn't have
10169 any negative impact on performance. This was done mostly in order
10170 to make it possible to share assembler modules, such as bn_mul_mont
10171 implementations, between 32- and 64-bit builds without hassle.
10175 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10176 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10181 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10182 dedicated Montgomery multiplication procedure, is introduced.
10183 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10184 "64-bit" performance on certain 32-bit targets.
10188 * New option SSL_OP_NO_COMP to disable use of compression selectively
10189 in SSL structures. New SSL ctrl to set maximum send fragment size.
10190 Save memory by setting the I/O buffer sizes dynamically instead of
10191 using the maximum available value.
10195 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10196 in addition to the text details.
10200 * Very, very preliminary EXPERIMENTAL support for printing of general
10201 ASN1 structures. This currently produces rather ugly output and doesn't
10202 handle several customised structures at all.
10206 * Integrated support for PVK file format and some related formats such
10207 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10208 these in the 'rsa' and 'dsa' utilities.
10212 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10216 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10217 place for the (very old) "NETSCAPE" format certificates which are now
10218 handled using new ASN1 code equivalents.
10222 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10223 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10224 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10228 * Modify CRL distribution points extension code to print out previously
10229 unsupported fields. Enhance extension setting code to allow setting of
10234 * Add print and set support for Issuing Distribution Point CRL extension.
10238 * Change 'Configure' script to enable Camellia by default.
10245 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10247 * When rejecting SSL/TLS records due to an incorrect version number, never
10248 update s->server with a new major version number. As of
10249 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10250 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10251 the previous behavior could result in a read attempt at NULL when
10252 receiving specific incorrect SSL/TLS records once record payload
10253 protection is active. ([CVE-2010-0740])
10255 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10257 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10258 could be crashed if the relevant tables were not present (e.g. chrooted).
10260 *Tomas Hoger <thoger@redhat.com>*
10262 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10264 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10266 *Martin Olsson, Neel Mehta*
10268 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10269 accommodate for stack sorting, always a write lock!).
10273 * On some versions of WIN32 Heap32Next is very slow. This can cause
10274 excessive delays in the RAND_poll(): over a minute. As a workaround
10275 include a time check in the inner Heap32Next loop too.
10279 * The code that handled flushing of data in SSL/TLS originally used the
10280 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10281 the problem outlined in PR#1949. The fix suggested there however can
10282 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10283 of Apache). So instead simplify the code to flush unconditionally.
10284 This should be fine since flushing with no data to flush is a no op.
10288 * Handle TLS versions 2.0 and later properly and correctly use the
10289 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10290 off ancient servers have a habit of sticking around for a while...
10294 * Modify compression code so it frees up structures without using the
10295 ex_data callbacks. This works around a problem where some applications
10296 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10297 restarting) then use compression (e.g. SSL with compression) later.
10298 This results in significant per-connection memory leaks and
10299 has caused some security issues including CVE-2008-1678 and
10304 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10305 change when encrypting or decrypting.
10309 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10310 connect and renegotiate with servers which do not support RI.
10311 Until RI is more widely deployed this option is enabled by default.
10315 * Add "missing" ssl ctrls to clear options and mode.
10319 * If client attempts to renegotiate and doesn't support RI respond with
10320 a no_renegotiation alert as required by RFC5746. Some renegotiating
10321 TLS clients will continue a connection gracefully when they receive
10322 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10323 waiting for a server hello which it will never receive. Now we treat a
10324 received no_renegotiation alert as a fatal error. This is because
10325 applications requesting a renegotiation might well expect it to succeed
10326 and would have no code in place to handle the server denying it so the
10327 only safe thing to do is to terminate the connection.
10331 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10332 peer supports secure renegotiation and 0 otherwise. Print out peer
10333 renegotiation support in s_client/s_server.
10337 * Replace the highly broken and deprecated SPKAC certification method with
10338 the updated NID creation version. This should correctly handle UTF8.
10342 * Implement RFC5746. Re-enable renegotiation but require the extension
10343 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10344 turns out to be a bad idea. It has been replaced by
10345 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10346 SSL_CTX_set_options(). This is really not recommended unless you
10347 know what you are doing.
10349 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10351 * Fixes to stateless session resumption handling. Use initial_ctx when
10352 issuing and attempting to decrypt tickets in case it has changed during
10353 servername handling. Use a non-zero length session ID when attempting
10354 stateless session resumption: this makes it possible to determine if
10355 a resumption has occurred immediately after receiving server hello
10356 (several places in OpenSSL subtly assume this) instead of later in
10361 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10362 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10363 fixes for a few places where the return code is not checked
10366 *Julia Lawall <julia@diku.dk>*
10368 * Add --strict-warnings option to Configure script to include devteam
10369 warnings in other configurations.
10373 * Add support for --libdir option and LIBDIR variable in makefiles. This
10374 makes it possible to install openssl libraries in locations which
10375 have names other than "lib", for example "/usr/lib64" which some
10378 *Steve Henson, based on patch from Jeremy Utley*
10380 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10381 X690 8.9.12 and can produce some misleading textual output of OIDs.
10383 *Steve Henson, reported by Dan Kaminsky*
10385 * Delete MD2 from algorithm tables. This follows the recommendation in
10386 several standards that it is not used in new applications due to
10387 several cryptographic weaknesses. For binary compatibility reasons
10388 the MD2 API is still compiled in by default.
10392 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10397 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10398 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10401 *Guenter <lists@gknw.net>*
10403 * Fix the server certificate chain building code to use X509_verify_cert(),
10404 it used to have an ad-hoc builder which was unable to cope with anything
10405 other than a simple chain.
10407 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10409 * Don't check self signed certificate signatures in X509_verify_cert()
10410 by default (a flag can override this): it just wastes time without
10411 adding any security. As a useful side effect self signed root CAs
10412 with non-FIPS digests are now usable in FIPS mode.
10416 * In dtls1_process_out_of_seq_message() the check if the current message
10417 is already buffered was missing. For every new message was memory
10418 allocated, allowing an attacker to perform an denial of service attack
10419 with sending out of seq handshake messages until there is no memory
10420 left. Additionally every future message was buffered, even if the
10421 sequence number made no sense and would be part of another handshake.
10422 So only messages with sequence numbers less than 10 in advance will be
10423 buffered. ([CVE-2009-1378])
10425 *Robin Seggelmann, discovered by Daniel Mentz*
10427 * Records are buffered if they arrive with a future epoch to be
10428 processed after finishing the corresponding handshake. There is
10429 currently no limitation to this buffer allowing an attacker to perform
10430 a DOS attack with sending records with future epochs until there is no
10431 memory left. This patch adds the pqueue_size() function to determine
10432 the size of a buffer and limits the record buffer to 100 entries.
10435 *Robin Seggelmann, discovered by Daniel Mentz*
10437 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10438 parent structure is freed. ([CVE-2009-1379])
10442 * Handle non-blocking I/O properly in SSL_shutdown() call.
10444 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10446 * Add `2.5.4.*` OIDs
10448 *Ilya O. <vrghost@gmail.com>*
10450 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10452 * Disable renegotiation completely - this fixes a severe security
10453 problem ([CVE-2009-3555]) at the cost of breaking all
10454 renegotiation. Renegotiation can be re-enabled by setting
10455 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10456 run-time. This is really not recommended unless you know what
10461 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10463 * Don't set val to NULL when freeing up structures, it is freed up by
10464 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10465 zeroing past the valid field. ([CVE-2009-0789])
10467 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10469 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10470 checked correctly. This would allow some invalid signed attributes to
10471 appear to verify correctly. ([CVE-2009-0591])
10473 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10475 * Reject UniversalString and BMPString types with invalid lengths. This
10476 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10477 a legal length. ([CVE-2009-0590])
10481 * Set S/MIME signing as the default purpose rather than setting it
10482 unconditionally. This allows applications to override it at the store
10487 * Permit restricted recursion of ASN1 strings. This is needed in practice
10488 to handle some structures.
10492 * Improve efficiency of mem_gets: don't search whole buffer each time
10495 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10497 * New -hex option for openssl rand.
10501 * Print out UTF8String and NumericString when parsing ASN1.
10505 * Support NumericString type for name components.
10509 * Allow CC in the environment to override the automatically chosen
10510 compiler. Note that nothing is done to ensure flags work with the
10515 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10517 * Properly check EVP_VerifyFinal() and similar return values
10520 *Ben Laurie, Bodo Moeller, Google Security Team*
10522 * Enable TLS extensions by default.
10526 * Allow the CHIL engine to be loaded, whether the application is
10527 multithreaded or not. (This does not release the developer from the
10528 obligation to set up the dynamic locking callbacks.)
10530 *Sander Temme <sander@temme.net>*
10532 * Use correct exit code if there is an error in dgst command.
10534 *Steve Henson; problem pointed out by Roland Dirlewanger*
10536 * Tweak Configure so that you need to say "experimental-jpake" to enable
10537 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10541 * Add experimental JPAKE support, including demo authentication in
10542 s_client and s_server.
10546 * Set the comparison function in v3_addr_canonize().
10548 *Rob Austein <sra@hactrn.net>*
10550 * Add support for XMPP STARTTLS in s_client.
10552 *Philip Paeps <philip@freebsd.org>*
10554 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10555 to ensure that even with this option, only ciphersuites in the
10556 server's preference list will be accepted. (Note that the option
10557 applies only when resuming a session, so the earlier behavior was
10558 just about the algorithm choice for symmetric cryptography.)
10562 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10564 * Fix NULL pointer dereference if a DTLS server received
10565 ChangeCipherSpec as first record ([CVE-2009-1386]).
10569 * Fix a state transition in s3_srvr.c and d1_srvr.c
10570 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10572 *Nagendra Modadugu*
10574 * The fix in 0.9.8c that supposedly got rid of unsafe
10575 double-checked locking was incomplete for RSA blinding,
10576 addressing just one layer of what turns out to have been
10577 doubly unsafe triple-checked locking.
10579 So now fix this for real by retiring the MONT_HELPER macro
10580 in crypto/rsa/rsa_eay.c.
10582 *Bodo Moeller; problem pointed out by Marius Schilder*
10584 * Various precautionary measures:
10586 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10588 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10589 (NB: This would require knowledge of the secret session ticket key
10590 to exploit, in which case you'd be SOL either way.)
10592 - Change bn_nist.c so that it will properly handle input BIGNUMs
10593 outside the expected range.
10595 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10598 *Neel Mehta, Bodo Moeller*
10600 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10601 the load fails. Useful for distros.
10603 *Ben Laurie and the FreeBSD team*
10605 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10609 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10613 * Expand ENGINE to support engine supplied SSL client certificate functions.
10615 This work was sponsored by Logica.
10619 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10620 keystores. Support for SSL/TLS client authentication too.
10621 Not compiled unless enable-capieng specified to Configure.
10623 This work was sponsored by Logica.
10627 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10628 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10629 attribute creation routines such as certificate requests and PKCS#12
10634 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10636 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10637 handshake which could lead to a client crash as found using the
10638 Codenomicon TLS test suite ([CVE-2008-1672])
10640 *Steve Henson, Mark Cox*
10642 * Fix double free in TLS server name extensions which could lead to
10643 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10647 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10649 Clear the error queue to ensure that error entries left from
10650 older function calls do not interfere with the correct operation.
10652 *Lutz Jaenicke, Erik de Castro Lopo*
10654 * Remove root CA certificates of commercial CAs:
10656 The OpenSSL project does not recommend any specific CA and does not
10657 have any policy with respect to including or excluding any CA.
10658 Therefore, it does not make any sense to ship an arbitrary selection
10659 of root CA certificates with the OpenSSL software.
10663 * RSA OAEP patches to fix two separate invalid memory reads.
10664 The first one involves inputs when 'lzero' is greater than
10665 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10666 before the beginning of from). The second one involves inputs where
10667 the 'db' section contains nothing but zeroes (there is a one-byte
10668 invalid read after the end of 'db').
10670 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10672 * Partial backport from 0.9.9-dev:
10674 Introduce bn_mul_mont (dedicated Montgomery multiplication
10675 procedure) as a candidate for BIGNUM assembler implementation.
10676 While 0.9.9-dev uses assembler for various architectures, only
10677 x86_64 is available by default here in the 0.9.8 branch, and
10678 32-bit x86 is available through a compile-time setting.
10680 To try the 32-bit x86 assembler implementation, use Configure
10681 option "enable-montasm" (which exists only for this backport).
10683 As "enable-montasm" for 32-bit x86 disclaims code stability
10684 anyway, in this constellation we activate additional code
10685 backported from 0.9.9-dev for further performance improvements,
10686 namely BN_from_montgomery_word. (To enable this otherwise,
10687 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10689 *Andy Polyakov (backport partially by Bodo Moeller)*
10691 * Add TLS session ticket callback. This allows an application to set
10692 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10693 values. This is useful for key rollover for example where several key
10694 sets may exist with different names.
10698 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10699 This was broken until now in 0.9.8 releases, such that the only way
10700 a registered ENGINE could be used (assuming it initialises
10701 successfully on the host) was to explicitly set it as the default
10702 for the relevant algorithms. This is in contradiction with 0.9.7
10703 behaviour and the documentation. With this fix, when an ENGINE is
10704 registered into a given algorithm's table of implementations, the
10705 'uptodate' flag is reset so that auto-discovery will be used next
10706 time a new context for that algorithm attempts to select an
10709 *Ian Lister (tweaked by Geoff Thorpe)*
10711 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10712 implementation in the following ways:
10714 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10717 Lack of BER streaming support means one pass streaming processing is
10718 only supported if data is detached: setting the streaming flag is
10719 ignored for embedded content.
10721 CMS support is disabled by default and must be explicitly enabled
10722 with the enable-cms configuration option.
10726 * Update the GMP engine glue to do direct copies between BIGNUM and
10727 mpz_t when openssl and GMP use the same limb size. Otherwise the
10728 existing "conversion via a text string export" trick is still used.
10730 *Paul Sheer <paulsheer@gmail.com>*
10732 * Zlib compression BIO. This is a filter BIO which compressed and
10733 uncompresses any data passed through it.
10737 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10738 RFC3394 compatible AES key wrapping.
10742 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10743 sets string data without copying. X509_ALGOR_set0() and
10744 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10745 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10746 from an X509_ATTRIBUTE structure optionally checking it occurs only
10747 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10752 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10753 to get the expected BN_FLG_CONSTTIME behavior.
10755 *Bodo Moeller (Google)*
10759 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10760 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10761 - added some more tests to do_tests.pl
10762 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10763 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10764 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10765 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10766 - various changes to netware.pl to enable gcc-cross builds on Win32
10768 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10769 - various changes to fix missing prototype warnings
10770 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10771 - added AES, WHIRLPOOL and CPUID assembler code to build files
10772 - added missing AES assembler make rules to mk1mf.pl
10773 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10775 *Guenter Knauf <eflash@gmx.net>*
10777 * Implement certificate status request TLS extension defined in RFC3546.
10778 A client can set the appropriate parameters and receive the encoded
10779 OCSP response via a callback. A server can query the supplied parameters
10780 and set the encoded OCSP response in the callback. Add simplified examples
10781 to s_client and s_server.
10785 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10787 * Fix various bugs:
10788 + Binary incompatibility of ssl_ctx_st structure
10789 + DTLS interoperation with non-compliant servers
10790 + Don't call get_session_cb() without proposed session
10791 + Fix ia64 assembler code
10793 *Andy Polyakov, Steve Henson*
10795 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10797 * DTLS Handshake overhaul. There were longstanding issues with
10798 OpenSSL DTLS implementation, which were making it impossible for
10799 RFC 4347 compliant client to communicate with OpenSSL server.
10800 Unfortunately just fixing these incompatibilities would "cut off"
10801 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10802 server keeps tolerating non RFC compliant syntax. The opposite is
10803 not true, 0.9.8f client can not communicate with earlier server.
10804 This update even addresses CVE-2007-4995.
10808 * Changes to avoid need for function casts in OpenSSL: some compilers
10809 (gcc 4.2 and later) reject their use.
10810 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10813 * Add RFC4507 support to OpenSSL. This includes the corrections in
10814 RFC4507bis. The encrypted ticket format is an encrypted encoded
10815 SSL_SESSION structure, that way new session features are automatically
10818 If a client application caches session in an SSL_SESSION structure
10819 support is transparent because tickets are now stored in the encoded
10822 The SSL_CTX structure automatically generates keys for ticket
10823 protection in servers so again support should be possible
10824 with no application modification.
10826 If a client or server wishes to disable RFC4507 support then the option
10827 SSL_OP_NO_TICKET can be set.
10829 Add a TLS extension debugging callback to allow the contents of any client
10830 or server extensions to be examined.
10832 This work was sponsored by Google.
10836 * Add initial support for TLS extensions, specifically for the server_name
10837 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10838 have new members for a hostname. The SSL data structure has an
10839 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10840 stored in that context to allow for session resumption, even after the
10841 SSL has been switched to a new SSL_CTX in reaction to a client's
10842 server_name extension.
10844 New functions (subject to change):
10846 SSL_get_servername()
10847 SSL_get_servername_type()
10850 New CTRL codes and macros (subject to change):
10852 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10853 - SSL_CTX_set_tlsext_servername_callback()
10854 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10855 - SSL_CTX_set_tlsext_servername_arg()
10856 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10858 openssl s_client has a new '-servername ...' option.
10860 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10861 '-key2 ...', '-servername_fatal' (subject to change). This allows
10862 testing the HostName extension for a specific single hostname ('-cert'
10863 and '-key' remain fallbacks for handshakes without HostName
10864 negotiation). If the unrecognized_name alert has to be sent, this by
10865 default is a warning; it becomes fatal with the '-servername_fatal'
10868 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10870 * Add AES and SSE2 assembly language support to VC++ build.
10874 * Mitigate attack on final subtraction in Montgomery reduction.
10878 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10879 (which previously caused an internal error).
10883 * Squeeze another 10% out of IGE mode when in != out.
10887 * AES IGE mode speedup.
10889 *Dean Gaudet (Google)*
10891 * Add the Korean symmetric 128-bit cipher SEED (see
10892 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10893 add SEED ciphersuites from RFC 4162:
10895 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10896 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10897 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10898 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10900 To minimize changes between patchlevels in the OpenSSL 0.9.8
10901 series, SEED remains excluded from compilation unless OpenSSL
10902 is configured with 'enable-seed'.
10904 *KISA, Bodo Moeller*
10906 * Mitigate branch prediction attacks, which can be practical if a
10907 single processor is shared, allowing a spy process to extract
10908 information. For detailed background information, see
10909 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10910 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10911 and Necessary Software Countermeasures"). The core of the change
10912 are new versions BN_div_no_branch() and
10913 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10914 respectively, which are slower, but avoid the security-relevant
10915 conditional branches. These are automatically called by BN_div()
10916 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10917 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10918 remove a conditional branch.
10920 BN_FLG_CONSTTIME is the new name for the previous
10921 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10922 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10923 in the exponent causes BN_mod_exp_mont() to use the alternative
10924 implementation in BN_mod_exp_mont_consttime().) The old name
10925 remains as a deprecated alias.
10927 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10928 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10929 constant-time implementations for more than just exponentiation.
10930 Here too the old name is kept as a deprecated alias.
10932 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10933 the BN_BLINDING structure gets an independent copy of the
10934 modulus. This means that the previous `BIGNUM *m` argument to
10935 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10936 essentially becomes `const BIGNUM *m`, although we can't actually
10937 change this in the header file before 0.9.9. It allows
10938 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10939 enable BN_FLG_CONSTTIME.
10941 *Matthew D Wood (Intel Corp)*
10943 * In the SSL/TLS server implementation, be strict about session ID
10944 context matching (which matters if an application uses a single
10945 external cache for different purposes). Previously,
10946 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10947 set. This did ensure strict client verification, but meant that,
10948 with applications using a single external cache for quite
10949 different requirements, clients could circumvent ciphersuite
10950 restrictions for a given session ID context by starting a session
10951 in a different context.
10955 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10956 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10957 authentication-only ciphersuites.
10961 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10962 not complete and could lead to a possible single byte overflow
10963 ([CVE-2007-5135]) [Ben Laurie]
10965 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10967 * Since AES128 and AES256 (and similarly Camellia128 and
10968 Camellia256) share a single mask bit in the logic of
10969 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10970 kludge to work properly if AES128 is available and AES256 isn't
10971 (or if Camellia128 is available and Camellia256 isn't).
10975 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10976 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10977 When a point or a seed is encoded in a BIT STRING, we need to
10978 prevent the removal of trailing zero bits to get the proper DER
10979 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10980 of a NamedBitList, for which trailing 0 bits need to be removed.)
10984 * Have SSL/TLS server implementation tolerate "mismatched" record
10985 protocol version while receiving ClientHello even if the
10986 ClientHello is fragmented. (The server can't insist on the
10987 particular protocol version it has chosen before the ServerHello
10988 message has informed the client about his choice.)
10992 * Add RFC 3779 support.
10994 *Rob Austein for ARIN, Ben Laurie*
10996 * Load error codes if they are not already present instead of using a
10997 static variable. This allows them to be cleanly unloaded and reloaded.
10998 Improve header file function name parsing.
11002 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
11003 or CAPABILITY handshake as required by RFCs.
11005 *Goetz Babin-Ebell*
11007 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
11009 * Introduce limits to prevent malicious keys being able to
11010 cause a denial of service. ([CVE-2006-2940])
11012 *Steve Henson, Bodo Moeller*
11014 * Fix ASN.1 parsing of certain invalid structures that can result
11015 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11017 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11018 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11020 * Fix SSL client code which could crash if connecting to a
11021 malicious SSLv2 server. ([CVE-2006-4343])
11023 *Tavis Ormandy and Will Drewry, Google Security Team*
11025 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
11026 match only those. Before that, "AES256-SHA" would be interpreted
11027 as a pattern and match "AES128-SHA" too (since AES128-SHA got
11028 the same strength classification in 0.9.7h) as we currently only
11029 have a single AES bit in the ciphersuite description bitmap.
11030 That change, however, also applied to ciphersuite strings such as
11031 "RC4-MD5" that intentionally matched multiple ciphersuites --
11032 namely, SSL 2.0 ciphersuites in addition to the more common ones
11033 from SSL 3.0/TLS 1.0.
11035 So we change the selection algorithm again: Naming an explicit
11036 ciphersuite selects this one ciphersuite, and any other similar
11037 ciphersuite (same bitmap) from *other* protocol versions.
11038 Thus, "RC4-MD5" again will properly select both the SSL 2.0
11039 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
11041 Since SSL 2.0 does not have any ciphersuites for which the
11042 128/256 bit distinction would be relevant, this works for now.
11043 The proper fix will be to use different bits for AES128 and
11044 AES256, which would have avoided the problems from the beginning;
11045 however, bits are scarce, so we can only do this in a new release
11046 (not just a patchlevel) when we can change the SSL_CIPHER
11047 definition to split the single 'unsigned long mask' bitmap into
11048 multiple values to extend the available space.
11052 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
11054 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11055 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11057 * Add AES IGE and biIGE modes.
11061 * Change the Unix randomness entropy gathering to use poll() when
11062 possible instead of select(), since the latter has some
11063 undesirable limitations.
11065 *Darryl Miles via Richard Levitte and Bodo Moeller*
11067 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
11068 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
11069 cannot be implicitly activated as part of, e.g., the "AES" alias.
11070 However, please upgrade to OpenSSL 0.9.9[-dev] for
11071 non-experimental use of the ECC ciphersuites to get TLS extension
11072 support, which is required for curve and point format negotiation
11073 to avoid potential handshake problems.
11077 * Disable rogue ciphersuites:
11079 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11080 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11081 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11083 The latter two were purportedly from
11084 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11087 Also deactivate the remaining ciphersuites from
11088 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11089 unofficial, and the ID has long expired.
11093 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11094 dual-core machines) and other potential thread-safety issues.
11098 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
11099 versions), which is now available for royalty-free use
11100 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
11101 Also, add Camellia TLS ciphersuites from RFC 4132.
11103 To minimize changes between patchlevels in the OpenSSL 0.9.8
11104 series, Camellia remains excluded from compilation unless OpenSSL
11105 is configured with 'enable-camellia'.
11109 * Disable the padding bug check when compression is in use. The padding
11110 bug check assumes the first packet is of even length, this is not
11111 necessarily true if compression is enabled and can result in false
11112 positives causing handshake failure. The actual bug test is ancient
11113 code so it is hoped that implementations will either have fixed it by
11114 now or any which still have the bug do not support compression.
11118 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
11120 * When applying a cipher rule check to see if string match is an explicit
11121 cipher suite and only match that one cipher suite if it is.
11125 * Link in manifests for VC++ if needed.
11127 *Austin Ziegler <halostatue@gmail.com>*
11129 * Update support for ECC-based TLS ciphersuites according to
11130 draft-ietf-tls-ecc-12.txt with proposed changes (but without
11131 TLS extensions, which are supported starting with the 0.9.9
11132 branch, not in the OpenSSL 0.9.8 branch).
11136 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
11137 opaque EVP_CIPHER_CTX handling.
11141 * Fixes and enhancements to zlib compression code. We now only use
11142 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
11143 to conform with the standards mentioned here:
11144 <http://www.zlib.net/DLL_FAQ.txt>
11145 Static zlib linking now works on Windows and the new --with-zlib-include
11146 --with-zlib-lib options to Configure can be used to supply the location
11147 of the headers and library. Gracefully handle case where zlib library
11152 * Several fixes and enhancements to the OID generation code. The old code
11153 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11154 handle numbers larger than ULONG_MAX, truncated printing and had a
11155 non standard OBJ_obj2txt() behaviour.
11159 * Add support for building of engines under engine/ as shared libraries
11160 under VC++ build system.
11164 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11165 Hopefully, we will not see any false combination of paths any more.
11169 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11171 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11172 (part of SSL_OP_ALL). This option used to disable the
11173 countermeasure against man-in-the-middle protocol-version
11174 rollback in the SSL 2.0 server implementation, which is a bad
11175 idea. ([CVE-2005-2969])
11177 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11178 for Information Security, National Institute of Advanced Industrial
11179 Science and Technology [AIST], Japan)*
11181 * Add two function to clear and return the verify parameter flags.
11185 * Keep cipherlists sorted in the source instead of sorting them at
11186 runtime, thus removing the need for a lock.
11190 * Avoid some small subgroup attacks in Diffie-Hellman.
11192 *Nick Mathewson and Ben Laurie*
11194 * Add functions for well-known primes.
11198 * Extended Windows CE support.
11200 *Satoshi Nakamura and Andy Polyakov*
11202 * Initialize SSL_METHOD structures at compile time instead of during
11203 runtime, thus removing the need for a lock.
11207 * Make PKCS7_decrypt() work even if no certificate is supplied by
11208 attempting to decrypt each encrypted key in turn. Add support to
11213 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11215 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11218 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11222 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11223 key into the same file any more.
11227 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11231 * Add -utf8 command line and config file option to 'ca'.
11233 *Stefan <stf@udoma.org*
11235 * Removed the macro des_crypt(), as it seems to conflict with some
11236 libraries. Use DES_crypt().
11240 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11241 involves renaming the source and generated shared-libs for
11242 both. The engines will accept the corrected or legacy ids
11243 ('ncipher' and '4758_cca' respectively) when binding. NB,
11244 this only applies when building 'shared'.
11246 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11248 * Add attribute functions to EVP_PKEY structure. Modify
11249 PKCS12_create() to recognize a CSP name attribute and
11250 use it. Make -CSP option work again in pkcs12 utility.
11254 * Add new functionality to the bn blinding code:
11255 - automatic re-creation of the BN_BLINDING parameters after
11256 a fixed number of uses (currently 32)
11257 - add new function for parameter creation
11258 - introduce flags to control the update behaviour of the
11259 BN_BLINDING parameters
11260 - hide BN_BLINDING structure
11261 Add a second BN_BLINDING slot to the RSA structure to improve
11262 performance when a single RSA object is shared among several
11267 * Add support for DTLS.
11269 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11271 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11272 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11276 * Remove buggy and incomplete DH cert support from
11277 ssl/ssl_rsa.c and ssl/s3_both.c
11281 * Use SHA-1 instead of MD5 as the default digest algorithm for
11282 the `apps/openssl` commands.
11286 * Compile clean with "-Wall -Wmissing-prototypes
11287 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11288 DEBUG_SAFESTACK must also be set.
11292 * Change ./Configure so that certain algorithms can be disabled by default.
11293 The new counterpiece to "no-xxx" is "enable-xxx".
11295 The patented RC5 and MDC2 algorithms will now be disabled unless
11296 "enable-rc5" and "enable-mdc2", respectively, are specified.
11298 (IDEA remains enabled despite being patented. This is because IDEA
11299 is frequently required for interoperability, and there is no license
11300 fee for non-commercial use. As before, "no-idea" can be used to
11301 avoid this algorithm.)
11305 * Add processing of proxy certificates (see RFC 3820). This work was
11306 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11307 EGEE (Enabling Grids for E-science in Europe).
11311 * RC4 performance overhaul on modern architectures/implementations, such
11312 as Intel P4, IA-64 and AMD64.
11316 * New utility extract-section.pl. This can be used specify an alternative
11317 section number in a pod file instead of having to treat each file as
11318 a separate case in Makefile. This can be done by adding two lines to the
11321 =for comment openssl_section:XXX
11323 The blank line is mandatory.
11327 * New arguments -certform, -keyform and -pass for s_client and s_server
11328 to allow alternative format key and certificate files and passphrase
11333 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11334 update associated structures and add various utility functions.
11336 Add new policy related verify parameters, include policy checking in
11337 standard verify code. Enhance 'smime' application with extra parameters
11338 to support policy checking and print out.
11342 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11343 Nehemiah processors. These extensions support AES encryption in hardware
11344 as well as RNG (though RNG support is currently disabled).
11346 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11348 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11352 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11354 *Andy Polyakov and a number of other people*
11356 * Improved PowerPC platform support. Most notably BIGNUM assembler
11357 implementation contributed by IBM.
11359 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11361 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11362 exponent rather than 'unsigned long'. There is a corresponding change to
11363 the new 'rsa_keygen' element of the RSA_METHOD structure.
11365 *Jelte Jansen, Geoff Thorpe*
11367 * Functionality for creating the initial serial number file is now
11368 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11370 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11371 number file to 1, which is bound to cause problems. To avoid
11372 the problems while respecting compatibility between different 0.9.7
11373 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11374 CA.pl for serial number initialization. With the new release 0.9.8,
11375 we can fix the problem directly in the 'ca' utility.)
11379 * Reduced header interdependencies by declaring more opaque objects in
11380 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11381 give fewer recursive includes, which could break lazy source code - so
11382 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11383 developers should define this symbol when building and using openssl to
11384 ensure they track the recommended behaviour, interfaces, [etc], but
11385 backwards-compatible behaviour prevails when this isn't defined.
11389 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11393 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11394 This will generate a random key of the appropriate length based on the
11395 cipher context. The EVP_CIPHER can provide its own random key generation
11396 routine to support keys of a specific form. This is used in the des and
11397 3des routines to generate a key of the correct parity. Update S/MIME
11398 code to use new functions and hence generate correct parity DES keys.
11399 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11400 valid (weak or incorrect parity).
11404 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11405 as looking them up. This is useful when the verified structure may contain
11406 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11407 present unless the new PKCS7_NO_CRL flag is asserted.
11411 * Extend ASN1 oid configuration module. It now additionally accepts the
11414 shortName = some long name, 1.2.3.4
11418 * Reimplemented the BN_CTX implementation. There is now no more static
11419 limitation on the number of variables it can handle nor the depth of the
11420 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11421 information can now expand as required, and rather than having a single
11422 static array of bignums, BN_CTX now uses a linked-list of such arrays
11423 allowing it to expand on demand whilst maintaining the usefulness of
11424 BN_CTX's "bundling".
11428 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11429 to allow all RSA operations to function using a single BN_CTX.
11433 * Preliminary support for certificate policy evaluation and checking. This
11434 is initially intended to pass the tests outlined in "Conformance Testing
11435 of Relying Party Client Certificate Path Processing Logic" v1.07.
11439 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11440 remained unused and not that useful. A variety of other little bignum
11441 tweaks and fixes have also been made continuing on from the audit (see
11446 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11447 associated ASN1, EVP and SSL functions and old ASN1 macros.
11451 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11452 and this should never fail. So the return value from the use of
11453 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11454 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11458 * BN_CTX_get() should return zero-valued bignums, providing the same
11459 initialised value as BN_new().
11461 *Geoff Thorpe, suggested by Ulf Möller*
11463 * Support for inhibitAnyPolicy certificate extension.
11467 * An audit of the BIGNUM code is underway, for which debugging code is
11468 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11469 is considered valid when processing BIGNUMs, and causes execution to
11470 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11471 further steps are taken to deliberately pollute unused data in BIGNUM
11472 structures to try and expose faulty code further on. For now, openssl will
11473 (in its default mode of operation) continue to tolerate the inconsistent
11474 forms that it has tolerated in the past, but authors and packagers should
11475 consider trying openssl and their own applications when compiled with
11476 these debugging symbols defined. It will help highlight potential bugs in
11477 their own code, and will improve the test coverage for OpenSSL itself. At
11478 some point, these tighter rules will become openssl's default to improve
11479 maintainability, though the assert()s and other overheads will remain only
11480 in debugging configurations. See bn.h for more details.
11482 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11484 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11485 that can only be obtained through BN_CTX_new() (which implicitly
11486 initialises it). The presence of this function only made it possible
11487 to overwrite an existing structure (and cause memory leaks).
11491 * Because of the callback-based approach for implementing LHASH as a
11492 template type, lh_insert() adds opaque objects to hash-tables and
11493 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11494 to clean up those corresponding objects before destroying the hash table
11495 (and losing the object pointers). So some over-zealous constifications in
11496 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11497 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11498 prototyped to have "const" restrictions on the object pointers they are
11499 given (and so aren't required to cast them away any more).
11503 * The tmdiff.h API was so ugly and minimal that our own timing utility
11504 (speed) prefers to use its own implementation. The two implementations
11505 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11506 its object type properly exposed (MS_TM) instead of casting to/from
11507 `char *`. This may still change yet if someone realises MS_TM and
11509 aren't necessarily the greatest nomenclatures - but this is what was used
11510 internally to the implementation so I've used that for now.
11514 * Ensure that deprecated functions do not get compiled when
11515 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11516 the self-tests were still using deprecated key-generation functions so
11517 these have been updated also.
11521 * Reorganise PKCS#7 code to separate the digest location functionality
11522 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11523 New function PKCS7_set_digest() to set the digest type for PKCS#7
11524 digestedData type. Add additional code to correctly generate the
11525 digestedData type and add support for this type in PKCS7 initialization
11530 * New function PKCS7_set0_type_other() this initializes a PKCS7
11531 structure of type "other".
11535 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11536 sure the loop does correctly stop and breaking ("division by zero")
11537 modulus operations are not performed. The (pre-generated) prime
11538 table crypto/bn/bn_prime.h was already correct, but it could not be
11539 re-generated on some platforms because of the "division by zero"
11540 situation in the script.
11542 *Ralf S. Engelschall*
11544 * Update support for ECC-based TLS ciphersuites according to
11545 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11546 SHA-1 now is only used for "small" curves (where the
11547 representation of a field element takes up to 24 bytes); for
11548 larger curves, the field element resulting from ECDH is directly
11549 used as premaster secret.
11551 *Douglas Stebila (Sun Microsystems Laboratories)*
11553 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11554 curve secp160r1 to the tests.
11556 *Douglas Stebila (Sun Microsystems Laboratories)*
11558 * Add the possibility to load symbols globally with DSO.
11560 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11562 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11563 control of the error stack.
11567 * Add support for STORE in ENGINE.
11571 * Add the STORE type. The intention is to provide a common interface
11572 to certificate and key stores, be they simple file-based stores, or
11573 HSM-type store, or LDAP stores, or...
11574 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11578 * Add a generic structure called OPENSSL_ITEM. This can be used to
11579 pass a list of arguments to any function as well as provide a way
11580 for a function to pass data back to the caller.
11584 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11585 works like BUF_strdup() but can be used to duplicate a portion of
11586 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11591 * Add the function sk_find_ex() which works like sk_find(), but will
11592 return an index to an element even if an exact match couldn't be
11593 found. The index is guaranteed to point at the element where the
11594 searched-for key would be inserted to preserve sorting order.
11598 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11599 takes an extra flags argument for optional functionality. Currently,
11600 the following flags are defined:
11602 OBJ_BSEARCH_VALUE_ON_NOMATCH
11603 This one gets OBJ_bsearch_ex() to return a pointer to the first
11604 element where the comparing function returns a negative or zero
11607 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11608 This one gets OBJ_bsearch_ex() to return a pointer to the first
11609 element where the comparing function returns zero. This is useful
11610 if there are more than one element where the comparing function
11615 * Make it possible to create self-signed certificates with 'openssl ca'
11616 in such a way that the self-signed certificate becomes part of the
11617 CA database and uses the same mechanisms for serial number generation
11618 as all other certificate signing. The new flag '-selfsign' enables
11619 this functionality. Adapt CA.sh and CA.pl.in.
11623 * Add functionality to check the public key of a certificate request
11624 against a given private. This is useful to check that a certificate
11625 request can be signed by that key (self-signing).
11629 * Make it possible to have multiple active certificates with the same
11630 subject in the CA index file. This is done only if the keyword
11631 'unique_subject' is set to 'no' in the main CA section (default
11632 if 'CA_default') of the configuration file. The value is saved
11633 with the database itself in a separate index attribute file,
11634 named like the index file with '.attr' appended to the name.
11638 * Generate multi-valued AVAs using '+' notation in config files for
11643 * Support for nameConstraints certificate extension.
11647 * Support for policyConstraints certificate extension.
11651 * Support for policyMappings certificate extension.
11655 * Make sure the default DSA_METHOD implementation only uses its
11656 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11657 and change its own handlers to be NULL so as to remove unnecessary
11658 indirection. This lets alternative implementations fallback to the
11659 default implementation more easily.
11663 * Support for directoryName in GeneralName related extensions
11668 * Make it possible to link applications using Makefile.shared.
11669 Make that possible even when linking against static libraries!
11673 * Support for single pass processing for S/MIME signing. This now
11674 means that S/MIME signing can be done from a pipe, in addition
11675 cleartext signing (multipart/signed type) is effectively streaming
11676 and the signed data does not need to be all held in memory.
11678 This is done with a new flag PKCS7_STREAM. When this flag is set
11679 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11680 is done after the data is output (and digests calculated) in
11681 SMIME_write_PKCS7().
11685 * Add full support for -rpath/-R, both in shared libraries and
11686 applications, at least on the platforms where it's known how
11691 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11692 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11693 will now compute a table of multiples of the generator that
11694 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11695 faster (notably in the case of a single point multiplication,
11696 scalar * generator).
11698 *Nils Larsch, Bodo Moeller*
11700 * IPv6 support for certificate extensions. The various extensions
11701 which use the IP:a.b.c.d can now take IPv6 addresses using the
11702 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11707 * Added an ENGINE that implements RSA by performing private key
11708 exponentiations with the GMP library. The conversions to and from
11709 GMP's mpz_t format aren't optimised nor are any montgomery forms
11710 cached, and on x86 it appears OpenSSL's own performance has caught up.
11711 However there are likely to be other architectures where GMP could
11712 provide a boost. This ENGINE is not built in by default, but it can be
11713 specified at Configure time and should be accompanied by the necessary
11714 linker additions, eg;
11715 ./config -DOPENSSL_USE_GMP -lgmp
11719 * "openssl engine" will not display ENGINE/DSO load failure errors when
11720 testing availability of engines with "-t" - the old behaviour is
11721 produced by increasing the feature's verbosity with "-tt".
11725 * ECDSA routines: under certain error conditions uninitialized BN objects
11726 could be freed. Solution: make sure initialization is performed early
11727 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11732 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11733 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11734 software implementations. For DSA and DH, parameter generation can
11735 also be overridden by providing the appropriate method callbacks.
11739 * Change the "progress" mechanism used in key-generation and
11740 primality testing to functions that take a new BN_GENCB pointer in
11741 place of callback/argument pairs. The new API functions have `_ex`
11742 postfixes and the older functions are reimplemented as wrappers for
11743 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11744 declarations of the old functions to help (graceful) attempts to
11745 migrate to the new functions. Also, the new key-generation API
11746 functions operate on a caller-supplied key-structure and return
11747 success/failure rather than returning a key or NULL - this is to
11748 help make "keygen" another member function of RSA_METHOD etc.
11750 Example for using the new callback interface:
11752 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11753 void *my_arg = ...;
11756 BN_GENCB_set(&my_cb, my_callback, my_arg);
11758 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11759 /* For the meaning of a, b in calls to my_callback(), see the
11760 * documentation of the function that calls the callback.
11761 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11762 * my_callback should return 1 if it wants BN_is_prime_ex()
11763 * to continue, or 0 to stop.
11768 * Change the ZLIB compression method to be stateful, and make it
11769 available to TLS with the number defined in
11770 draft-ietf-tls-compression-04.txt.
11774 * Add the ASN.1 structures and functions for CertificatePair, which
11775 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11777 CertificatePair ::= SEQUENCE {
11778 forward [0] Certificate OPTIONAL,
11779 reverse [1] Certificate OPTIONAL,
11780 -- at least one of the pair shall be present -- }
11782 Also implement the PEM functions to read and write certificate
11783 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11785 This needed to be defined, mostly for the sake of the LDAP
11786 attribute crossCertificatePair, but may prove useful elsewhere as
11791 * Make it possible to inhibit symlinking of shared libraries in
11792 Makefile.shared, for Cygwin's sake.
11796 * Extend the BIGNUM API by creating a function
11797 void BN_set_negative(BIGNUM *a, int neg);
11798 and a macro that behave like
11799 int BN_is_negative(const BIGNUM *a);
11801 to avoid the need to access 'a->neg' directly in applications.
11805 * Implement fast modular reduction for pseudo-Mersenne primes
11806 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11807 EC_GROUP_new_curve_GFp() will now automatically use this
11810 *Nils Larsch <nla@trustcenter.de>*
11812 * Add new lock type (CRYPTO_LOCK_BN).
11816 * Change the ENGINE framework to automatically load engines
11817 dynamically from specific directories unless they could be
11818 found to already be built in or loaded. Move all the
11819 current engines except for the cryptodev one to a new
11820 directory engines/.
11821 The engines in engines/ are built as shared libraries if
11822 the "shared" options was given to ./Configure or ./config.
11823 Otherwise, they are inserted in libcrypto.a.
11824 /usr/local/ssl/engines is the default directory for dynamic
11825 engines, but that can be overridden at configure time through
11826 the usual use of --prefix and/or --openssldir, and at run
11827 time with the environment variable OPENSSL_ENGINES.
11829 *Geoff Thorpe and Richard Levitte*
11831 * Add Makefile.shared, a helper makefile to build shared
11832 libraries. Adapt Makefile.org.
11836 * Add version info to Win32 DLLs.
11838 *Peter 'Luna' Runestig" <peter@runestig.com>*
11840 * Add new 'medium level' PKCS#12 API. Certificates and keys
11841 can be added using this API to created arbitrary PKCS#12
11842 files while avoiding the low-level API.
11844 New options to PKCS12_create(), key or cert can be NULL and
11845 will then be omitted from the output file. The encryption
11846 algorithm NIDs can be set to -1 for no encryption, the mac
11847 iteration count can be set to 0 to omit the mac.
11849 Enhance pkcs12 utility by making the -nokeys and -nocerts
11850 options work when creating a PKCS#12 file. New option -nomac
11851 to omit the mac, NONE can be set for an encryption algorithm.
11852 New code is modified to use the enhanced PKCS12_create()
11853 instead of the low-level API.
11857 * Extend ASN1 encoder to support indefinite length constructed
11858 encoding. This can output sequences tags and octet strings in
11859 this form. Modify pk7_asn1.c to support indefinite length
11860 encoding. This is experimental and needs additional code to
11861 be useful, such as an ASN1 bio and some enhanced streaming
11864 Extend template encode functionality so that tagging is passed
11865 down to the template encoder.
11869 * Let 'openssl req' fail if an argument to '-newkey' is not
11870 recognized instead of using RSA as a default.
11874 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11875 As these are not official, they are not included in "ALL";
11876 the "ECCdraft" ciphersuite group alias can be used to select them.
11878 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11880 * Add ECDH engine support.
11882 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11884 * Add ECDH in new directory crypto/ecdh/.
11886 *Douglas Stebila (Sun Microsystems Laboratories)*
11888 * Let BN_rand_range() abort with an error after 100 iterations
11889 without success (which indicates a broken PRNG).
11893 * Change BN_mod_sqrt() so that it verifies that the input value
11894 is really the square of the return value. (Previously,
11895 BN_mod_sqrt would show GIGO behaviour.)
11899 * Add named elliptic curves over binary fields from X9.62, SECG,
11900 and WAP/WTLS; add OIDs that were still missing.
11902 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11904 * Extend the EC library for elliptic curves over binary fields
11905 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11908 EC_GF2m_simple_method
11912 EC_GROUP_new_curve_GF2m
11913 EC_GROUP_set_curve_GF2m
11914 EC_GROUP_get_curve_GF2m
11915 EC_POINT_set_affine_coordinates_GF2m
11916 EC_POINT_get_affine_coordinates_GF2m
11917 EC_POINT_set_compressed_coordinates_GF2m
11919 Point compression for binary fields is disabled by default for
11920 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11923 As binary polynomials are represented as BIGNUMs, various members
11924 of the EC_GROUP and EC_POINT data structures can be shared
11925 between the implementations for prime fields and binary fields;
11926 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11927 are essentially identical to their `..._GFp` counterparts.
11928 (For simplicity, the `..._GFp` prefix has been dropped from
11929 various internal method names.)
11931 An internal 'field_div' method (similar to 'field_mul' and
11932 'field_sqr') has been added; this is used only for binary fields.
11934 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11936 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11937 through methods ('mul', 'precompute_mult').
11939 The generic implementations (now internally called 'ec_wNAF_mul'
11940 and 'ec_wNAF_precomputed_mult') remain the default if these
11941 methods are undefined.
11943 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11945 * New function EC_GROUP_get_degree, which is defined through
11946 EC_METHOD. For curves over prime fields, this returns the bit
11947 length of the modulus.
11949 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11951 * New functions EC_GROUP_dup, EC_POINT_dup.
11952 (These simply call ..._new and ..._copy).
11954 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11956 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11957 Polynomials are represented as BIGNUMs (where the sign bit is not
11958 used) in the following functions [macros]:
11961 BN_GF2m_sub [= BN_GF2m_add]
11962 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11963 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11964 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11966 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11967 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11968 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11969 BN_GF2m_cmp [= BN_ucmp]
11971 (Note that only the 'mod' functions are actually for fields GF(2^m).
11972 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11974 For some functions, an the irreducible polynomial defining a
11975 field can be given as an 'unsigned int[]' with strictly
11976 decreasing elements giving the indices of those bits that are set;
11977 i.e., p[] represents the polynomial
11978 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11980 p[0] > p[1] > ... > p[k] = 0.
11981 This applies to the following functions:
11984 BN_GF2m_mod_mul_arr
11985 BN_GF2m_mod_sqr_arr
11986 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11987 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11988 BN_GF2m_mod_exp_arr
11989 BN_GF2m_mod_sqrt_arr
11990 BN_GF2m_mod_solve_quad_arr
11994 Conversion can be performed by the following functions:
11999 bntest.c has additional tests for binary polynomial arithmetic.
12001 Two implementations for BN_GF2m_mod_div() are available.
12002 The default algorithm simply uses BN_GF2m_mod_inv() and
12003 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
12004 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
12005 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
12007 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12009 * Add new error code 'ERR_R_DISABLED' that can be used when some
12010 functionality is disabled at compile-time.
12012 *Douglas Stebila <douglas.stebila@sun.com>*
12014 * Change default behaviour of 'openssl asn1parse' so that more
12015 information is visible when viewing, e.g., a certificate:
12017 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
12018 mode the content of non-printable OCTET STRINGs is output in a
12019 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
12020 avoid the appearance of a printable string.
12022 *Nils Larsch <nla@trustcenter.de>*
12024 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
12026 EC_GROUP_set_asn1_flag()
12027 EC_GROUP_get_asn1_flag()
12028 EC_GROUP_set_point_conversion_form()
12029 EC_GROUP_get_point_conversion_form()
12030 These control ASN1 encoding details:
12031 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
12032 has been set to OPENSSL_EC_NAMED_CURVE.
12033 - Points are encoded in uncompressed form by default; options for
12034 asn1_for are as for point2oct, namely
12035 POINT_CONVERSION_COMPRESSED
12036 POINT_CONVERSION_UNCOMPRESSED
12037 POINT_CONVERSION_HYBRID
12039 Also add 'seed' and 'seed_len' members to EC_GROUP with access
12041 EC_GROUP_set_seed()
12042 EC_GROUP_get0_seed()
12043 EC_GROUP_get_seed_len()
12044 This is used only for ASN1 purposes (so far).
12046 *Nils Larsch <nla@trustcenter.de>*
12048 * Add 'field_type' member to EC_METHOD, which holds the NID
12049 of the appropriate field type OID. The new function
12050 EC_METHOD_get_field_type() returns this value.
12052 *Nils Larsch <nla@trustcenter.de>*
12055 EC_POINT_point2bn()
12056 EC_POINT_bn2point()
12057 EC_POINT_point2hex()
12058 EC_POINT_hex2point()
12059 providing useful interfaces to EC_POINT_point2oct() and
12060 EC_POINT_oct2point().
12062 *Nils Larsch <nla@trustcenter.de>*
12064 * Change internals of the EC library so that the functions
12065 EC_GROUP_set_generator()
12066 EC_GROUP_get_generator()
12067 EC_GROUP_get_order()
12068 EC_GROUP_get_cofactor()
12069 are implemented directly in crypto/ec/ec_lib.c and not dispatched
12070 to methods, which would lead to unnecessary code duplication when
12071 adding different types of curves.
12073 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
12075 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
12076 arithmetic, and such that modified wNAFs are generated
12077 (which avoid length expansion in many cases).
12081 * Add a function EC_GROUP_check_discriminant() (defined via
12082 EC_METHOD) that verifies that the curve discriminant is non-zero.
12084 Add a function EC_GROUP_check() that makes some sanity tests
12085 on a EC_GROUP, its generator and order. This includes
12086 EC_GROUP_check_discriminant().
12088 *Nils Larsch <nla@trustcenter.de>*
12090 * Add ECDSA in new directory crypto/ecdsa/.
12092 Add applications 'openssl ecparam' and 'openssl ecdsa'
12093 (these are based on 'openssl dsaparam' and 'openssl dsa').
12095 ECDSA support is also included in various other files across the
12096 library. Most notably,
12097 - 'openssl req' now has a '-newkey ecdsa:file' option;
12098 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
12099 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
12100 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
12101 them suitable for ECDSA where domain parameters must be
12102 extracted before the specific public key;
12103 - ECDSA engine support has been added.
12105 *Nils Larsch <nla@trustcenter.de>*
12107 * Include some named elliptic curves, and add OIDs from X9.62,
12108 SECG, and WAP/WTLS. Each curve can be obtained from the new
12110 EC_GROUP_new_by_curve_name(),
12111 and the list of available named curves can be obtained with
12112 EC_get_builtin_curves().
12113 Also add a 'curve_name' member to EC_GROUP objects, which can be
12115 EC_GROUP_set_curve_name()
12116 EC_GROUP_get_curve_name()
12118 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
12120 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12121 was actually never needed) and in BN_mul(). The removal in BN_mul()
12122 required a small change in bn_mul_part_recursive() and the addition
12123 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12124 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12125 bn_sub_words() and bn_add_words() except they take arrays with
12130 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
12132 * Cleanse PEM buffers before freeing them since they may contain
12135 *Benjamin Bennett <ben@psc.edu>*
12137 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
12138 a ciphersuite string such as "DEFAULT:RSA" cannot enable
12139 authentication-only ciphersuites.
12143 * Since AES128 and AES256 share a single mask bit in the logic of
12144 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
12145 kludge to work properly if AES128 is available and AES256 isn't.
12149 * Expand security boundary to match 1.1.1 module.
12153 * Remove redundant features: hash file source, editing of test vectors
12154 modify fipsld to use external fips_premain.c signature.
12158 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12159 run algorithm test programs.
12163 * Make algorithm test programs more tolerant of whitespace.
12167 * Have SSL/TLS server implementation tolerate "mismatched" record
12168 protocol version while receiving ClientHello even if the
12169 ClientHello is fragmented. (The server can't insist on the
12170 particular protocol version it has chosen before the ServerHello
12171 message has informed the client about his choice.)
12175 * Load error codes if they are not already present instead of using a
12176 static variable. This allows them to be cleanly unloaded and reloaded.
12180 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12182 * Introduce limits to prevent malicious keys being able to
12183 cause a denial of service. ([CVE-2006-2940])
12185 *Steve Henson, Bodo Moeller*
12187 * Fix ASN.1 parsing of certain invalid structures that can result
12188 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12190 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12191 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12193 * Fix SSL client code which could crash if connecting to a
12194 malicious SSLv2 server. ([CVE-2006-4343])
12196 *Tavis Ormandy and Will Drewry, Google Security Team*
12198 * Change ciphersuite string processing so that an explicit
12199 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12200 will no longer include "AES128-SHA"), and any other similar
12201 ciphersuite (same bitmap) from *other* protocol versions (so that
12202 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12203 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12204 changes from 0.9.8b and 0.9.8d.
12208 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12210 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12211 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12213 * Change the Unix randomness entropy gathering to use poll() when
12214 possible instead of select(), since the latter has some
12215 undesirable limitations.
12217 *Darryl Miles via Richard Levitte and Bodo Moeller*
12219 * Disable rogue ciphersuites:
12221 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12222 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12223 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12225 The latter two were purportedly from
12226 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12229 Also deactivate the remaining ciphersuites from
12230 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12231 unofficial, and the ID has long expired.
12235 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12236 dual-core machines) and other potential thread-safety issues.
12240 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12242 * Adapt fipsld and the build system to link against the validated FIPS
12243 module in FIPS mode.
12247 * Fixes for VC++ 2005 build under Windows.
12251 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12252 from a Windows bash shell such as MSYS. It is autodetected from the
12253 "config" script when run from a VC++ environment. Modify standard VC++
12254 build to use fipscanister.o from the GNU make build.
12258 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12260 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12261 The value now differs depending on if you build for FIPS or not.
12262 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12263 safely run with a non-FIPSed libcrypto, as it may crash because of
12264 the difference induced by this change.
12268 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12270 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12271 (part of SSL_OP_ALL). This option used to disable the
12272 countermeasure against man-in-the-middle protocol-version
12273 rollback in the SSL 2.0 server implementation, which is a bad
12274 idea. ([CVE-2005-2969])
12276 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12277 for Information Security, National Institute of Advanced Industrial
12278 Science and Technology [AIST, Japan)]*
12280 * Minimal support for X9.31 signatures and PSS padding modes. This is
12281 mainly for FIPS compliance and not fully integrated at this stage.
12285 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12286 the exponentiation using a fixed-length exponent. (Otherwise,
12287 the information leaked through timing could expose the secret key
12288 after many signatures; cf. Bleichenbacher's attack on DSA with
12293 * Make a new fixed-window mod_exp implementation the default for
12294 RSA, DSA, and DH private-key operations so that the sequence of
12295 squares and multiplies and the memory access pattern are
12296 independent of the particular secret key. This will mitigate
12297 cache-timing and potential related attacks.
12299 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12300 and this is automatically used by BN_mod_exp_mont() if the new flag
12301 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12302 will use this BN flag for private exponents unless the flag
12303 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12304 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12306 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12308 * Change the client implementation for SSLv23_method() and
12309 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12310 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12311 (Previously, the SSL 2.0 backwards compatible Client Hello
12312 message format would be used even with SSL_OP_NO_SSLv2.)
12316 * Add support for smime-type MIME parameter in S/MIME messages which some
12321 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12322 a threadsafe manner. Modify rsa code to use new function and add calls
12323 to dsa and dh code (which had race conditions before).
12327 * Include the fixed error library code in the C error file definitions
12328 instead of fixing them up at runtime. This keeps the error code
12329 structures constant.
12333 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12335 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12338 * Fixes for newer kerberos headers. NB: the casts are needed because
12339 the 'length' field is signed on one version and unsigned on another
12340 with no (?) obvious way to tell the difference, without these VC++
12341 complains. Also the "definition" of FAR (blank) is no longer included
12342 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12343 some needed definitions.
12347 * Undo Cygwin change.
12351 * Added support for proxy certificates according to RFC 3820.
12352 Because they may be a security thread to unaware applications,
12353 they must be explicitly allowed in run-time. See
12354 docs/HOWTO/proxy_certificates.txt for further information.
12358 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12360 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12361 server and client random values. Previously
12362 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12363 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12365 This change has negligible security impact because:
12367 1. Server and client random values still have 24 bytes of pseudo random
12370 2. Server and client random values are sent in the clear in the initial
12373 3. The master secret is derived using the premaster secret (48 bytes in
12374 size for static RSA ciphersuites) as well as client server and random
12377 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12380 *Stephen Henson, reported by UK NISCC*
12382 * Use Windows randomness collection on Cygwin.
12386 * Fix hang in EGD/PRNGD query when communication socket is closed
12387 prematurely by EGD/PRNGD.
12389 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12391 * Prompt for pass phrases when appropriate for PKCS12 input format.
12395 * Back-port of selected performance improvements from development
12396 branch, as well as improved support for PowerPC platforms.
12400 * Add lots of checks for memory allocation failure, error codes to indicate
12401 failure and freeing up memory if a failure occurs.
12403 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12405 * Add new -passin argument to dgst.
12409 * Perform some character comparisons of different types in X509_NAME_cmp:
12410 this is needed for some certificates that re-encode DNs into UTF8Strings
12411 (in violation of RFC3280) and can't or won't issue name rollover
12416 * Make an explicit check during certificate validation to see that
12417 the CA setting in each certificate on the chain is correct. As a
12418 side effect always do the following basic checks on extensions,
12419 not just when there's an associated purpose to the check:
12421 - if there is an unhandled critical extension (unless the user
12422 has chosen to ignore this fault)
12423 - if the path length has been exceeded (if one is set at all)
12424 - that certain extensions fit the associated purpose (if one has
12429 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12431 * Avoid a race condition when CRLs are checked in a multi threaded
12432 environment. This would happen due to the reordering of the revoked
12433 entries during signature checking and serial number lookup. Now the
12434 encoding is cached and the serial number sort performed under a lock.
12435 Add new STACK function sk_is_sorted().
12439 * Add Delta CRL to the extension code.
12443 * Various fixes to s3_pkt.c so alerts are sent properly.
12445 *David Holmes <d.holmes@f5.com>*
12447 * Reduce the chances of duplicate issuer name and serial numbers (in
12448 violation of RFC3280) using the OpenSSL certificate creation utilities.
12449 This is done by creating a random 64 bit value for the initial serial
12450 number when a serial number file is created or when a self signed
12451 certificate is created using 'openssl req -x509'. The initial serial
12452 number file is created using 'openssl x509 -next_serial' in CA.pl
12453 rather than being initialized to 1.
12457 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12459 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12460 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12462 *Joe Orton, Steve Henson*
12464 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12467 *Joe Orton, Steve Henson*
12469 * Make it possible to have multiple active certificates with the same
12470 subject in the CA index file. This is done only if the keyword
12471 'unique_subject' is set to 'no' in the main CA section (default
12472 if 'CA_default') of the configuration file. The value is saved
12473 with the database itself in a separate index attribute file,
12474 named like the index file with '.attr' appended to the name.
12478 * X509 verify fixes. Disable broken certificate workarounds when
12479 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12480 keyUsage extension present. Don't accept CRLs with unhandled critical
12481 extensions: since verify currently doesn't process CRL extensions this
12482 rejects a CRL with *any* critical extensions. Add new verify error codes
12487 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12488 A clarification of RFC2560 will require the use of OCTET STRINGs and
12489 some implementations cannot handle the current raw format. Since OpenSSL
12490 copies and compares OCSP nonces as opaque blobs without any attempt at
12491 parsing them this should not create any compatibility issues.
12495 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12496 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12497 this HMAC (and other) operations are several times slower than OpenSSL
12502 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12504 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12506 * Use the correct content when signing type "other".
12510 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12512 * Fix various bugs revealed by running the NISCC test suite:
12514 Stop out of bounds reads in the ASN1 code when presented with
12515 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12517 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12519 If verify callback ignores invalid public key errors don't try to check
12520 certificate signature with the NULL public key.
12524 * New -ignore_err option in ocsp application to stop the server
12525 exiting on the first error in a request.
12529 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12530 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12535 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12536 extra data after the compression methods not only for TLS 1.0
12537 but also for SSL 3.0 (as required by the specification).
12539 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12541 * Change X509_certificate_type() to mark the key as exported/exportable
12542 when it's 512 *bits* long, not 512 bytes.
12546 * Change AES_cbc_encrypt() so it outputs exact multiple of
12547 blocks during encryption.
12551 * Various fixes to base64 BIO and non blocking I/O. On write
12552 flushes were not handled properly if the BIO retried. On read
12553 data was not being buffered properly and had various logic bugs.
12554 This also affects blocking I/O when the data being decoded is a
12559 * Various S/MIME bugfixes and compatibility changes:
12560 output correct application/pkcs7 MIME type if
12561 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12562 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12563 of files as .eml work). Correctly handle very long lines in MIME
12568 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12570 * Countermeasure against the Klima-Pokorny-Rosa extension of
12571 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12572 a protocol version number mismatch like a decryption error
12573 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12577 * Turn on RSA blinding by default in the default implementation
12578 to avoid a timing attack. Applications that don't want it can call
12579 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12580 They would be ill-advised to do so in most cases.
12582 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12584 * Change RSA blinding code so that it works when the PRNG is not
12585 seeded (in this case, the secret RSA exponent is abused as
12586 an unpredictable seed -- if it is not unpredictable, there
12587 is no point in blinding anyway). Make RSA blinding thread-safe
12588 by remembering the creator's thread ID in rsa->blinding and
12589 having all other threads use local one-time blinding factors
12590 (this requires more computation than sharing rsa->blinding, but
12591 avoids excessive locking; and if an RSA object is not shared
12592 between threads, blinding will still be very fast).
12596 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12597 ENGINE as defaults for all supported algorithms irrespective of
12598 the 'flags' parameter. 'flags' is now honoured, so applications
12599 should make sure they are passing it correctly.
12603 * Target "mingw" now allows native Windows code to be generated in
12604 the Cygwin environment as well as with the MinGW compiler.
12608 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12610 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12611 via timing by performing a MAC computation even if incorrect
12612 block cipher padding has been found. This is a countermeasure
12613 against active attacks where the attacker has to distinguish
12614 between bad padding and a MAC verification error. ([CVE-2003-0078])
12616 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12617 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12618 Martin Vuagnoux (EPFL, Ilion)*
12620 * Make the no-err option work as intended. The intention with no-err
12621 is not to have the whole error stack handling routines removed from
12622 libcrypto, it's only intended to remove all the function name and
12623 reason texts, thereby removing some of the footprint that may not
12624 be interesting if those errors aren't displayed anyway.
12626 NOTE: it's still possible for any application or module to have its
12627 own set of error texts inserted. The routines are there, just not
12628 used by default when no-err is given.
12632 * Add support for FreeBSD on IA64.
12634 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12636 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12637 Kerberos function mit_des_cbc_cksum(). Before this change,
12638 the value returned by DES_cbc_cksum() was like the one from
12639 mit_des_cbc_cksum(), except the bytes were swapped.
12641 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12643 * Allow an application to disable the automatic SSL chain building.
12644 Before this a rather primitive chain build was always performed in
12645 ssl3_output_cert_chain(): an application had no way to send the
12646 correct chain if the automatic operation produced an incorrect result.
12648 Now the chain builder is disabled if either:
12650 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12652 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12654 The reasoning behind this is that an application would not want the
12655 auto chain building to take place if extra chain certificates are
12656 present and it might also want a means of sending no additional
12657 certificates (for example the chain has two certificates and the
12662 * Add the possibility to build without the ENGINE framework.
12664 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12666 * Under Win32 gmtime() can return NULL: check return value in
12667 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12671 * DSA routines: under certain error conditions uninitialized BN objects
12672 could be freed. Solution: make sure initialization is performed early
12673 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12674 Nils Larsch <nla@trustcenter.de> via PR#459)
12678 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12679 checked on reconnect on the client side, therefore session resumption
12680 could still fail with a "ssl session id is different" error. This
12681 behaviour is masked when SSL_OP_ALL is used due to
12682 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12683 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12684 followup to PR #377.
12688 * IA-32 assembler support enhancements: unified ELF targets, support
12689 for SCO/Caldera platforms, fix for Cygwin shared build.
12693 * Add support for FreeBSD on sparc64. As a consequence, support for
12694 FreeBSD on non-x86 processors is separate from x86 processors on
12695 the config script, much like the NetBSD support.
12697 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12699 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12701 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12704 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12705 code (06) was taken as the first octet of the session ID and the last
12706 octet was ignored consequently. As a result SSLv2 client side session
12707 caching could not have worked due to the session ID mismatch between
12709 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12714 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12715 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12720 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12721 seems that in spite of existing for more than a year, many application
12722 author have done nothing to provide the necessary callbacks, which
12723 means that this particular engine will not work properly anywhere.
12724 This is a very unfortunate situation which forces us, in the name
12725 of usability, to give the hw_ncipher.c a static lock, which is part
12727 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12728 appear in 0.9.8 or later. We EXPECT application authors to have
12729 dealt properly with this when 0.9.8 is released (unless we actually
12730 make such changes in the libcrypto locking code that changes will
12731 have to be made anyway).
12735 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12736 octets have been read, EOF or an error occurs. Without this change
12737 some truncated ASN1 structures will not produce an error.
12741 * Disable Heimdal support, since it hasn't been fully implemented.
12742 Still give the possibility to force the use of Heimdal, but with
12743 warnings and a request that patches get sent to openssl-dev.
12747 * Add the VC-CE target, introduce the WINCE sysname, and add
12748 INSTALL.WCE and appropriate conditionals to make it build.
12750 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12752 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12753 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12754 edit numbers of the version.
12756 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12758 * Introduce safe string copy and catenation functions
12759 (BUF_strlcpy() and BUF_strlcat()).
12761 *Ben Laurie (CHATS) and Richard Levitte*
12763 * Avoid using fixed-size buffers for one-line DNs.
12765 *Ben Laurie (CHATS)*
12767 * Add BUF_MEM_grow_clean() to avoid information leakage when
12768 resizing buffers containing secrets, and use where appropriate.
12770 *Ben Laurie (CHATS)*
12772 * Avoid using fixed size buffers for configuration file location.
12774 *Ben Laurie (CHATS)*
12776 * Avoid filename truncation for various CA files.
12778 *Ben Laurie (CHATS)*
12780 * Use sizeof in preference to magic numbers.
12782 *Ben Laurie (CHATS)*
12784 * Avoid filename truncation in cert requests.
12786 *Ben Laurie (CHATS)*
12788 * Add assertions to check for (supposedly impossible) buffer
12791 *Ben Laurie (CHATS)*
12793 * Don't cache truncated DNS entries in the local cache (this could
12794 potentially lead to a spoofing attack).
12796 *Ben Laurie (CHATS)*
12798 * Fix various buffers to be large enough for hex/decimal
12799 representations in a platform independent manner.
12801 *Ben Laurie (CHATS)*
12803 * Add CRYPTO_realloc_clean() to avoid information leakage when
12804 resizing buffers containing secrets, and use where appropriate.
12806 *Ben Laurie (CHATS)*
12808 * Add BIO_indent() to avoid much slightly worrying code to do
12811 *Ben Laurie (CHATS)*
12813 * Convert sprintf()/BIO_puts() to BIO_printf().
12815 *Ben Laurie (CHATS)*
12817 * buffer_gets() could terminate with the buffer only half
12820 *Ben Laurie (CHATS)*
12822 * Add assertions to prevent user-supplied crypto functions from
12823 overflowing internal buffers by having large block sizes, etc.
12825 *Ben Laurie (CHATS)*
12827 * New OPENSSL_assert() macro (similar to assert(), but enabled
12830 *Ben Laurie (CHATS)*
12832 * Eliminate unused copy of key in RC4.
12834 *Ben Laurie (CHATS)*
12836 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12838 *Ben Laurie (CHATS)*
12840 * Fix off-by-one error in EGD path.
12842 *Ben Laurie (CHATS)*
12844 * If RANDFILE path is too long, ignore instead of truncating.
12846 *Ben Laurie (CHATS)*
12848 * Eliminate unused and incorrectly sized X.509 structure
12851 *Ben Laurie (CHATS)*
12853 * Eliminate unused and dangerous function knumber().
12855 *Ben Laurie (CHATS)*
12857 * Eliminate unused and dangerous structure, KSSL_ERR.
12859 *Ben Laurie (CHATS)*
12861 * Protect against overlong session ID context length in an encoded
12862 session object. Since these are local, this does not appear to be
12865 *Ben Laurie (CHATS)*
12867 * Change from security patch (see 0.9.6e below) that did not affect
12868 the 0.9.6 release series:
12870 Remote buffer overflow in SSL3 protocol - an attacker could
12871 supply an oversized master key in Kerberos-enabled versions.
12874 *Ben Laurie (CHATS)*
12876 * Change the SSL kerb5 codes to match RFC 2712.
12880 * Make -nameopt work fully for req and add -reqopt switch.
12882 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12884 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12886 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12888 * Make sure tests can be performed even if the corresponding algorithms
12889 have been removed entirely. This was also the last step to make
12890 OpenSSL compilable with DJGPP under all reasonable conditions.
12892 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12894 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12895 to allow version independent disabling of normally unselected ciphers,
12896 which may be activated as a side-effect of selecting a single cipher.
12898 (E.g., cipher list string "RSA" enables ciphersuites that are left
12899 out of "ALL" because they do not provide symmetric encryption.
12900 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12902 *Lutz Jaenicke, Bodo Moeller*
12904 * Add appropriate support for separate platform-dependent build
12905 directories. The recommended way to make a platform-dependent
12906 build directory is the following (tested on Linux), maybe with
12909 # Place yourself outside of the OpenSSL source tree. In
12910 # this example, the environment variable OPENSSL_SOURCE
12911 # is assumed to contain the absolute OpenSSL source directory.
12912 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12913 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12914 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12915 mkdir -p `dirname $F`
12916 ln -s $OPENSSL_SOURCE/$F $F
12919 To be absolutely sure not to disturb the source tree, a "make clean"
12920 is a good thing. If it isn't successful, don't worry about it,
12921 it probably means the source directory is very clean.
12925 * Make sure any ENGINE control commands make local copies of string
12926 pointers passed to them whenever necessary. Otherwise it is possible
12927 the caller may have overwritten (or deallocated) the original string
12928 data when a later ENGINE operation tries to use the stored values.
12930 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12932 * Improve diagnostics in file reading and command-line digests.
12934 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12936 * Add AES modes CFB and OFB to the object database. Correct an
12937 error in AES-CFB decryption.
12941 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12942 allows existing EVP_CIPHER_CTX structures to be reused after
12943 calling `EVP_*Final()`. This behaviour is used by encryption
12944 BIOs and some applications. This has the side effect that
12945 applications must explicitly clean up cipher contexts with
12946 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12950 * Check the values of dna and dnb in bn_mul_recursive before calling
12951 bn_mul_comba (a non zero value means the a or b arrays do not contain
12952 n2 elements) and fallback to bn_mul_normal if either is not zero.
12956 * Fix escaping of non-ASCII characters when using the -subj option
12957 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12961 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12962 form for "surname", serialNumber has no short form.
12963 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12964 therefore remove "mail" short name for "internet 7".
12965 The OID for unique identifiers in X509 certificates is
12966 x500UniqueIdentifier, not uniqueIdentifier.
12967 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12971 * Add an "init" command to the ENGINE config module and auto initialize
12972 ENGINEs. Without any "init" command the ENGINE will be initialized
12973 after all ctrl commands have been executed on it. If init=1 the
12974 ENGINE is initialized at that point (ctrls before that point are run
12975 on the uninitialized ENGINE and after on the initialized one). If
12976 init=0 then the ENGINE will not be initialized at all.
12980 * Fix the 'app_verify_callback' interface so that the user-defined
12981 argument is actually passed to the callback: In the
12982 SSL_CTX_set_cert_verify_callback() prototype, the callback
12983 declaration has been changed from
12986 int (*cb)(X509_STORE_CTX *,void *);
12987 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12988 i=s->ctx->app_verify_callback(&ctx)
12989 has been changed into
12990 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12992 To update applications using SSL_CTX_set_cert_verify_callback(),
12993 a dummy argument can be added to their callback functions.
12995 *D. K. Smetters <smetters@parc.xerox.com>*
12997 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12999 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
13001 * Add and OPENSSL_LOAD_CONF define which will cause
13002 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
13003 This allows older applications to transparently support certain
13004 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
13005 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
13006 load the config file and OPENSSL_add_all_algorithms_conf() which will
13007 always load it have also been added.
13011 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
13012 Adjust NIDs and EVP layer.
13014 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13016 * Config modules support in openssl utility.
13018 Most commands now load modules from the config file,
13019 though in a few (such as version) this isn't done
13020 because it couldn't be used for anything.
13022 In the case of ca and req the config file used is
13023 the same as the utility itself: that is the -config
13024 command line option can be used to specify an
13029 * Move default behaviour from OPENSSL_config(). If appname is NULL
13030 use "openssl_conf" if filename is NULL use default openssl config file.
13034 * Add an argument to OPENSSL_config() to allow the use of an alternative
13035 config section name. Add a new flag to tolerate a missing config file
13036 and move code to CONF_modules_load_file().
13040 * Support for crypto accelerator cards from Accelerated Encryption
13041 Processing, www.aep.ie. (Use engine 'aep')
13042 The support was copied from 0.9.6c [engine] and adapted/corrected
13043 to work with the new engine framework.
13045 *AEP Inc. and Richard Levitte*
13047 * Support for SureWare crypto accelerator cards from Baltimore
13048 Technologies. (Use engine 'sureware')
13049 The support was copied from 0.9.6c [engine] and adapted
13050 to work with the new engine framework.
13054 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
13055 make the newer ENGINE framework commands for the CHIL engine work.
13057 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
13059 * Make it possible to produce shared libraries on ReliantUNIX.
13061 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
13063 * Add the configuration target debug-linux-ppro.
13064 Make 'openssl rsa' use the general key loading routines
13065 implemented in `apps.c`, and make those routines able to
13066 handle the key format FORMAT_NETSCAPE and the variant
13069 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13071 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13073 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13075 * Add -keyform to rsautl, and document -engine.
13077 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
13079 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
13080 BIO_R_NO_SUCH_FILE error code rather than the generic
13081 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
13085 * Add new functions
13086 ERR_peek_last_error
13087 ERR_peek_last_error_line
13088 ERR_peek_last_error_line_data.
13089 These are similar to
13091 ERR_peek_error_line
13092 ERR_peek_error_line_data,
13093 but report on the latest error recorded rather than the first one
13094 still in the error queue.
13096 *Ben Laurie, Bodo Moeller*
13098 * default_algorithms option in ENGINE config module. This allows things
13100 default_algorithms = ALL
13101 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
13105 * Preliminary ENGINE config module.
13109 * New experimental application configuration code.
13113 * Change the AES code to follow the same name structure as all other
13114 symmetric ciphers, and behave the same way. Move everything to
13115 the directory crypto/aes, thereby obsoleting crypto/rijndael.
13117 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13119 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
13121 *Ben Laurie and Theo de Raadt*
13123 * Add option to output public keys in req command.
13125 *Massimiliano Pala madwolf@openca.org*
13127 * Use wNAFs in EC_POINTs_mul() for improved efficiency
13128 (up to about 10% better than before for P-192 and P-224).
13132 * New functions/macros
13134 SSL_CTX_set_msg_callback(ctx, cb)
13135 SSL_CTX_set_msg_callback_arg(ctx, arg)
13136 SSL_set_msg_callback(ssl, cb)
13137 SSL_set_msg_callback_arg(ssl, arg)
13139 to request calling a callback function
13141 void cb(int write_p, int version, int content_type,
13142 const void *buf, size_t len, SSL *ssl, void *arg)
13144 whenever a protocol message has been completely received
13145 (write_p == 0) or sent (write_p == 1). Here 'version' is the
13146 protocol version according to which the SSL library interprets
13147 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
13148 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
13149 the content type as defined in the SSL 3.0/TLS 1.0 protocol
13150 specification (change_cipher_spec(20), alert(21), handshake(22)).
13151 'buf' and 'len' point to the actual message, 'ssl' to the
13152 SSL object, and 'arg' is the application-defined value set by
13153 SSL[_CTX]_set_msg_callback_arg().
13155 'openssl s_client' and 'openssl s_server' have new '-msg' options
13156 to enable a callback that displays all protocol messages.
13160 * Change the shared library support so shared libraries are built as
13161 soon as the corresponding static library is finished, and thereby get
13162 openssl and the test programs linked against the shared library.
13163 This still only happens when the keyword "shard" has been given to
13164 the configuration scripts.
13166 NOTE: shared library support is still an experimental thing, and
13167 backward binary compatibility is still not guaranteed.
13169 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13171 * Add support for Subject Information Access extension.
13173 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13175 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13176 additional bytes when new memory had to be allocated, not just
13177 when reusing an existing buffer.
13181 * New command line and configuration option 'utf8' for the req command.
13182 This allows field values to be specified as UTF8 strings.
13186 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13187 runs for the former and machine-readable output for the latter.
13191 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13192 of the e-mail address in the DN (i.e., it will go into a certificate
13193 extension only). The new configuration file option 'email_in_dn = no'
13194 has the same effect.
13196 *Massimiliano Pala madwolf@openca.org*
13198 * Change all functions with names starting with `des_` to be starting
13199 with `DES_` instead. Add wrappers that are compatible with libdes,
13200 but are named `_ossl_old_des_*`. Finally, add macros that map the
13201 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13202 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13203 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13206 Since we provide two compatibility mappings, the user needs to
13207 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13208 compatibility is desired. The default (i.e., when that macro
13209 isn't defined) is OpenSSL 0.9.6c compatibility.
13211 There are also macros that enable and disable the support of old
13212 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13213 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13214 are defined, the default will apply: to support the old des routines.
13216 In either case, one must include openssl/des.h to get the correct
13217 definitions. Do not try to just include openssl/des_old.h, that
13220 NOTE: This is a major break of an old API into a new one. Software
13221 authors are encouraged to switch to the `DES_` style functions. Some
13222 time in the future, des_old.h and the libdes compatibility functions
13223 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13224 default), and then completely removed.
13228 * Test for certificates which contain unsupported critical extensions.
13229 If such a certificate is found during a verify operation it is
13230 rejected by default: this behaviour can be overridden by either
13231 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13232 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13233 X509_supported_extension() has also been added which returns 1 if a
13234 particular extension is supported.
13238 * Modify the behaviour of EVP cipher functions in similar way to digests
13239 to retain compatibility with existing code.
13243 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13244 compatibility with existing code. In particular the 'ctx' parameter does
13245 not have to be to be initialized before the call to EVP_DigestInit() and
13246 it is tidied up after a call to EVP_DigestFinal(). New function
13247 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13248 EVP_MD_CTX_copy() changed to not require the destination to be
13249 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13250 requires the destination to be valid.
13252 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13253 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13257 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13258 so that complete 'Handshake' protocol structures are kept in memory
13259 instead of overwriting 'msg_type' and 'length' with 'body' data.
13263 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13265 *Massimo Santin via Richard Levitte*
13267 * Major restructuring to the underlying ENGINE code. This includes
13268 reduction of linker bloat, separation of pure "ENGINE" manipulation
13269 (initialisation, etc) from functionality dealing with implementations
13270 of specific crypto interfaces. This change also introduces integrated
13271 support for symmetric ciphers and digest implementations - so ENGINEs
13272 can now accelerate these by providing EVP_CIPHER and EVP_MD
13273 implementations of their own. This is detailed in
13274 [crypto/engine/README.md](crypto/engine/README.md)
13275 as it couldn't be adequately described here. However, there are a few
13276 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13277 were changed in the original introduction of ENGINE code have now
13278 reverted back - the hooking from this code to ENGINE is now a good
13279 deal more passive and at run-time, operations deal directly with
13280 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13281 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13282 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13283 they were not being used by the framework as there is no concept of a
13284 BIGNUM_METHOD and they could not be generalised to the new
13285 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13286 ENGINE_cpy() has been removed as it cannot be consistently defined in
13291 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13295 * Change mkdef.pl to sort symbols that get the same entry number,
13296 and make sure the automatically generated functions `ERR_load_*`
13297 become part of libeay.num as well.
13301 * New function SSL_renegotiate_pending(). This returns true once
13302 renegotiation has been requested (either SSL_renegotiate() call
13303 or HelloRequest/ClientHello received from the peer) and becomes
13304 false once a handshake has been completed.
13305 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13306 sends a HelloRequest, but does not ensure that a handshake takes
13307 place. SSL_renegotiate_pending() is useful for checking if the
13308 client has followed the request.)
13312 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13313 By default, clients may request session resumption even during
13314 renegotiation (if session ID contexts permit); with this option,
13315 session resumption is possible only in the first handshake.
13317 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13318 more bits available for options that should not be part of
13319 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13323 * Add some demos for certificate and certificate request creation.
13327 * Make maximum certificate chain size accepted from the peer application
13328 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13329 "Douglas E. Engert" <deengert@anl.gov>.
13333 * Add support for shared libraries for Unixware-7
13334 (Boyd Lynn Gerber <gerberb@zenez.com>).
13338 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13339 be done prior to destruction. Use this to unload error strings from
13340 ENGINEs that load their own error strings. NB: This adds two new API
13341 functions to "get" and "set" this destroy handler in an ENGINE.
13345 * Alter all existing ENGINE implementations (except "openssl" and
13346 "openbsd") to dynamically instantiate their own error strings. This
13347 makes them more flexible to be built both as statically-linked ENGINEs
13348 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13349 Also, add stub code to each that makes building them as self-contained
13350 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13354 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13355 implementations into applications that are completely implemented in
13356 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13357 commands that can be used to configure what shared-library to load and
13358 to control aspects of the way it is handled. Also, made an update to
13359 the [README-Engine.md](README-Engine.md) file
13360 that brings its information up-to-date and
13361 provides some information and instructions on the "dynamic" ENGINE
13362 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13366 * Make it possible to unload ranges of ERR strings with a new
13367 "ERR_unload_strings" function.
13371 * Add a copy() function to EVP_MD.
13375 * Make EVP_MD routines take a context pointer instead of just the
13376 md_data void pointer.
13380 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13381 that the digest can only process a single chunk of data
13382 (typically because it is provided by a piece of
13383 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13384 is only going to provide a single chunk of data, and hence the
13385 framework needn't accumulate the data for oneshot drivers.
13389 * As with "ERR", make it possible to replace the underlying "ex_data"
13390 functions. This change also alters the storage and management of global
13391 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13392 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13393 index counters. The API functions that use this state have been changed
13394 to take a "class_index" rather than pointers to the class's local STACK
13395 and counter, and there is now an API function to dynamically create new
13396 classes. This centralisation allows us to (a) plug a lot of the
13397 thread-safety problems that existed, and (b) makes it possible to clean
13398 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13399 such data would previously have always leaked in application code and
13400 workarounds were in place to make the memory debugging turn a blind eye
13401 to it. Application code that doesn't use this new function will still
13402 leak as before, but their memory debugging output will announce it now
13403 rather than letting it slide.
13405 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13406 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13407 has a return value to indicate success or failure.
13411 * Make it possible to replace the underlying "ERR" functions such that the
13412 global state (2 LHASH tables and 2 locks) is only used by the "default"
13413 implementation. This change also adds two functions to "get" and "set"
13414 the implementation prior to it being automatically set the first time
13415 any other ERR function takes place. Ie. an application can call "get",
13416 pass the return value to a module it has just loaded, and that module
13417 can call its own "set" function using that value. This means the
13418 module's "ERR" operations will use (and modify) the error state in the
13419 application and not in its own statically linked copy of OpenSSL code.
13423 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13424 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13425 the operation, and provides a more encapsulated way for external code
13426 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13427 to use these functions rather than manually incrementing the counts.
13429 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13433 * Add EVP test program.
13437 * Add symmetric cipher support to ENGINE. Expect the API to change!
13441 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13442 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13443 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13444 These allow a CRL to be built without having to access X509_CRL fields
13445 directly. Modify 'ca' application to use new functions.
13449 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13450 bug workarounds. Rollback attack detection is a security feature.
13451 The problem will only arise on OpenSSL servers when TLSv1 is not
13452 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13453 Software authors not wanting to support TLSv1 will have special reasons
13454 for their choice and can explicitly enable this option.
13456 *Bodo Moeller, Lutz Jaenicke*
13458 * Rationalise EVP so it can be extended: don't include a union of
13459 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13460 (similar to those existing for EVP_CIPHER_CTX).
13465 EVP_MD_CTX_init(&md); /* new function call */
13466 EVP_DigestInit(&md, EVP_sha1());
13467 EVP_DigestUpdate(&md, in, len);
13468 EVP_DigestFinal(&md, out, NULL);
13469 EVP_MD_CTX_cleanup(&md); /* new function call */
13473 * Make DES key schedule conform to the usual scheme, as well as
13474 correcting its structure. This means that calls to DES functions
13475 now have to pass a pointer to a des_key_schedule instead of a
13476 plain des_key_schedule (which was actually always a pointer
13479 des_key_schedule ks;
13481 des_set_key_checked(..., &ks);
13482 des_ncbc_encrypt(..., &ks, ...);
13484 (Note that a later change renames 'des_...' into 'DES_...'.)
13488 * Initial reduction of linker bloat: the use of some functions, such as
13489 PEM causes large amounts of unused functions to be linked in due to
13490 poor organisation. For example pem_all.c contains every PEM function
13491 which has a knock on effect of linking in large amounts of (unused)
13492 ASN1 code. Grouping together similar functions and splitting unrelated
13493 functions prevents this.
13497 * Cleanup of EVP macros.
13501 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13502 correct `_ecb suffix`.
13506 * Add initial OCSP responder support to ocsp application. The
13507 revocation information is handled using the text based index
13508 use by the ca application. The responder can either handle
13509 requests generated internally, supplied in files (for example
13510 via a CGI script) or using an internal minimal server.
13514 * Add configuration choices to get zlib compression for TLS.
13518 * Changes to Kerberos SSL for RFC 2712 compliance:
13519 1. Implemented real KerberosWrapper, instead of just using
13520 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13521 2. Implemented optional authenticator field of KerberosWrapper.
13523 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13524 and authenticator structs; see crypto/krb5/.
13526 Generalized Kerberos calls to support multiple Kerberos libraries.
13527 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13528 via Richard Levitte*
13530 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13531 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13532 values for each of the key sizes rather than having just
13533 parameters (and 'speed' generating keys each time).
13537 * Speed up EVP routines.
13540 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13541 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13542 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13543 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13545 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13546 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13547 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13550 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13552 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13556 * Added the OS2-EMX target.
13558 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13560 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13561 New functions to support `NCONF` routines in extension code.
13562 New function `CONF_set_nconf()`
13563 to allow functions which take an `NCONF` to also handle the old `LHASH`
13564 structure: this means that the old `CONF` compatible routines can be
13565 retained (in particular w.rt. extensions) without having to duplicate the
13566 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13570 * Enhance the general user interface with mechanisms for inner control
13571 and with possibilities to have yes/no kind of prompts.
13575 * Change all calls to low-level digest routines in the library and
13576 applications to use EVP. Add missing calls to HMAC_cleanup() and
13577 don't assume HMAC_CTX can be copied using memcpy().
13579 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13581 * Add the possibility to control engines through control names but with
13582 arbitrary arguments instead of just a string.
13583 Change the key loaders to take a UI_METHOD instead of a callback
13584 function pointer. NOTE: this breaks binary compatibility with earlier
13585 versions of OpenSSL [engine].
13586 Adapt the nCipher code for these new conditions and add a card insertion
13591 * Enhance the general user interface with mechanisms to better support
13592 dialog box interfaces, application-defined prompts, the possibility
13593 to use defaults (for example default passwords from somewhere else)
13594 and interrupts/cancellations.
13598 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13599 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13603 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13604 tidy up some unnecessarily weird code in 'sk_new()').
13606 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13608 * Change the key loading routines for ENGINEs to use the same kind
13609 callback (pem_password_cb) as all other routines that need this
13614 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13615 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13616 than this minimum value is recommended.
13620 * New random seeder for OpenVMS, using the system process statistics
13621 that are easily reachable.
13625 * Windows apparently can't transparently handle global
13626 variables defined in DLLs. Initialisations such as:
13628 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13630 won't compile. This is used by the any applications that need to
13631 declare their own ASN1 modules. This was fixed by adding the option
13632 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13633 needed for static libraries under Win32.
13637 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13638 setting of purpose and trust fields. New X509_STORE trust and
13639 purpose functions and tidy up setting in other SSL functions.
13643 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13644 structure. These are inherited by X509_STORE_CTX when it is
13645 initialised. This allows various defaults to be set in the
13646 X509_STORE structure (such as flags for CRL checking and custom
13647 purpose or trust settings) for functions which only use X509_STORE_CTX
13648 internally such as S/MIME.
13650 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13651 trust settings if they are not set in X509_STORE. This allows X509_STORE
13652 purposes and trust (in S/MIME for example) to override any set by default.
13654 Add command line options for CRL checking to smime, s_client and s_server
13659 * Initial CRL based revocation checking. If the CRL checking flag(s)
13660 are set then the CRL is looked up in the X509_STORE structure and
13661 its validity and signature checked, then if the certificate is found
13662 in the CRL the verify fails with a revoked error.
13664 Various new CRL related callbacks added to X509_STORE_CTX structure.
13666 Command line options added to 'verify' application to support this.
13668 This needs some additional work, such as being able to handle multiple
13669 CRLs with different times, extension based lookup (rather than just
13670 by subject name) and ultimately more complete V2 CRL extension
13675 * Add a general user interface API (crypto/ui/). This is designed
13676 to replace things like des_read_password and friends (backward
13677 compatibility functions using this new API are provided).
13678 The purpose is to remove prompting functions from the DES code
13679 section as well as provide for prompting through dialog boxes in
13680 a window system and the like.
13684 * Add "ex_data" support to ENGINE so implementations can add state at a
13685 per-structure level rather than having to store it globally.
13689 * Make it possible for ENGINE structures to be copied when retrieved by
13690 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13691 This causes the "original" ENGINE structure to act like a template,
13692 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13693 operational state can be localised to each ENGINE structure, despite the
13694 fact they all share the same "methods". New ENGINE structures returned in
13695 this case have no functional references and the return value is the single
13696 structural reference. This matches the single structural reference returned
13697 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13702 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13703 needs to match any other type at all we need to manually clear the
13708 * Changes to the "openssl engine" utility to include;
13709 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13710 about an ENGINE's available control commands.
13711 - executing control commands from command line arguments using the
13712 '-pre' and '-post' switches. '-post' is only used if '-t' is
13713 specified and the ENGINE is successfully initialised. The syntax for
13714 the individual commands are colon-separated, for example;
13715 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13719 * New dynamic control command support for ENGINEs. ENGINEs can now
13720 declare their own commands (numbers), names (strings), descriptions,
13721 and input types for run-time discovery by calling applications. A
13722 subset of these commands are implicitly classed as "executable"
13723 depending on their input type, and only these can be invoked through
13724 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13725 can be based on user input, config files, etc). The distinction is
13726 that "executable" commands cannot return anything other than a boolean
13727 result and can only support numeric or string input, whereas some
13728 discoverable commands may only be for direct use through
13729 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13730 pointers, or other custom uses. The "executable" commands are to
13731 support parameterisations of ENGINE behaviour that can be
13732 unambiguously defined by ENGINEs and used consistently across any
13733 OpenSSL-based application. Commands have been added to all the
13734 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13735 control over shared-library paths without source code alterations.
13739 * Changed all ENGINE implementations to dynamically allocate their
13740 ENGINEs rather than declaring them statically. Apart from this being
13741 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13742 this also allows the implementations to compile without using the
13743 internal engine_int.h header.
13747 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13748 'const' value. Any code that should be able to modify a RAND_METHOD
13749 should already have non-const pointers to it (ie. they should only
13750 modify their own ones).
13754 * Made a variety of little tweaks to the ENGINE code.
13755 - "atalla" and "ubsec" string definitions were moved from header files
13756 to C code. "nuron" string definitions were placed in variables
13757 rather than hard-coded - allowing parameterisation of these values
13758 later on via ctrl() commands.
13759 - Removed unused "#if 0"'d code.
13760 - Fixed engine list iteration code so it uses ENGINE_free() to release
13761 structural references.
13762 - Constified the RAND_METHOD element of ENGINE structures.
13763 - Constified various get/set functions as appropriate and added
13764 missing functions (including a catch-all ENGINE_cpy that duplicates
13765 all ENGINE values onto a new ENGINE except reference counts/state).
13766 - Removed NULL parameter checks in get/set functions. Setting a method
13767 or function to NULL is a way of cancelling out a previously set
13768 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13769 and doesn't justify the extra error symbols and code.
13770 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13771 flags from engine_int.h to engine.h.
13772 - Changed prototypes for ENGINE handler functions (init(), finish(),
13773 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13777 * Implement binary inversion algorithm for BN_mod_inverse in addition
13778 to the algorithm using long division. The binary algorithm can be
13779 used only if the modulus is odd. On 32-bit systems, it is faster
13780 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13781 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13782 up to 450 bits. In 64-bit environments, the binary algorithm
13783 appears to be advantageous for much longer moduli; here we use it
13784 for moduli up to 2048 bits.
13788 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13789 could not support the combine flag in choice fields.
13793 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13794 extensions from a certificate request to the certificate.
13798 * Allow multiple 'certopt' and 'nameopt' options to be separated
13799 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13800 file: this allows the display of the certificate about to be
13801 signed to be customised, to allow certain fields to be included
13802 or excluded and extension details. The old system didn't display
13803 multicharacter strings properly, omitted fields not in the policy
13804 and couldn't display additional details such as extensions.
13808 * Function EC_POINTs_mul for multiple scalar multiplication
13809 of an arbitrary number of elliptic curve points
13810 \sum scalars[i]*points[i],
13811 optionally including the generator defined for the EC_GROUP:
13812 scalar*generator + \sum scalars[i]*points[i].
13814 EC_POINT_mul is a simple wrapper function for the typical case
13815 that the point list has just one item (besides the optional
13820 * First EC_METHODs for curves over GF(p):
13822 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13823 operations and provides various method functions that can also
13824 operate with faster implementations of modular arithmetic.
13826 EC_GFp_mont_method() reuses most functions that are part of
13827 EC_GFp_simple_method, but uses Montgomery arithmetic.
13829 *Bodo Moeller; point addition and point doubling
13830 implementation directly derived from source code provided by
13831 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13833 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13834 crypto/ec/ec_lib.c):
13836 Curves are EC_GROUP objects (with an optional group generator)
13837 based on EC_METHODs that are built into the library.
13839 Points are EC_POINT objects based on EC_GROUP objects.
13841 Most of the framework would be able to handle curves over arbitrary
13842 finite fields, but as there are no obvious types for fields other
13843 than GF(p), some functions are limited to that for now.
13847 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13848 that the file contains a complete HTTP response.
13852 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13853 change the def and num file printf format specifier from "%-40sXXX"
13854 to "%-39s XXX". The latter will always guarantee a space after the
13855 field while the former will cause them to run together if the field
13856 is 40 of more characters long.
13860 * Constify the cipher and digest 'method' functions and structures
13861 and modify related functions to take constant EVP_MD and EVP_CIPHER
13866 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13867 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13871 * Modify `EVP_Digest*()` routines so they now return values. Although the
13872 internal software routines can never fail additional hardware versions
13877 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13879 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13880 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13883 ERR_R_NESTED_ASN1_ERROR
13885 ERR_R_MISSING_ASN1_EOS
13886 were 4 .. 9, conflicting with
13887 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13889 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13890 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13892 Add new error code 'ERR_R_INTERNAL_ERROR'.
13896 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13901 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13902 sets the subject name for a new request or supersedes the
13903 subject name in a given request. Formats that can be parsed are
13904 'CN=Some Name, OU=myOU, C=IT'
13906 'CN=Some Name/OU=myOU/C=IT'.
13908 Add options '-batch' and '-verbose' to 'openssl req'.
13910 *Massimiliano Pala <madwolf@hackmasters.net>*
13912 * Introduce the possibility to access global variables through
13913 functions on platform were that's the best way to handle exporting
13914 global variables in shared libraries. To enable this functionality,
13915 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13916 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13917 is normally done by Configure or something similar).
13919 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13920 in the source file (foo.c) like this:
13922 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13923 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13925 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13926 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13928 OPENSSL_DECLARE_GLOBAL(int,foo);
13929 #define foo OPENSSL_GLOBAL_REF(foo)
13930 OPENSSL_DECLARE_GLOBAL(double,bar);
13931 #define bar OPENSSL_GLOBAL_REF(bar)
13933 The #defines are very important, and therefore so is including the
13934 header file everywhere where the defined globals are used.
13936 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13937 of ASN.1 items, but that structure is a bit different.
13939 The largest change is in util/mkdef.pl which has been enhanced with
13940 better and easier to understand logic to choose which symbols should
13941 go into the Windows .def files as well as a number of fixes and code
13942 cleanup (among others, algorithm keywords are now sorted
13943 lexicographically to avoid constant rewrites).
13947 * In BN_div() keep a copy of the sign of 'num' before writing the
13948 result to 'rm' because if rm==num the value will be overwritten
13949 and produce the wrong result if 'num' is negative: this caused
13950 problems with BN_mod() and BN_nnmod().
13954 * Function OCSP_request_verify(). This checks the signature on an
13955 OCSP request and verifies the signer certificate. The signer
13956 certificate is just checked for a generic purpose and OCSP request
13961 * Add OCSP_check_validity() function to check the validity of OCSP
13962 responses. OCSP responses are prepared in real time and may only
13963 be a few seconds old. Simply checking that the current time lies
13964 between thisUpdate and nextUpdate max reject otherwise valid responses
13965 caused by either OCSP responder or client clock inaccuracy. Instead
13966 we allow thisUpdate and nextUpdate to fall within a certain period of
13967 the current time. The age of the response can also optionally be
13968 checked. Two new options -validity_period and -status_age added to
13973 * If signature or public key algorithm is unrecognized print out its
13974 OID rather that just UNKNOWN.
13978 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13979 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13980 ID to be generated from the issuer certificate alone which can then be
13981 passed to OCSP_id_issuer_cmp().
13985 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13986 ASN1 modules to export functions returning ASN1_ITEM pointers
13987 instead of the ASN1_ITEM structures themselves. This adds several
13988 new macros which allow the underlying ASN1 function/structure to
13989 be accessed transparently. As a result code should not use ASN1_ITEM
13990 references directly (such as &X509_it) but instead use the relevant
13991 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13992 use of the new ASN1 code on platforms where exporting structures
13993 is problematical (for example in shared libraries) but exporting
13994 functions returning pointers to structures is not.
13998 * Add support for overriding the generation of SSL/TLS session IDs.
13999 These callbacks can be registered either in an SSL_CTX or per SSL.
14000 The purpose of this is to allow applications to control, if they wish,
14001 the arbitrary values chosen for use as session IDs, particularly as it
14002 can be useful for session caching in multiple-server environments. A
14003 command-line switch for testing this (and any client code that wishes
14004 to use such a feature) has been added to "s_server".
14006 *Geoff Thorpe, Lutz Jaenicke*
14008 * Modify mkdef.pl to recognise and parse preprocessor conditionals
14009 of the form `#if defined(...) || defined(...) || ...` and
14010 `#if !defined(...) && !defined(...) && ...`. This also avoids
14011 the growing number of special cases it was previously handling.
14015 * Make all configuration macros available for application by making
14016 sure they are available in opensslconf.h, by giving them names starting
14017 with `OPENSSL_` to avoid conflicts with other packages and by making
14018 sure e_os2.h will cover all platform-specific cases together with
14020 Additionally, it is now possible to define configuration/platform-
14021 specific names (called "system identities"). In the C code, these
14022 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
14023 macro with the name beginning with `OPENSSL_SYS_`, which is determined
14024 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
14029 * New option -set_serial to 'req' and 'x509' this allows the serial
14030 number to use to be specified on the command line. Previously self
14031 signed certificates were hard coded with serial number 0 and the
14032 CA options of 'x509' had to use a serial number in a file which was
14037 * New options to 'ca' utility to support V2 CRL entry extensions.
14038 Currently CRL reason, invalidity date and hold instruction are
14039 supported. Add new CRL extensions to V3 code and some new objects.
14043 * New function EVP_CIPHER_CTX_set_padding() this is used to
14044 disable standard block padding (aka PKCS#5 padding) in the EVP
14045 API, which was previously mandatory. This means that the data is
14046 not padded in any way and so the total length much be a multiple
14047 of the block size, otherwise an error occurs.
14051 * Initial (incomplete) OCSP SSL support.
14055 * New function OCSP_parse_url(). This splits up a URL into its host,
14056 port and path components: primarily to parse OCSP URLs. New -url
14057 option to ocsp utility.
14061 * New nonce behavior. The return value of OCSP_check_nonce() now
14062 reflects the various checks performed. Applications can decide
14063 whether to tolerate certain situations such as an absent nonce
14064 in a response when one was present in a request: the ocsp application
14065 just prints out a warning. New function OCSP_add1_basic_nonce()
14066 this is to allow responders to include a nonce in a response even if
14067 the request is nonce-less.
14071 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
14072 skipped when using openssl x509 multiple times on a single input file,
14073 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
14077 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
14078 set string type: to handle setting ASN1_TIME structures. Fix ca
14079 utility to correctly initialize revocation date of CRLs.
14083 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
14084 the clients preferred ciphersuites and rather use its own preferences.
14085 Should help to work around M$ SGC (Server Gated Cryptography) bug in
14086 Internet Explorer by ensuring unchanged hash method during stepup.
14087 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
14091 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
14092 to aes and add a new 'exist' option to print out symbols that don't
14097 * Additional options to ocsp utility to allow flags to be set and
14098 additional certificates supplied.
14102 * Add the option -VAfile to 'openssl ocsp', so the user can give the
14103 OCSP client a number of certificate to only verify the response
14108 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
14109 handle the new API. Currently only ECB, CBC modes supported. Add new
14112 Add TLS AES ciphersuites as described in RFC3268, "Advanced
14113 Encryption Standard (AES) Ciphersuites for Transport Layer
14114 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
14115 not enabled by default and were not part of the "ALL" ciphersuite
14116 alias because they were not yet official; they could be
14117 explicitly requested by specifying the "AESdraft" ciphersuite
14118 group alias. In the final release of OpenSSL 0.9.7, the group
14119 alias is called "AES" and is part of "ALL".)
14121 *Ben Laurie, Steve Henson, Bodo Moeller*
14123 * New function OCSP_copy_nonce() to copy nonce value (if present) from
14124 request to response.
14128 * Functions for OCSP responders. OCSP_request_onereq_count(),
14129 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
14130 extract information from a certificate request. OCSP_response_create()
14131 creates a response and optionally adds a basic response structure.
14132 OCSP_basic_add1_status() adds a complete single response to a basic
14133 response and returns the OCSP_SINGLERESP structure just added (to allow
14134 extensions to be included for example). OCSP_basic_add1_cert() adds a
14135 certificate to a basic response and OCSP_basic_sign() signs a basic
14136 response with various flags. New helper functions ASN1_TIME_check()
14137 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
14138 (converts ASN1_TIME to GeneralizedTime).
14142 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
14143 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
14144 structure from a certificate. X509_pubkey_digest() digests the public_key
14145 contents: this is used in various key identifiers.
14149 * Make sk_sort() tolerate a NULL argument.
14151 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14153 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14154 passed by the function are trusted implicitly. If any of them signed the
14155 response then it is assumed to be valid and is not verified.
14159 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14160 to data. This was previously part of the PKCS7 ASN1 code. This
14161 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14162 *Steve Henson, reported by Kenneth R. Robinette
14163 <support@securenetterm.com>*
14165 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14166 routines: without these tracing memory leaks is very painful.
14167 Fix leaks in PKCS12 and PKCS7 routines.
14171 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14172 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14173 effectively meant GeneralizedTime would never be used. Now it
14174 is initialised to -1 but X509_time_adj() now has to check the value
14175 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14176 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14177 *Steve Henson, reported by Kenneth R. Robinette
14178 <support@securenetterm.com>*
14180 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14181 result in a zero length in the ASN1_INTEGER structure which was
14182 not consistent with the structure when d2i_ASN1_INTEGER() was used
14183 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14184 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14185 where it did not print out a minus for negative ASN1_INTEGER.
14189 * Add summary printout to ocsp utility. The various functions which
14190 convert status values to strings have been renamed to:
14191 OCSP_response_status_str(), OCSP_cert_status_str() and
14192 OCSP_crl_reason_str() and are no longer static. New options
14193 to verify nonce values and to disable verification. OCSP response
14194 printout format cleaned up.
14198 * Add additional OCSP certificate checks. These are those specified
14199 in RFC2560. This consists of two separate checks: the CA of the
14200 certificate being checked must either be the OCSP signer certificate
14201 or the issuer of the OCSP signer certificate. In the latter case the
14202 OCSP signer certificate must contain the OCSP signing extended key
14203 usage. This check is performed by attempting to match the OCSP
14204 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14205 in the OCSP_CERTID structures of the response.
14209 * Initial OCSP certificate verification added to OCSP_basic_verify()
14210 and related routines. This uses the standard OpenSSL certificate
14211 verify routines to perform initial checks (just CA validity) and
14212 to obtain the certificate chain. Then additional checks will be
14213 performed on the chain. Currently the root CA is checked to see
14214 if it is explicitly trusted for OCSP signing. This is used to set
14215 a root CA as a global signing root: that is any certificate that
14216 chains to that CA is an acceptable OCSP signing certificate.
14220 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14221 extensions from a separate configuration file.
14222 As when reading extensions from the main configuration file,
14223 the '-extensions ...' option may be used for specifying the
14226 *Massimiliano Pala <madwolf@comune.modena.it>*
14228 * New OCSP utility. Allows OCSP requests to be generated or
14229 read. The request can be sent to a responder and the output
14230 parsed, outputted or printed in text form. Not complete yet:
14231 still needs to check the OCSP response validity.
14235 * New subcommands for 'openssl ca':
14236 `openssl ca -status <serial>` prints the status of the cert with
14237 the given serial number (according to the index file).
14238 `openssl ca -updatedb` updates the expiry status of certificates
14241 *Massimiliano Pala <madwolf@comune.modena.it>*
14243 * New '-newreq-nodes' command option to CA.pl. This is like
14244 '-newreq', but calls 'openssl req' with the '-nodes' option
14245 so that the resulting key is not encrypted.
14247 *Damien Miller <djm@mindrot.org>*
14249 * New configuration for the GNU Hurd.
14251 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14253 * Initial code to implement OCSP basic response verify. This
14254 is currently incomplete. Currently just finds the signer's
14255 certificate and verifies the signature on the response.
14259 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14260 value of OPENSSLDIR. This is available via the new '-d' option
14261 to 'openssl version', and is also included in 'openssl version -a'.
14265 * Allowing defining memory allocation callbacks that will be given
14266 file name and line number information in additional arguments
14267 (a `const char*` and an int). The basic functionality remains, as
14268 well as the original possibility to just replace malloc(),
14269 realloc() and free() by functions that do not know about these
14270 additional arguments. To register and find out the current
14271 settings for extended allocation functions, the following
14272 functions are provided:
14274 CRYPTO_set_mem_ex_functions
14275 CRYPTO_set_locked_mem_ex_functions
14276 CRYPTO_get_mem_ex_functions
14277 CRYPTO_get_locked_mem_ex_functions
14279 These work the same way as CRYPTO_set_mem_functions and friends.
14280 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14281 extended allocation function is enabled.
14282 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14283 a conventional allocation function is enabled.
14285 *Richard Levitte, Bodo Moeller*
14287 * Finish off removing the remaining LHASH function pointer casts.
14288 There should no longer be any prototype-casting required when using
14289 the LHASH abstraction, and any casts that remain are "bugs". See
14290 the callback types and macros at the head of lhash.h for details
14291 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14295 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14296 If /dev/[u]random devices are not available or do not return enough
14297 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14299 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14300 /etc/entropy will be queried once each in this sequence, querying stops
14301 when enough entropy was collected without querying more sockets.
14305 * Change the Unix RAND_poll() variant to be able to poll several
14306 random devices, as specified by DEVRANDOM, until a sufficient amount
14307 of data has been collected. We spend at most 10 ms on each file
14308 (select timeout) and read in non-blocking mode. DEVRANDOM now
14309 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14310 (previously it was just the string "/dev/urandom"), so on typical
14311 platforms the 10 ms delay will never occur.
14312 Also separate out the Unix variant to its own file, rand_unix.c.
14313 For VMS, there's a currently-empty rand_vms.c.
14317 * Move OCSP client related routines to ocsp_cl.c. These
14318 provide utility functions which an application needing
14319 to issue a request to an OCSP responder and analyse the
14320 response will typically need: as opposed to those which an
14321 OCSP responder itself would need which will be added later.
14323 OCSP_request_sign() signs an OCSP request with an API similar
14324 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14325 response. OCSP_response_get1_basic() extracts basic response
14326 from response. OCSP_resp_find_status(): finds and extracts status
14327 information from an OCSP_CERTID structure (which will be created
14328 when the request structure is built). These are built from lower
14329 level functions which work on OCSP_SINGLERESP structures but
14330 won't normally be used unless the application wishes to examine
14331 extensions in the OCSP response for example.
14333 Replace nonce routines with a pair of functions.
14334 OCSP_request_add1_nonce() adds a nonce value and optionally
14335 generates a random value. OCSP_check_nonce() checks the
14336 validity of the nonce in an OCSP response.
14340 * Change function OCSP_request_add() to OCSP_request_add0_id().
14341 This doesn't copy the supplied OCSP_CERTID and avoids the
14342 need to free up the newly created id. Change return type
14343 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14344 This can then be used to add extensions to the request.
14345 Deleted OCSP_request_new(), since most of its functionality
14346 is now in OCSP_REQUEST_new() (and the case insensitive name
14347 clash) apart from the ability to set the request name which
14348 will be added elsewhere.
14352 * Update OCSP API. Remove obsolete extensions argument from
14353 various functions. Extensions are now handled using the new
14354 OCSP extension code. New simple OCSP HTTP function which
14355 can be used to send requests and parse the response.
14359 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14360 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14361 uses the special reorder version of SET OF to sort the attributes
14362 and reorder them to match the encoded order. This resolves a long
14363 standing problem: a verify on a PKCS7 structure just after signing
14364 it used to fail because the attribute order did not match the
14365 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14366 it uses the received order. This is necessary to tolerate some broken
14367 software that does not order SET OF. This is handled by encoding
14368 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14369 to produce the required SET OF.
14373 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14374 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14375 files to get correct declarations of the ASN.1 item variables.
14379 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14380 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14381 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14382 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14383 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14384 ASN1_ITEM and no wrapper functions.
14388 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14389 replace the old function pointer based I/O routines. Change most of
14390 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14394 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14395 lines, recognize more "algorithms" that can be deselected, and make
14396 it complain about algorithm deselection that isn't recognised.
14400 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14401 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14402 to use new functions. Add NO_ASN1_OLD which can be set to remove
14403 some old style ASN1 functions: this can be used to determine if old
14404 code will still work when these eventually go away.
14408 * New extension functions for OCSP structures, these follow the
14409 same conventions as certificates and CRLs.
14413 * New function X509V3_add1_i2d(). This automatically encodes and
14414 adds an extension. Its behaviour can be customised with various
14415 flags to append, replace or delete. Various wrappers added for
14416 certificates and CRLs.
14420 * Fix to avoid calling the underlying ASN1 print routine when
14421 an extension cannot be parsed. Correct a typo in the
14422 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14426 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14427 entries for variables.
14431 * Add functionality to `apps/openssl.c` for detecting locking
14432 problems: As the program is single-threaded, all we have
14433 to do is register a locking callback using an array for
14434 storing which locks are currently held by the program.
14438 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14439 SSL_get_ex_data_X509_STORE_idx(), which is used in
14440 ssl_verify_cert_chain() and thus can be called at any time
14441 during TLS/SSL handshakes so that thread-safety is essential.
14442 Unfortunately, the ex_data design is not at all suited
14443 for multi-threaded use, so it probably should be abolished.
14447 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14449 *Broadcom, tweaked and integrated by Geoff Thorpe*
14451 * Move common extension printing code to new function
14452 X509V3_print_extensions(). Reorganise OCSP print routines and
14453 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14457 * New function X509_signature_print() to remove duplication in some
14462 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14463 set (this was treated exactly the same as SET OF previously). This
14464 is used to reorder the STACK representing the structure to match the
14465 encoding. This will be used to get round a problem where a PKCS7
14466 structure which was signed could not be verified because the STACK
14467 order did not reflect the encoded order.
14471 * Reimplement the OCSP ASN1 module using the new code.
14475 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14476 for its ASN1 operations. The old style function pointers still exist
14477 for now but they will eventually go away.
14481 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14482 completely replaces the old ASN1 functionality with a table driven
14483 encoder and decoder which interprets an ASN1_ITEM structure describing
14484 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14485 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14486 has also been converted to the new form.
14490 * Change BN_mod_exp_recp so that negative moduli are tolerated
14491 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14492 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14493 for negative moduli.
14497 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14498 of not touching the result's sign bit.
14502 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14507 * Changed the LHASH code to use prototypes for callbacks, and created
14508 macros to declare and implement thin (optionally static) functions
14509 that provide type-safety and avoid function pointer casting for the
14510 type-specific callbacks.
14514 * Added Kerberos Cipher Suites to be used with TLS, as written in
14516 *Veers Staats <staatsvr@asc.hpc.mil>,
14517 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14519 * Reformat the FAQ so the different questions and answers can be divided
14520 in sections depending on the subject.
14524 * Have the zlib compression code load ZLIB.DLL dynamically under
14529 * New function BN_mod_sqrt for computing square roots modulo a prime
14530 (using the probabilistic Tonelli-Shanks algorithm unless
14531 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14532 be handled deterministically).
14534 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14536 * Make BN_mod_inverse faster by explicitly handling small quotients
14537 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14538 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14542 * New function BN_kronecker.
14546 * Fix BN_gcd so that it works on negative inputs; the result is
14547 positive unless both parameters are zero.
14548 Previously something reasonably close to an infinite loop was
14549 possible because numbers could be growing instead of shrinking
14550 in the implementation of Euclid's algorithm.
14554 * Fix BN_is_word() and BN_is_one() macros to take into account the
14555 sign of the number in question.
14557 Fix BN_is_word(a,w) to work correctly for w == 0.
14559 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14560 because its test if the absolute value of 'a' equals 'w'.
14561 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14562 it exists mostly for use in the implementations of BN_is_zero(),
14563 BN_is_one(), and BN_is_word().
14567 * New function BN_swap.
14571 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14572 the exponentiation functions are more likely to produce reasonable
14573 results on negative inputs.
14577 * Change BN_mod_mul so that the result is always non-negative.
14578 Previously, it could be negative if one of the factors was negative;
14579 I don't think anyone really wanted that behaviour.
14583 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14584 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14585 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14586 and add new functions:
14595 BN_mod_lshift1_quick
14597 BN_mod_lshift_quick
14599 These functions always generate non-negative results.
14601 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14602 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14604 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14605 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14606 be reduced modulo `m`.
14608 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14611 The following entry accidentally appeared in the CHANGES file
14612 distributed with OpenSSL 0.9.7. The modifications described in
14613 it do *not* apply to OpenSSL 0.9.7.
14615 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14616 was actually never needed) and in BN_mul(). The removal in BN_mul()
14617 required a small change in bn_mul_part_recursive() and the addition
14618 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14619 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14620 bn_sub_words() and bn_add_words() except they take arrays with
14626 * In 'openssl passwd', verify passwords read from the terminal
14627 unless the '-salt' option is used (which usually means that
14628 verification would just waste user's time since the resulting
14629 hash is going to be compared with some given password hash)
14630 or the new '-noverify' option is used.
14632 This is an incompatible change, but it does not affect
14633 non-interactive use of 'openssl passwd' (passwords on the command
14634 line, '-stdin' option, '-in ...' option) and thus should not
14635 cause any problems.
14639 * Remove all references to RSAref, since there's no more need for it.
14643 * Make DSO load along a path given through an environment variable
14644 (SHLIB_PATH) with shl_load().
14648 * Constify the ENGINE code as a result of BIGNUM constification.
14649 Also constify the RSA code and most things related to it. In a
14650 few places, most notable in the depth of the ASN.1 code, ugly
14651 casts back to non-const were required (to be solved at a later
14656 * Make it so the openssl application has all engines loaded by default.
14660 * Constify the BIGNUM routines a little more.
14664 * Add the following functions:
14666 ENGINE_load_cswift()
14668 ENGINE_load_atalla()
14669 ENGINE_load_nuron()
14670 ENGINE_load_builtin_engines()
14672 That way, an application can itself choose if external engines that
14673 are built-in in OpenSSL shall ever be used or not. The benefit is
14674 that applications won't have to be linked with libdl or other dso
14675 libraries unless it's really needed.
14677 Changed 'openssl engine' to load all engines on demand.
14678 Changed the engine header files to avoid the duplication of some
14679 declarations (they differed!).
14683 * 'openssl engine' can now list capabilities.
14687 * Better error reporting in 'openssl engine'.
14691 * Never call load_dh_param(NULL) in s_server.
14695 * Add engine application. It can currently list engines by name and
14696 identity, and test if they are actually available.
14700 * Improve RPM specification file by forcing symbolic linking and making
14701 sure the installed documentation is also owned by root.root.
14703 *Damien Miller <djm@mindrot.org>*
14705 * Give the OpenSSL applications more possibilities to make use of
14706 keys (public as well as private) handled by engines.
14710 * Add OCSP code that comes from CertCo.
14714 * Add VMS support for the Rijndael code.
14718 * Added untested support for Nuron crypto accelerator.
14722 * Add support for external cryptographic devices. This code was
14723 previously distributed separately as the "engine" branch.
14725 *Geoff Thorpe, Richard Levitte*
14727 * Rework the filename-translation in the DSO code. It is now possible to
14728 have far greater control over how a "name" is turned into a filename
14729 depending on the operating environment and any oddities about the
14730 different shared library filenames on each system.
14734 * Support threads on FreeBSD-elf in Configure.
14738 * Fix for SHA1 assembly problem with MASM: it produces
14739 warnings about corrupt line number information when assembling
14740 with debugging information. This is caused by the overlapping
14743 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14746 NCONF_get_number() has no error checking at all. As a replacement,
14747 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14748 promoted strongly. The old NCONF_get_number is kept around for
14749 binary backward compatibility.
14750 Make it possible for methods to load from something other than a BIO,
14751 by providing a function pointer that is given a name instead of a BIO.
14752 For example, this could be used to load configuration data from an
14757 * Fix for non blocking accept BIOs. Added new I/O special reason
14758 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14759 with non blocking I/O was not possible because no retry code was
14760 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14765 * Added the beginnings of Rijndael support.
14769 * Fix for bug in DirectoryString mask setting. Add support for
14770 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14771 to allow certificate printing to more controllable, additional
14772 'certopt' option to 'x509' to allow new printing options to be
14777 * Clean old EAY MD5 hack from e_os.h.
14781 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14783 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14784 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14786 *Joe Orton, Steve Henson*
14788 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14790 * Fix additional bug revealed by the NISCC test suite:
14792 Stop bug triggering large recursion when presented with
14793 certain ASN.1 tags ([CVE-2003-0851])
14797 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14799 * Fix various bugs revealed by running the NISCC test suite:
14801 Stop out of bounds reads in the ASN1 code when presented with
14802 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14804 If verify callback ignores invalid public key errors don't try to check
14805 certificate signature with the NULL public key.
14809 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14810 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14815 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14816 extra data after the compression methods not only for TLS 1.0
14817 but also for SSL 3.0 (as required by the specification).
14819 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14821 * Change X509_certificate_type() to mark the key as exported/exportable
14822 when it's 512 *bits* long, not 512 bytes.
14826 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14828 * Countermeasure against the Klima-Pokorny-Rosa extension of
14829 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14830 a protocol version number mismatch like a decryption error
14831 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14835 * Turn on RSA blinding by default in the default implementation
14836 to avoid a timing attack. Applications that don't want it can call
14837 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14838 They would be ill-advised to do so in most cases.
14840 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14842 * Change RSA blinding code so that it works when the PRNG is not
14843 seeded (in this case, the secret RSA exponent is abused as
14844 an unpredictable seed -- if it is not unpredictable, there
14845 is no point in blinding anyway). Make RSA blinding thread-safe
14846 by remembering the creator's thread ID in rsa->blinding and
14847 having all other threads use local one-time blinding factors
14848 (this requires more computation than sharing rsa->blinding, but
14849 avoids excessive locking; and if an RSA object is not shared
14850 between threads, blinding will still be very fast).
14854 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14856 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14857 via timing by performing a MAC computation even if incorrect
14858 block cipher padding has been found. This is a countermeasure
14859 against active attacks where the attacker has to distinguish
14860 between bad padding and a MAC verification error. ([CVE-2003-0078])
14862 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14863 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14864 Martin Vuagnoux (EPFL, Ilion)*
14866 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14868 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14869 memory from its contents. This is done with a counter that will
14870 place alternating values in each byte. This can be used to solve
14871 two issues: 1) the removal of calls to memset() by highly optimizing
14872 compilers, and 2) cleansing with other values than 0, since those can
14873 be read through on certain media, for example a swap space on disk.
14877 * Bugfix: client side session caching did not work with external caching,
14878 because the session->cipher setting was not restored when reloading
14879 from the external cache. This problem was masked, when
14880 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14881 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14885 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14886 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14888 *Zeev Lieber <zeev-l@yahoo.com>*
14890 * Undo an undocumented change introduced in 0.9.6e which caused
14891 repeated calls to OpenSSL_add_all_ciphers() and
14892 OpenSSL_add_all_digests() to be ignored, even after calling
14897 * Change the default configuration reader to deal with last line not
14898 being properly terminated.
14902 * Change X509_NAME_cmp() so it applies the special rules on handling
14903 DN values that are of type PrintableString, as well as RDNs of type
14904 emailAddress where the value has the type ia5String.
14906 *stefank@valicert.com via Richard Levitte*
14908 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14909 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14910 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14911 the bitwise-OR of the two for use by the majority of applications
14912 wanting this behaviour, and update the docs. The documented
14913 behaviour and actual behaviour were inconsistent and had been
14914 changing anyway, so this is more a bug-fix than a behavioural
14917 *Geoff Thorpe, diagnosed by Nadav Har'El*
14919 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14920 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14924 * Fix initialization code race conditions in
14925 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14926 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14927 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14928 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14929 ssl2_get_cipher_by_char(),
14930 ssl3_get_cipher_by_char().
14932 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14934 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14935 the cached sessions are flushed, as the remove_cb() might use ex_data
14936 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14937 (see [openssl.org #212]).
14939 *Geoff Thorpe, Lutz Jaenicke*
14941 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14942 length, instead of the encoding length to d2i_ASN1_OBJECT.
14946 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14948 * [In 0.9.6g-engine release:]
14949 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14951 *Lynn Gazis <lgazis@rainbow.com>*
14953 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14955 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14956 and get fix the header length calculation.
14957 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14958 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14960 * Use proper error handling instead of 'assertions' in buffer
14961 overflow checks added in 0.9.6e. This prevents DoS (the
14962 assertions could call abort()).
14964 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14966 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14968 * Add various sanity checks to asn1_get_length() to reject
14969 the ASN1 length bytes if they exceed sizeof(long), will appear
14970 negative or the content length exceeds the length of the
14973 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14975 * Fix cipher selection routines: ciphers without encryption had no flags
14976 for the cipher strength set and where therefore not handled correctly
14977 by the selection routines (PR #130).
14981 * Fix EVP_dsa_sha macro.
14986 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14987 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14988 that was added in OpenSSL 0.9.6d.
14990 As the countermeasure turned out to be incompatible with some
14991 broken SSL implementations, the new option is part of SSL_OP_ALL.
14992 SSL_OP_ALL is usually employed when compatibility with weird SSL
14993 implementations is desired (e.g. '-bugs' option to 's_client' and
14994 's_server'), so the new option is automatically set in many
14999 * Changes in security patch:
15001 Changes marked "(CHATS)" were sponsored by the Defense Advanced
15002 Research Projects Agency (DARPA) and Air Force Research Laboratory,
15003 Air Force Materiel Command, USAF, under agreement number
15006 * Add various sanity checks to asn1_get_length() to reject
15007 the ASN1 length bytes if they exceed sizeof(long), will appear
15008 negative or the content length exceeds the length of the
15009 supplied buffer. ([CVE-2002-0659])
15011 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15013 * Assertions for various potential buffer overflows, not known to
15014 happen in practice.
15016 *Ben Laurie (CHATS)*
15018 * Various temporary buffers to hold ASCII versions of integers were
15019 too small for 64 bit platforms. ([CVE-2002-0655])
15020 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
15022 * Remote buffer overflow in SSL3 protocol - an attacker could
15023 supply an oversized session ID to a client. ([CVE-2002-0656])
15025 *Ben Laurie (CHATS)*
15027 * Remote buffer overflow in SSL2 protocol - an attacker could
15028 supply an oversized client master key. ([CVE-2002-0656])
15030 *Ben Laurie (CHATS)*
15032 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
15034 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
15035 encoded as NULL) with id-dsa-with-sha1.
15037 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
15039 * Check various `X509_...()` return values in `apps/req.c`.
15041 *Nils Larsch <nla@trustcenter.de>*
15043 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
15044 an end-of-file condition would erroneously be flagged, when the CRLF
15045 was just at the end of a processed block. The bug was discovered when
15046 processing data through a buffering memory BIO handing the data to a
15047 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
15048 <ptsekov@syntrex.com> and Nedelcho Stanev.
15052 * Implement a countermeasure against a vulnerability recently found
15053 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
15054 before application data chunks to avoid the use of known IVs
15055 with data potentially chosen by the attacker.
15059 * Fix length checks in ssl3_get_client_hello().
15063 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
15064 to prevent ssl3_read_internal() from incorrectly assuming that
15065 ssl3_read_bytes() found application data while handshake
15066 processing was enabled when in fact s->s3->in_read_app_data was
15067 merely automatically cleared during the initial handshake.
15069 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
15071 * Fix object definitions for Private and Enterprise: they were not
15072 recognized in their shortname (=lowercase) representation. Extend
15073 obj_dat.pl to issue an error when using undefined keywords instead
15074 of silently ignoring the problem (Svenning Sorensen
15075 <sss@sss.dnsalias.net>).
15079 * Fix DH_generate_parameters() so that it works for 'non-standard'
15080 generators, i.e. generators other than 2 and 5. (Previously, the
15081 code did not properly initialise the 'add' and 'rem' values to
15082 BN_generate_prime().)
15084 In the new general case, we do not insist that 'generator' is
15085 actually a primitive root: This requirement is rather pointless;
15086 a generator of the order-q subgroup is just as good, if not
15091 * Map new X509 verification errors to alerts. Discovered and submitted by
15092 Tom Wu <tom@arcot.com>.
15096 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
15097 returning non-zero before the data has been completely received
15098 when using non-blocking I/O.
15100 *Bodo Moeller; problem pointed out by John Hughes*
15102 * Some of the ciphers missed the strength entry (SSL_LOW etc).
15104 *Ben Laurie, Lutz Jaenicke*
15106 * Fix bug in SSL_clear(): bad sessions were not removed (found by
15107 Yoram Zahavi <YoramZ@gilian.com>).
15111 * Add information about CygWin 1.3 and on, and preserve proper
15112 configuration for the versions before that.
15114 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
15116 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
15117 check whether we deal with a copy of a session and do not delete from
15118 the cache in this case. Problem reported by "Izhar Shoshani Levi"
15119 <izhar@checkpoint.com>.
15123 * Do not store session data into the internal session cache, if it
15124 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
15125 flag is set). Proposed by Aslam <aslam@funk.com>.
15129 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
15134 * [In 0.9.6d-engine release:]
15135 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
15137 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
15139 * Add the configuration target linux-s390x.
15141 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
15143 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
15144 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
15145 variable as an indication that a ClientHello message has been
15146 received. As the flag value will be lost between multiple
15147 invocations of ssl3_accept when using non-blocking I/O, the
15148 function may not be aware that a handshake has actually taken
15149 place, thus preventing a new session from being added to the
15152 To avoid this problem, we now set s->new_session to 2 instead of
15153 using a local variable.
15155 *Lutz Jaenicke, Bodo Moeller*
15157 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15158 if the SSL_R_LENGTH_MISMATCH error is detected.
15160 *Geoff Thorpe, Bodo Moeller*
15162 * New 'shared_ldflag' column in Configure platform table.
15166 * Fix EVP_CIPHER_mode macro.
15168 *"Dan S. Camper" <dan@bti.net>*
15170 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15171 type, we must throw them away by setting rr->length to 0.
15173 *D P Chang <dpc@qualys.com>*
15175 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15177 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15178 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15179 worked incorrectly for those cases where range = `10..._2` and
15180 `3*range` is two bits longer than range.)
15184 * Only add signing time to PKCS7 structures if it is not already
15189 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15190 OBJ_ld_ce should be OBJ_id_ce.
15191 Also some ip-pda OIDs in crypto/objects/objects.txt were
15192 incorrect (cf. RFC 3039).
15194 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15196 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15197 returns early because it has nothing to do.
15199 *Andy Schneider <andy.schneider@bjss.co.uk>*
15201 * [In 0.9.6c-engine release:]
15202 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15204 *Andy Schneider <andy.schneider@bjss.co.uk>*
15206 * [In 0.9.6c-engine release:]
15207 Add support for Cryptographic Appliance's keyserver technology.
15208 (Use engine 'keyclient')
15210 *Cryptographic Appliances and Geoff Thorpe*
15212 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15213 is called via tools/c89.sh because arguments have to be
15214 rearranged (all '-L' options must appear before the first object
15217 *Richard Shapiro <rshapiro@abinitio.com>*
15219 * [In 0.9.6c-engine release:]
15220 Add support for Broadcom crypto accelerator cards, backported
15223 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15225 * [In 0.9.6c-engine release:]
15226 Add support for SureWare crypto accelerator cards from
15227 Baltimore Technologies. (Use engine 'sureware')
15229 *Baltimore Technologies and Mark Cox*
15231 * [In 0.9.6c-engine release:]
15232 Add support for crypto accelerator cards from Accelerated
15233 Encryption Processing, www.aep.ie. (Use engine 'aep')
15235 *AEP Inc. and Mark Cox*
15237 * Add a configuration entry for gcc on UnixWare.
15239 *Gary Benson <gbenson@redhat.com>*
15241 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15242 messages are stored in a single piece (fixed-length part and
15243 variable-length part combined) and fix various bugs found on the way.
15247 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15248 instead. BIO_gethostbyname() does not know what timeouts are
15249 appropriate, so entries would stay in cache even when they have
15251 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15253 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15254 faced with a pathologically small ClientHello fragment that does
15255 not contain client_version: Instead of aborting with an error,
15256 simply choose the highest available protocol version (i.e.,
15257 TLS 1.0 unless it is disabled). In practice, ClientHello
15258 messages are never sent like this, but this change gives us
15259 strictly correct behaviour at least for TLS.
15263 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15264 never resets s->method to s->ctx->method when called from within
15265 one of the SSL handshake functions.
15267 *Bodo Moeller; problem pointed out by Niko Baric*
15269 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15270 (sent using the client's version number) if client_version is
15271 smaller than the protocol version in use. Also change
15272 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15273 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15274 the client will at least see that alert.
15278 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15283 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15284 client receives HelloRequest while in a handshake.
15286 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15288 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15289 should end in 'break', not 'goto end' which circumvents various
15290 cleanups done in state SSL_ST_OK. But session related stuff
15291 must be disabled for SSL_ST_OK in the case that we just sent a
15294 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15295 before just sending a HelloRequest.
15297 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15299 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15300 reveal whether illegal block cipher padding was found or a MAC
15301 verification error occurred. (Neither SSLerr() codes nor alerts
15302 are directly visible to potential attackers, but the information
15303 may leak via logfiles.)
15305 Similar changes are not required for the SSL 2.0 implementation
15306 because the number of padding bytes is sent in clear for SSL 2.0,
15307 and the extra bytes are just ignored. However ssl/s2_pkt.c
15308 failed to verify that the purported number of padding bytes is in
15313 * Add OpenUNIX-8 support including shared libraries
15314 (Boyd Lynn Gerber <gerberb@zenez.com>).
15318 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15319 'wristwatch attack' using huge encoding parameters (cf.
15320 James H. Manger's CRYPTO 2001 paper). Note that the
15321 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15322 encoding parameters and hence was not vulnerable.
15326 * BN_sqr() bug fix.
15328 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15330 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15331 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15332 followed by modular reduction.
15334 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15336 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15337 equivalent based on BN_pseudo_rand() instead of BN_rand().
15341 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15342 This function was broken, as the check for a new client hello message
15343 to handle SGC did not allow these large messages.
15344 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15348 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15352 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15353 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15357 * Rework the configuration and shared library support for Tru64 Unix.
15358 The configuration part makes use of modern compiler features and
15359 still retains old compiler behavior for those that run older versions
15360 of the OS. The shared library support part includes a variant that
15361 uses the RPATH feature, and is available through the special
15362 configuration target "alpha-cc-rpath", which will never be selected
15365 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15367 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15368 with the same message size as in ssl3_get_certificate_request().
15369 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15370 messages might inadvertently be reject as too long.
15372 *Petr Lampa <lampa@fee.vutbr.cz>*
15374 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15378 * Modified SSL library such that the verify_callback that has been set
15379 specifically for an SSL object with SSL_set_verify() is actually being
15380 used. Before the change, a verify_callback set with this function was
15381 ignored and the verify_callback() set in the SSL_CTX at the time of
15382 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15383 to allow the necessary settings.
15387 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15388 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15389 done automatically (in contradiction to the requirements of the C
15390 standard). This made problems when used from OpenSSH.
15394 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15395 dh->length and always used
15397 BN_rand_range(priv_key, dh->p).
15399 BN_rand_range() is not necessary for Diffie-Hellman, and this
15400 specific range makes Diffie-Hellman unnecessarily inefficient if
15401 dh->length (recommended exponent length) is much smaller than the
15402 length of dh->p. We could use BN_rand_range() if the order of
15403 the subgroup was stored in the DH structure, but we only have
15408 BN_rand(priv_key, l, ...)
15410 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15417 RSA_eay_public_encrypt
15418 RSA_eay_private_decrypt
15419 RSA_eay_private_encrypt (signing)
15420 RSA_eay_public_decrypt (signature verification)
15422 (default implementations for RSA_public_encrypt,
15423 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15424 always reject numbers >= n.
15428 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15429 to synchronize access to 'locking_thread'. This is necessary on
15430 systems where access to 'locking_thread' (an 'unsigned long'
15431 variable) is not atomic.
15435 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15436 *before* setting the 'crypto_lock_rand' flag. The previous code had
15437 a race condition if 0 is a valid thread ID.
15439 *Travis Vitek <vitek@roguewave.com>*
15441 * Add support for shared libraries under Irix.
15443 *Albert Chin-A-Young <china@thewrittenword.com>*
15445 * Add configuration option to build on Linux on both big-endian and
15446 little-endian MIPS.
15448 *Ralf Baechle <ralf@uni-koblenz.de>*
15450 * Add the possibility to create shared libraries on HP-UX.
15454 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15456 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15457 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15458 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15459 PRNG state recovery was possible based on the output of
15460 one PRNG request appropriately sized to gain knowledge on
15461 'md' followed by enough consecutive 1-byte PRNG requests
15462 to traverse all of 'state'.
15464 1. When updating 'md_local' (the current thread's copy of 'md')
15465 during PRNG output generation, hash all of the previous
15466 'md_local' value, not just the half used for PRNG output.
15468 2. Make the number of bytes from 'state' included into the hash
15469 independent from the number of PRNG bytes requested.
15471 The first measure alone would be sufficient to avoid
15472 Markku-Juhani's attack. (Actually it had never occurred
15473 to me that the half of 'md_local' used for chaining was the
15474 half from which PRNG output bytes were taken -- I had always
15475 assumed that the secret half would be used.) The second
15476 measure makes sure that additional data from 'state' is never
15477 mixed into 'md_local' in small portions; this heuristically
15478 further strengthens the PRNG.
15482 * Fix crypto/bn/asm/mips3.s.
15486 * When only the key is given to "enc", the IV is undefined. Print out
15487 an error message in this case.
15491 * Handle special case when X509_NAME is empty in X509 printing routines.
15495 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15496 positive and less than q.
15500 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15501 used: it isn't thread safe and the add_lock_callback should handle
15504 *Paul Rose <Paul.Rose@bridge.com>*
15506 * Verify that incoming data obeys the block size in
15507 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15513 *Ulf Möller, Bodo Möller*
15515 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15516 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15517 when fixing the server behaviour for backwards-compatible 'client
15518 hello' messages. (Note that the attack is impractical against
15519 SSL 3.0 and TLS 1.0 anyway because length and version checking
15520 means that the probability of guessing a valid ciphertext is
15521 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15524 Before 0.9.5, the countermeasure (hide the error by generating a
15525 random 'decryption result') did not work properly because
15526 ERR_clear_error() was missing, meaning that SSL_get_error() would
15527 detect the supposedly ignored error.
15529 Both problems are now fixed.
15533 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15534 (previously it was 1024).
15538 * Fix for compatibility mode trust settings: ignore trust settings
15539 unless some valid trust or reject settings are present.
15543 * Fix for blowfish EVP: its a variable length cipher.
15547 * Fix various bugs related to DSA S/MIME verification. Handle missing
15548 parameters in DSA public key structures and return an error in the
15549 DSA routines if parameters are absent.
15553 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15554 in the current directory if neither $RANDFILE nor $HOME was set.
15555 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15556 caused some confusion to Windows users who haven't defined $HOME.
15557 Thus RAND_file_name() is changed again: e_os.h can define a
15558 DEFAULT_HOME, which will be used if $HOME is not set.
15559 For Windows, we use "C:"; on other platforms, we still require
15560 environment variables.
15562 * Move 'if (!initialized) RAND_poll()' into regions protected by
15563 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15564 having multiple threads call RAND_poll() concurrently.
15568 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15569 combination of a flag and a thread ID variable.
15570 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15571 flag), *other* threads can enter ssleay_add_bytes without obeying
15572 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15573 that they do not hold after the first thread unsets add_do_not_lock).
15577 * Change bctest again: '-x' expressions are not available in all
15578 versions of 'test'.
15582 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15584 * Fix a couple of memory leaks in PKCS7_dataDecode()
15586 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15588 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15589 the default extension for executables, if any. Also, make the perl
15590 scripts that use symlink() to test if it really exists and use "cp"
15591 if it doesn't. All this made OpenSSL compilable and installable in
15596 * Fix for asn1_GetSequence() for indefinite length constructed data.
15597 If SEQUENCE is length is indefinite just set c->slen to the total
15598 amount of data available.
15600 *Steve Henson, reported by shige@FreeBSD.org*
15602 *This change does not apply to 0.9.7.*
15604 * Change bctest to avoid here-documents inside command substitution
15605 (workaround for FreeBSD /bin/sh bug).
15606 For compatibility with Ultrix, avoid shell functions (introduced
15607 in the bctest version that searches along $PATH).
15611 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15612 with des_encrypt() defined on some operating systems, like Solaris
15617 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15618 On the Importance of Eliminating Errors in Cryptographic
15619 Computations, J. Cryptology 14 (2001) 2, 101-119,
15620 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15624 * MIPS assembler BIGNUM division bug fix.
15628 * Disabled incorrect Alpha assembler code.
15632 * Fix PKCS#7 decode routines so they correctly update the length
15633 after reading an EOC for the EXPLICIT tag.
15637 *This change does not apply to 0.9.7.*
15639 * Fix bug in PKCS#12 key generation routines. This was triggered
15640 if a 3DES key was generated with a 0 initial byte. Include
15641 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15642 (but broken) behaviour.
15646 * Enhance bctest to search for a working bc along $PATH and print
15649 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15651 * Fix memory leaks in err.c: free err_data string if necessary;
15652 don't write to the wrong index in ERR_set_error_data.
15656 * Implement ssl23_peek (analogous to ssl23_read), which previously
15661 * Replace rdtsc with `_emit` statements for VC++ version 5.
15663 *Jeremy Cooper <jeremy@baymoo.org>*
15665 * Make it possible to reuse SSLv2 sessions.
15669 * In copy_email() check for >= 0 as a return value for
15670 X509_NAME_get_index_by_NID() since 0 is a valid index.
15672 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15674 * Avoid coredump with unsupported or invalid public keys by checking if
15675 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15676 PKCS7_verify() fails with non detached data.
15680 * Don't use getenv in library functions when run as setuid/setgid.
15681 New function OPENSSL_issetugid().
15685 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15686 due to incorrect handling of multi-threading:
15688 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15690 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15692 3. Count how many times MemCheck_off() has been called so that
15693 nested use can be treated correctly. This also avoids
15694 inband-signalling in the previous code (which relied on the
15695 assumption that thread ID 0 is impossible).
15699 * Add "-rand" option also to s_client and s_server.
15703 * Fix CPU detection on Irix 6.x.
15704 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15705 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15707 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15712 *This change does not apply to 0.9.7.*
15714 * Use the cached encoding of an X509_NAME structure rather than
15715 copying it. This is apparently the reason for the libsafe "errors"
15716 but the code is actually correct.
15720 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15721 Bleichenbacher's DSA attack.
15722 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15723 to be set and top=0 forces the highest bit to be set; top=-1 is new
15724 and leaves the highest bit random.
15726 *Ulf Moeller, Bodo Moeller*
15728 * In the `NCONF_...`-based implementations for `CONF_...` queries
15729 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15730 a temporary CONF structure with the data component set to NULL
15731 (which gives segmentation faults in lh_retrieve).
15732 Instead, use NULL for the CONF pointer in CONF_get_string and
15733 CONF_get_number (which may use environment variables) and directly
15734 return NULL from CONF_get_section.
15738 * Fix potential buffer overrun for EBCDIC.
15742 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15743 keyUsage if basicConstraints absent for a CA.
15747 * Make SMIME_write_PKCS7() write mail header values with a format that
15748 is more generally accepted (no spaces before the semicolon), since
15749 some programs can't parse those values properly otherwise. Also make
15750 sure BIO's that break lines after each write do not create invalid
15755 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15756 macros previously used would not encode an empty SEQUENCE OF
15757 and break the signature.
15761 *This change does not apply to 0.9.7.*
15763 * Zero the premaster secret after deriving the master secret in
15768 * Add some EVP_add_digest_alias registrations (as found in
15769 OpenSSL_add_all_digests()) to SSL_library_init()
15770 aka OpenSSL_add_ssl_algorithms(). This provides improved
15771 compatibility with peers using X.509 certificates
15772 with unconventional AlgorithmIdentifier OIDs.
15776 * Fix for Irix with NO_ASM.
15778 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15780 * ./config script fixes.
15782 *Ulf Moeller, Richard Levitte*
15784 * Fix 'openssl passwd -1'.
15788 * Change PKCS12_key_gen_asc() so it can cope with non null
15789 terminated strings whose length is passed in the passlen
15790 parameter, for example from PEM callbacks. This was done
15791 by adding an extra length parameter to asc2uni().
15793 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15795 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15796 call failed, free the DSA structure.
15800 * Fix to uni2asc() to cope with zero length Unicode strings.
15801 These are present in some PKCS#12 files.
15805 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15806 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15807 when writing a 32767 byte record.
15809 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15811 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15812 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15814 (RSA objects have a reference count access to which is protected
15815 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15816 so they are meant to be shared between threads.)
15817 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15818 "Reddie, Steven" <Steven.Reddie@ca.com>*
15820 * Fix a deadlock in CRYPTO_mem_leaks().
15824 * Use better test patterns in bntest.
15828 * rand_win.c fix for Borland C.
15832 * BN_rshift bugfix for n == 0.
15836 * Add a 'bctest' script that checks for some known 'bc' bugs
15837 so that 'make test' does not abort just because 'bc' is broken.
15841 * Store verify_result within SSL_SESSION also for client side to
15842 avoid potential security hole. (Re-used sessions on the client side
15843 always resulted in verify_result==X509_V_OK, not using the original
15844 result of the server certificate verification.)
15848 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15849 SSL3_RT_APPLICATION_DATA, return 0.
15850 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15855 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15856 releases, have been re-implemented by renaming the previous
15857 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15858 and ssl3_read_internal, respectively, and adding 'peek' parameters
15859 to them. The new ssl[23]_{read,peek} functions are calls to
15860 ssl[23]_read_internal with the 'peek' flag set appropriately.
15861 A 'peek' parameter has also been added to ssl3_read_bytes, which
15862 does the actual work for ssl3_read_internal.
15866 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15867 the method-specific "init()" handler. Also clean up ex_data after
15868 calling the method-specific "finish()" handler. Previously, this was
15869 happening the other way round.
15873 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15874 The previous value, 12, was not always sufficient for BN_mod_exp().
15878 * Make sure that shared libraries get the internal name engine with
15879 the full version number and not just 0. This should mark the
15880 shared libraries as not backward compatible. Of course, this should
15881 be changed again when we can guarantee backward binary compatibility.
15885 * Fix typo in get_cert_by_subject() in by_dir.c
15887 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15889 * Rework the system to generate shared libraries:
15891 - Make note of the expected extension for the shared libraries and
15892 if there is a need for symbolic links from for example libcrypto.so.0
15893 to libcrypto.so.0.9.7. There is extended info in Configure for
15896 - Make as few rebuilds of the shared libraries as possible.
15898 - Still avoid linking the OpenSSL programs with the shared libraries.
15900 - When installing, install the shared libraries separately from the
15905 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15907 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15908 and not in SSL_clear because the latter is also used by the
15909 accept/connect functions; previously, the settings made by
15910 SSL_set_read_ahead would be lost during the handshake.
15912 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15914 * Correct util/mkdef.pl to be selective about disabled algorithms.
15915 Previously, it would create entries for disabled algorithms no
15920 * Added several new manual pages for SSL_* function.
15924 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15926 * In ssl23_get_client_hello, generate an error message when faced
15927 with an initial SSL 3.0/TLS record that is too small to contain the
15928 first two bytes of the ClientHello message, i.e. client_version.
15929 (Note that this is a pathologic case that probably has never happened
15930 in real life.) The previous approach was to use the version number
15931 from the record header as a substitute; but our protocol choice
15932 should not depend on that one because it is not authenticated
15933 by the Finished messages.
15937 * More robust randomness gathering functions for Windows.
15939 *Jeffrey Altman <jaltman@columbia.edu>*
15941 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15942 not set then we don't setup the error code for issuer check errors
15943 to avoid possibly overwriting other errors which the callback does
15944 handle. If an application does set the flag then we assume it knows
15945 what it is doing and can handle the new informational codes
15950 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15951 a general "ANY" type, as such it should be able to decode anything
15952 including tagged types. However it didn't check the class so it would
15953 wrongly interpret tagged types in the same way as their universal
15954 counterpart and unknown types were just rejected. Changed so that the
15955 tagged and unknown types are handled in the same way as a SEQUENCE:
15956 that is the encoding is stored intact. There is also a new type
15957 "V_ASN1_OTHER" which is used when the class is not universal, in this
15958 case we have no idea what the actual type is so we just lump them all
15963 * On VMS, stdout may very well lead to a file that is written to
15964 in a record-oriented fashion. That means that every write() will
15965 write a separate record, which will be read separately by the
15966 programs trying to read from it. This can be very confusing.
15968 The solution is to put a BIO filter in the way that will buffer
15969 text until a linefeed is reached, and then write everything a
15970 line at a time, so every record written will be an actual line,
15971 not chunks of lines and not (usually doesn't happen, but I've
15972 seen it once) several lines in one record. BIO_f_linebuffer() is
15975 Currently, it's a VMS-only method, because that's where it has
15976 been tested well enough.
15980 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15981 it can return incorrect results.
15982 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15983 but it was in 0.9.6-beta[12].)
15987 * Disable the check for content being present when verifying detached
15988 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15989 include zero length content when signing messages.
15993 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15994 BIO_ctrl (for BIO pairs).
15998 * Add DSO method for VMS.
16002 * Bug fix: Montgomery multiplication could produce results with the
16007 * Add RPM specification openssl.spec and modify it to build three
16008 packages. The default package contains applications, application
16009 documentation and run-time libraries. The devel package contains
16010 include files, static libraries and function documentation. The
16011 doc package contains the contents of the doc directory. The original
16012 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
16016 * Add a large number of documentation files for many SSL routines.
16018 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16020 * Add a configuration entry for Sony News 4.
16022 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
16024 * Don't set the two most significant bits to one when generating a
16025 random number < q in the DSA library.
16029 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
16030 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
16031 the underlying transport is blocking) if a handshake took place.
16032 (The default behaviour is needed by applications such as s_client
16033 and s_server that use select() to determine when to use SSL_read;
16034 but for applications that know in advance when to expect data, it
16035 just makes things more complicated.)
16039 * Add RAND_egd_bytes(), which gives control over the number of bytes read
16044 * Add a few more EBCDIC conditionals that make `req` and `x509`
16045 work better on such systems.
16047 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16049 * Add two demo programs for PKCS12_parse() and PKCS12_create().
16050 Update PKCS12_parse() so it copies the friendlyName and the
16051 keyid to the certificates aux info.
16055 * Fix bug in PKCS7_verify() which caused an infinite loop
16056 if there was more than one signature.
16058 *Sven Uszpelkat <su@celocom.de>*
16060 * Major change in util/mkdef.pl to include extra information
16061 about each symbol, as well as presenting variables as well
16062 as functions. This change means that there's n more need
16063 to rebuild the .num files when some algorithms are excluded.
16067 * Allow the verify time to be set by an application,
16068 rather than always using the current time.
16072 * Phase 2 verify code reorganisation. The certificate
16073 verify code now looks up an issuer certificate by a
16074 number of criteria: subject name, authority key id
16075 and key usage. It also verifies self signed certificates
16076 by the same criteria. The main comparison function is
16077 X509_check_issued() which performs these checks.
16079 Lot of changes were necessary in order to support this
16080 without completely rewriting the lookup code.
16082 Authority and subject key identifier are now cached.
16084 The LHASH 'certs' is X509_STORE has now been replaced
16085 by a STACK_OF(X509_OBJECT). This is mainly because an
16086 LHASH can't store or retrieve multiple objects with
16087 the same hash value.
16089 As a result various functions (which were all internal
16090 use only) have changed to handle the new X509_STORE
16091 structure. This will break anything that messed round
16092 with X509_STORE internally.
16094 The functions X509_STORE_add_cert() now checks for an
16095 exact match, rather than just subject name.
16097 The X509_STORE API doesn't directly support the retrieval
16098 of multiple certificates matching a given criteria, however
16099 this can be worked round by performing a lookup first
16100 (which will fill the cache with candidate certificates)
16101 and then examining the cache for matches. This is probably
16102 the best we can do without throwing out X509_LOOKUP
16103 entirely (maybe later...).
16105 The X509_VERIFY_CTX structure has been enhanced considerably.
16107 All certificate lookup operations now go via a get_issuer()
16108 callback. Although this currently uses an X509_STORE it
16109 can be replaced by custom lookups. This is a simple way
16110 to bypass the X509_STORE hackery necessary to make this
16111 work and makes it possible to use more efficient techniques
16112 in future. A very simple version which uses a simple
16113 STACK for its trusted certificate store is also provided
16114 using X509_STORE_CTX_trusted_stack().
16116 The verify_cb() and verify() callbacks now have equivalents
16117 in the X509_STORE_CTX structure.
16119 X509_STORE_CTX also has a 'flags' field which can be used
16120 to customise the verify behaviour.
16124 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
16125 excludes S/MIME capabilities.
16129 * When a certificate request is read in keep a copy of the
16130 original encoding of the signed data and use it when outputting
16131 again. Signatures then use the original encoding rather than
16132 a decoded, encoded version which may cause problems if the
16133 request is improperly encoded.
16137 * For consistency with other BIO_puts implementations, call
16138 buffer_write(b, ...) directly in buffer_puts instead of calling
16141 In BIO_puts, increment b->num_write as in BIO_write.
16143 *Peter.Sylvester@EdelWeb.fr*
16145 * Fix BN_mul_word for the case where the word is 0. (We have to use
16146 BN_zero, we may not return a BIGNUM with an array consisting of
16147 words set to zero.)
16151 * Avoid calling abort() from within the library when problems are
16152 detected, except if preprocessor symbols have been defined
16153 (such as REF_CHECK, BN_DEBUG etc.).
16157 * New openssl application 'rsautl'. This utility can be
16158 used for low-level RSA operations. DER public key
16159 BIO/fp routines also added.
16163 * New Configure entry and patches for compiling on QNX 4.
16165 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16167 * A demo state-machine implementation was sponsored by
16168 Nuron (<http://www.nuron.com/>) and is now available in
16169 demos/state_machine.
16173 * New options added to the 'dgst' utility for signature
16174 generation and verification.
16178 * Unrecognized PKCS#7 content types are now handled via a
16179 catch all ASN1_TYPE structure. This allows unsupported
16180 types to be stored as a "blob" and an application can
16181 encode and decode it manually.
16185 * Fix various signed/unsigned issues to make a_strex.c
16186 compile under VC++.
16188 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16190 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16191 length if passed a buffer. ASN1_INTEGER_to_BN failed
16192 if passed a NULL BN and its argument was negative.
16194 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16196 * Modification to PKCS#7 encoding routines to output definite
16197 length encoding. Since currently the whole structures are in
16198 memory there's not real point in using indefinite length
16199 constructed encoding. However if OpenSSL is compiled with
16200 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16204 * Added BIO_vprintf() and BIO_vsnprintf().
16208 * Added more prefixes to parse for in the strings written
16209 through a logging bio, to cover all the levels that are available
16210 through syslog. The prefixes are now:
16212 PANIC, EMERG, EMR => LOG_EMERG
16213 ALERT, ALR => LOG_ALERT
16214 CRIT, CRI => LOG_CRIT
16215 ERROR, ERR => LOG_ERR
16216 WARNING, WARN, WAR => LOG_WARNING
16217 NOTICE, NOTE, NOT => LOG_NOTICE
16218 INFO, INF => LOG_INFO
16219 DEBUG, DBG => LOG_DEBUG
16221 and as before, if none of those prefixes are present at the
16222 beginning of the string, LOG_ERR is chosen.
16224 On Win32, the `LOG_*` levels are mapped according to this:
16226 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16227 LOG_WARNING => EVENTLOG_WARNING_TYPE
16228 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16232 * Made it possible to reconfigure with just the configuration
16233 argument "reconf" or "reconfigure". The command line arguments
16234 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16235 and are retrieved from there when reconfiguring.
16241 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16243 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16247 * The obj_dat.pl script was messing up the sorting of object
16248 names. The reason was that it compared the quoted version
16249 of strings as a result "OCSP" > "OCSP Signing" because
16250 " > SPACE. Changed script to store unquoted versions of
16251 names and add quotes on output. It was also omitting some
16252 names from the lookup table if they were given a default
16253 value (that is if SN is missing it is given the same
16254 value as LN and vice versa), these are now added on the
16255 grounds that if an object has a name we should be able to
16256 look it up. Finally added warning output when duplicate
16257 short or long names are found.
16261 * Changes needed for Tandem NSK.
16263 *Scott Uroff <scott@xypro.com>*
16265 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16266 RSA_padding_check_SSLv23(), special padding was never detected
16267 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16268 version rollback attacks was not effective.
16270 In s23_clnt.c, don't use special rollback-attack detection padding
16271 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16272 client; similarly, in s23_srvr.c, don't do the rollback check if
16273 SSL 2.0 is the only protocol enabled in the server.
16277 * Make it possible to get hexdumps of unprintable data with 'openssl
16278 asn1parse'. By implication, the functions ASN1_parse_dump() and
16279 BIO_dump_indent() are added.
16283 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16284 these print out strings and name structures based on various
16285 flags including RFC2253 support and proper handling of
16286 multibyte characters. Added options to the 'x509' utility
16287 to allow the various flags to be set.
16291 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16292 Also change the functions X509_cmp_current_time() and
16293 X509_gmtime_adj() work with an ASN1_TIME structure,
16294 this will enable certificates using GeneralizedTime in validity
16295 dates to be checked.
16299 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16300 negative public key encodings) on by default,
16301 NO_NEG_PUBKEY_BUG can be set to disable it.
16305 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16306 content octets. An i2c_ASN1_OBJECT is unnecessary because
16307 the encoding can be trivially obtained from the structure.
16311 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16312 not read locks (`CRYPTO_r_[un]lock`).
16316 * A first attempt at creating official support for shared
16317 libraries through configuration. I've kept it so the
16318 default is static libraries only, and the OpenSSL programs
16319 are always statically linked for now, but there are
16320 preparations for dynamic linking in place.
16321 This has been tested on Linux and Tru64.
16325 * Randomness polling function for Win9x, as described in:
16326 Peter Gutmann, Software Generation of Practically Strong
16331 * Fix so PRNG is seeded in req if using an already existing
16336 * New options to smime application. -inform and -outform
16337 allow alternative formats for the S/MIME message including
16338 PEM and DER. The -content option allows the content to be
16339 specified separately. This should allow things like Netscape
16340 form signing output easier to verify.
16344 * Fix the ASN1 encoding of tags using the 'long form'.
16348 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16349 STRING types. These convert content octets to and from the
16350 underlying type. The actual tag and length octets are
16351 already assumed to have been read in and checked. These
16352 are needed because all other string types have virtually
16353 identical handling apart from the tag. By having versions
16354 of the ASN1 functions that just operate on content octets
16355 IMPLICIT tagging can be handled properly. It also allows
16356 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16357 and ASN1_INTEGER are identical apart from the tag.
16361 * Change the handling of OID objects as follows:
16363 - New object identifiers are inserted in objects.txt, following
16364 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16365 - objects.pl is used to process obj_mac.num and create a new
16367 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16370 This is currently kind of a hack, and the perl code in objects.pl
16371 isn't very elegant, but it works as I intended. The simplest way
16372 to check that it worked correctly is to look in obj_dat.h and
16373 check the array nid_objs and make sure the objects haven't moved
16374 around (this is important!). Additions are OK, as well as
16375 consistent name changes.
16379 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16383 * Addition of the command line parameter '-rand file' to 'openssl req'.
16384 The given file adds to whatever has already been seeded into the
16385 random pool through the RANDFILE configuration file option or
16386 environment variable, or the default random state file.
16390 * mkstack.pl now sorts each macro group into lexical order.
16391 Previously the output order depended on the order the files
16392 appeared in the directory, resulting in needless rewriting
16397 * Patches to make OpenSSL compile under Win32 again. Mostly
16398 work arounds for the VC++ problem that it treats func() as
16399 func(void). Also stripped out the parts of mkdef.pl that
16400 added extra typesafe functions: these no longer exist.
16404 * Reorganisation of the stack code. The macros are now all
16405 collected in safestack.h . Each macro is defined in terms of
16406 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16407 DEBUG_SAFESTACK is now handled in terms of function casts,
16408 this has the advantage of retaining type safety without the
16409 use of additional functions. If DEBUG_SAFESTACK is not defined
16410 then the non typesafe macros are used instead. Also modified the
16411 mkstack.pl script to handle the new form. Needs testing to see
16412 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16413 the default if no major problems. Similar behaviour for ASN1_SET_OF
16414 and PKCS12_STACK_OF.
16418 * When some versions of IIS use the 'NET' form of private key the
16419 key derivation algorithm is different. Normally MD5(password) is
16420 used as a 128 bit RC4 key. In the modified case
16421 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16422 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16423 as the old Netscape_RSA functions except they have an additional
16424 'sgckey' parameter which uses the modified algorithm. Also added
16425 an -sgckey command line option to the rsa utility. Thanks to
16426 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16427 algorithm to openssl-dev.
16431 * The evp_local.h macros were using 'c.##kname' which resulted in
16432 invalid expansion on some systems (SCO 5.0.5 for example).
16433 Corrected to 'c.kname'.
16435 *Phillip Porch <root@theporch.com>*
16437 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16438 a STACK of email addresses from a certificate or request, these look
16439 in the subject name and the subject alternative name extensions and
16440 omit any duplicate addresses.
16444 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16445 This makes DSA verification about 2 % faster.
16449 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16450 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16451 plus overhead for 1024 bit moduli).
16452 This makes exponentiations about 0.5 % faster for 1024 bit
16453 exponents (as measured by "openssl speed rsa2048").
16457 * Rename memory handling macros to avoid conflicts with other
16459 Malloc => OPENSSL_malloc
16460 Malloc_locked => OPENSSL_malloc_locked
16461 Realloc => OPENSSL_realloc
16462 Free => OPENSSL_free
16466 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16467 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16471 * CygWin32 support.
16473 *John Jarvie <jjarvie@newsguy.com>*
16475 * The type-safe stack code has been rejigged. It is now only compiled
16476 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16477 by default all type-specific stack functions are "#define"d back to
16478 standard stack functions. This results in more streamlined output
16479 but retains the type-safety checking possibilities of the original
16484 * The STACK code has been cleaned up, and certain type declarations
16485 that didn't make a lot of sense have been brought in line. This has
16486 also involved a cleanup of sorts in safestack.h to more correctly
16487 map type-safe stack functions onto their plain stack counterparts.
16488 This work has also resulted in a variety of "const"ifications of
16489 lots of the code, especially `_cmp` operations which should normally
16490 be prototyped with "const" parameters anyway.
16494 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16495 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16496 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16497 where all of 'md' is used each time the PRNG is used, but 'state'
16498 is used only indexed by a cyclic counter. As entropy may not be
16499 well distributed from the beginning, 'md' is important as a
16500 chaining variable. However, the output function chains only half
16501 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16502 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16503 in all of 'state' being rewritten, with the new values depending
16504 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16508 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16509 the handshake is continued after ssl_verify_cert_chain();
16510 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16511 can lead to 'unexplainable' connection aborts later.
16513 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16515 * Major EVP API cipher revision.
16516 Add hooks for extra EVP features. This allows various cipher
16517 parameters to be set in the EVP interface. Support added for variable
16518 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16519 setting of RC2 and RC5 parameters.
16521 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16524 Remove lots of duplicated code from the EVP library. For example *every*
16525 cipher init() function handles the 'iv' in the same way according to the
16526 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16527 for CFB and OFB modes they zero ctx->num.
16529 New functionality allows removal of S/MIME code RC2 hack.
16531 Most of the routines have the same form and so can be declared in terms
16534 By shifting this to the top level EVP_CipherInit() it can be removed from
16535 all individual ciphers. If the cipher wants to handle IVs or keys
16536 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16539 Change lots of functions like EVP_EncryptUpdate() to now return a
16540 value: although software versions of the algorithms cannot fail
16541 any installed hardware versions can.
16545 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16546 this option is set, tolerate broken clients that send the negotiated
16547 protocol version number instead of the requested protocol version
16552 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16553 i.e. non-zero for export ciphersuites, zero otherwise.
16554 Previous versions had this flag inverted, inconsistent with
16555 rsa_tmp_cb (..._TMP_RSA_CB).
16557 *Bodo Moeller; problem reported by Amit Chopra*
16559 * Add missing DSA library text string. Work around for some IIS
16560 key files with invalid SEQUENCE encoding.
16564 * Add a document (doc/standards.txt) that list all kinds of standards
16565 and so on that are implemented in OpenSSL.
16569 * Enhance c_rehash script. Old version would mishandle certificates
16570 with the same subject name hash and wouldn't handle CRLs at all.
16571 Added -fingerprint option to crl utility, to support new c_rehash
16576 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16580 * Fix for SSL server purpose checking. Server checking was
16581 rejecting certificates which had extended key usage present
16582 but no ssl client purpose.
16584 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16586 * Make PKCS#12 code work with no password. The PKCS#12 spec
16587 is a little unclear about how a blank password is handled.
16588 Since the password in encoded as a BMPString with terminating
16589 double NULL a zero length password would end up as just the
16590 double NULL. However no password at all is different and is
16591 handled differently in the PKCS#12 key generation code. NS
16592 treats a blank password as zero length. MSIE treats it as no
16593 password on export: but it will try both on import. We now do
16594 the same: PKCS12_parse() tries zero length and no password if
16595 the password is set to "" or NULL (NULL is now a valid password:
16596 it wasn't before) as does the pkcs12 application.
16600 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16601 perror when PEM_read_bio_X509_REQ fails, the error message must
16602 be obtained from the error queue.
16606 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16607 it in ERR_remove_state if appropriate, and change ERR_get_state
16608 accordingly to avoid race conditions (this is necessary because
16609 thread_hash is no longer constant once set).
16613 * Bugfix for linux-elf makefile.one.
16617 * RSA_get_default_method() will now cause a default
16618 RSA_METHOD to be chosen if one doesn't exist already.
16619 Previously this was only set during a call to RSA_new()
16620 or RSA_new_method(NULL) meaning it was possible for
16621 RSA_get_default_method() to return NULL.
16625 * Added native name translation to the existing DSO code
16626 that will convert (if the flag to do so is set) filenames
16627 that are sufficiently small and have no path information
16628 into a canonical native form. Eg. "blah" converted to
16629 "libblah.so" or "blah.dll" etc.
16633 * New function ERR_error_string_n(e, buf, len) which is like
16634 ERR_error_string(e, buf), but writes at most 'len' bytes
16635 including the 0 terminator. For ERR_error_string_n, 'buf'
16638 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16640 * CONF library reworked to become more general. A new CONF
16641 configuration file reader "class" is implemented as well as a
16642 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16643 old `CONF_*` functions are still there, but are reimplemented to
16644 work in terms of the new functions. Also, a set of functions
16645 to handle the internal storage of the configuration data is
16646 provided to make it easier to write new configuration file
16647 reader "classes" (I can definitely see something reading a
16648 configuration file in XML format, for example), called `_CONF_*`,
16649 or "the configuration storage API"...
16651 The new configuration file reading functions are:
16653 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16654 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16656 NCONF_default, NCONF_WIN32
16658 NCONF_dump_fp, NCONF_dump_bio
16660 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16661 NCONF_new creates a new CONF object. This works in the same way
16662 as other interfaces in OpenSSL, like the BIO interface.
16663 `NCONF_dump_*` dump the internal storage of the configuration file,
16664 which is useful for debugging. All other functions take the same
16665 arguments as the old `CONF_*` functions with the exception of the
16666 first that must be a `CONF *` instead of a `LHASH *`.
16668 To make it easier to use the new classes with the old `CONF_*` functions,
16669 the function CONF_set_default_method is provided.
16673 * Add '-tls1' option to 'openssl ciphers', which was already
16674 mentioned in the documentation but had not been implemented.
16675 (This option is not yet really useful because even the additional
16676 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16680 * Initial DSO code added into libcrypto for letting OpenSSL (and
16681 OpenSSL-based applications) load shared libraries and bind to
16682 them in a portable way.
16684 *Geoff Thorpe, with contributions from Richard Levitte*
16686 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16688 * Make sure _lrotl and _lrotr are only used with MSVC.
16690 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16691 (the default implementation of RAND_status).
16693 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16694 to '-clrext' (= clear extensions), as intended and documented.
16695 *Bodo Moeller; inconsistency pointed out by Michael Attili
16696 <attili@amaxo.com>*
16698 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16699 was larger than the MD block size.
16701 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16703 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16704 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16705 using the passed key: if the passed key was a private key the result
16706 of X509_print(), for example, would be to print out all the private key
16711 * des_quad_cksum() byte order bug fix.
16712 *Ulf Möller, using the problem description in krb4-0.9.7, where
16713 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16715 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16718 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16720 * For easily testing in shell scripts whether some command
16721 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16722 returns with exit code 0 iff no command of the given name is available.
16723 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16724 the output goes to stdout and nothing is printed to stderr.
16725 Additional arguments are always ignored.
16727 Since for each cipher there is a command of the same name,
16728 the 'no-cipher' compilation switches can be tested this way.
16730 ('openssl no-XXX' is not able to detect pseudo-commands such
16731 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16735 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16739 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16740 is set; it will be thrown away anyway because each handshake creates
16742 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16743 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16744 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16745 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16749 * New s_client option -ign_eof: EOF at stdin is ignored, and
16750 'Q' and 'R' lose their special meanings (quit/renegotiate).
16751 This is part of what -quiet does; unlike -quiet, -ign_eof
16752 does not suppress any output.
16756 * Add compatibility options to the purpose and trust code. The
16757 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16758 accepts a certificate or CA, this was the previous behaviour,
16759 with all the associated security issues.
16761 X509_TRUST_COMPAT is the old trust behaviour: only and
16762 automatically trust self signed roots in certificate store. A
16763 new trust setting X509_TRUST_DEFAULT is used to specify that
16764 a purpose has no associated trust setting and it should instead
16765 use the value in the default purpose.
16769 * Fix the PKCS#8 DSA private key code so it decodes keys again
16770 and fix a memory leak.
16774 * In util/mkerr.pl (which implements 'make errors'), preserve
16775 reason strings from the previous version of the .c file, as
16776 the default to have only downcase letters (and digits) in
16777 automatically generated reasons codes is not always appropriate.
16781 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16782 using strerror. Previously, ERR_reason_error_string() returned
16783 library names as reason strings for SYSerr; but SYSerr is a special
16784 case where small numbers are errno values, not library numbers.
16788 * Add '-dsaparam' option to 'openssl dhparam' application. This
16789 converts DSA parameters into DH parameters. (When creating parameters,
16790 DSA_generate_parameters is used.)
16794 * Include 'length' (recommended exponent length) in C code generated
16795 by 'openssl dhparam -C'.
16799 * The second argument to set_label in perlasm was already being used
16800 so couldn't be used as a "file scope" flag. Moved to third argument
16805 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16806 instead of RAND_bytes for encryption IVs and salts.
16810 * Include RAND_status() into RAND_METHOD instead of implementing
16811 it only for md_rand.c Otherwise replacing the PRNG by calling
16812 RAND_set_rand_method would be impossible.
16816 * Don't let DSA_generate_key() enter an infinite loop if the random
16817 number generation fails.
16821 * New 'rand' application for creating pseudo-random output.
16825 * Added configuration support for Linux/IA64
16827 *Rolf Haberrecker <rolf@suse.de>*
16829 * Assembler module support for Mingw32.
16833 * Shared library support for HPUX (in shlib/).
16835 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16837 * Shared library support for Solaris gcc.
16839 *Lutz Behnke <behnke@trustcenter.de>*
16841 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16843 * PKCS7_encrypt() was adding text MIME headers twice because they
16844 were added manually and by SMIME_crlf_copy().
16848 * In bntest.c don't call BN_rand with zero bits argument.
16850 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16852 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16853 case was implemented. This caused BN_div_recp() to fail occasionally.
16857 * Add an optional second argument to the set_label() in the perl
16858 assembly language builder. If this argument exists and is set
16859 to 1 it signals that the assembler should use a symbol whose
16860 scope is the entire file, not just the current function. This
16861 is needed with MASM which uses the format label:: for this scope.
16863 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16865 * Change the ASN1 types so they are typedefs by default. Before
16866 almost all types were #define'd to ASN1_STRING which was causing
16867 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16872 * Change names of new functions to the new get1/get0 naming
16873 convention: After 'get1', the caller owns a reference count
16874 and has to call `..._free`; 'get0' returns a pointer to some
16875 data structure without incrementing reference counters.
16876 (Some of the existing 'get' functions increment a reference
16877 counter, some don't.)
16878 Similarly, 'set1' and 'add1' functions increase reference
16879 counters or duplicate objects.
16883 * Allow for the possibility of temp RSA key generation failure:
16884 the code used to assume it always worked and crashed on failure.
16888 * Fix potential buffer overrun problem in BIO_printf().
16889 *Ulf Möller, using public domain code by Patrick Powell; problem
16890 pointed out by David Sacerdote <das33@cornell.edu>*
16892 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16893 RAND_egd() and RAND_status(). In the command line application,
16894 the EGD socket can be specified like a seed file using RANDFILE
16899 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16900 Some CAs (e.g. Verisign) distribute certificates in this form.
16904 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16905 list to exclude them. This means that no special compilation option
16906 is needed to use anonymous DH: it just needs to be included in the
16911 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16912 EVP_MD_type. The old functionality is available in a new macro called
16913 EVP_MD_md(). Change code that uses it and update docs.
16917 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16918 where the `void *` argument is replaced by a function pointer argument.
16919 Previously `void *` was abused to point to functions, which works on
16920 many platforms, but is not correct. As these functions are usually
16921 called by macros defined in OpenSSL header files, most source code
16922 should work without changes.
16926 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16927 sections with information on -D... compiler switches used for
16928 compiling the library so that applications can see them. To enable
16929 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16930 must be defined. E.g.,
16931 #define OPENSSL_ALGORITHM_DEFINES
16932 #include <openssl/opensslconf.h>
16933 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16935 *Richard Levitte, Ulf and Bodo Möller*
16937 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16942 * Change the 'other' type in certificate aux info to a STACK_OF
16943 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16944 the required ASN1 format: arbitrary types determined by an OID.
16948 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16949 argument to 'req'. This is not because the function is newer or
16950 better than others it just uses the work 'NEW' in the certificate
16951 request header lines. Some software needs this.
16955 * Reorganise password command line arguments: now passwords can be
16956 obtained from various sources. Delete the PEM_cb function and make
16957 it the default behaviour: i.e. if the callback is NULL and the
16958 usrdata argument is not NULL interpret it as a null terminated pass
16959 phrase. If usrdata and the callback are NULL then the pass phrase
16960 is prompted for as usual.
16964 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16965 the support is automatically enabled. The resulting binaries will
16966 autodetect the card and use it if present.
16968 *Ben Laurie and Compaq Inc.*
16970 * Work around for Netscape hang bug. This sends certificate request
16971 and server done in one record. Since this is perfectly legal in the
16972 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16973 the bugs/SSLv3 entry for more info.
16977 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16981 * Add -rand argument to smime and pkcs12 applications and read/write
16986 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16990 * Add command line password options to the remaining applications.
16994 * Bug fix for BN_div_recp() for numerators with an even number of
16999 * More tests in bntest.c, and changed test_bn output.
17003 * ./config recognizes MacOS X now.
17007 * Bug fix for BN_div() when the first words of num and divisor are
17008 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
17012 * Add support for various broken PKCS#8 formats, and command line
17013 options to produce them.
17017 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
17018 get temporary BIGNUMs from a BN_CTX.
17022 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
17027 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
17028 include a #define from the old name to the new. The original intent
17029 was that statically linked binaries could for example just call
17030 SSLeay_add_all_ciphers() to just add ciphers to the table and not
17031 link with digests. This never worked because SSLeay_add_all_digests()
17032 and SSLeay_add_all_ciphers() were in the same source file so calling
17033 one would link with the other. They are now in separate source files.
17037 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
17041 * Use a less unusual form of the Miller-Rabin primality test (it used
17042 a binary algorithm for exponentiation integrated into the Miller-Rabin
17043 loop, our standard modexp algorithms are faster).
17047 * Support for the EBCDIC character set completed.
17049 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
17051 * Source code cleanups: use const where appropriate, eliminate casts,
17052 use `void *` instead of `char *` in lhash.
17056 * Bugfix: ssl3_send_server_key_exchange was not restartable
17057 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
17058 this the server could overwrite ephemeral keys that the client
17063 * Turn DSA_is_prime into a macro that calls BN_is_prime,
17064 using 50 iterations of the Rabin-Miller test.
17066 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
17067 iterations of the Rabin-Miller test as required by the appendix
17068 to FIPS PUB 186[-1]) instead of DSA_is_prime.
17069 As BN_is_prime_fasttest includes trial division, DSA parameter
17070 generation becomes much faster.
17072 This implies a change for the callback functions in DSA_is_prime
17073 and DSA_generate_parameters: The callback function is called once
17074 for each positive witness in the Rabin-Miller test, not just
17075 occasionally in the inner loop; and the parameters to the
17076 callback function now provide an iteration count for the outer
17077 loop rather than for the current invocation of the inner loop.
17078 DSA_generate_parameters additionally can call the callback
17079 function with an 'iteration count' of -1, meaning that a
17080 candidate has passed the trial division test (when q is generated
17081 from an application-provided seed, trial division is skipped).
17085 * New function BN_is_prime_fasttest that optionally does trial
17086 division before starting the Rabin-Miller test and has
17087 an additional BN_CTX * argument (whereas BN_is_prime always
17088 has to allocate at least one BN_CTX).
17089 'callback(1, -1, cb_arg)' is called when a number has passed the
17090 trial division stage.
17094 * Fix for bug in CRL encoding. The validity dates weren't being handled
17099 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
17103 * New function BN_pseudo_rand().
17107 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
17108 bignum version of BN_from_montgomery() with the working code from
17109 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
17114 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
17115 made it impossible to use the same SSL_SESSION data structure in
17116 SSL2 clients in multiple threads.
17120 * The return value of RAND_load_file() no longer counts bytes obtained
17121 by stat(). RAND_load_file(..., -1) is new and uses the complete file
17122 to seed the PRNG (previously an explicit byte count was required).
17124 *Ulf Möller, Bodo Möller*
17126 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
17127 used `char *` instead of `void *` and had casts all over the place.
17131 * Make BN_generate_prime() return NULL on error if ret!=NULL.
17135 * Retain source code compatibility for BN_prime_checks macro:
17136 BN_is_prime(..., BN_prime_checks, ...) now uses
17137 BN_prime_checks_for_size to determine the appropriate number of
17138 Rabin-Miller iterations.
17142 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
17143 DH_CHECK_P_NOT_SAFE_PRIME.
17144 (Check if this is true? OpenPGP calls them "strong".)
17148 * Merge the functionality of "dh" and "gendh" programs into a new program
17149 "dhparam". The old programs are retained for now but will handle DH keys
17150 (instead of parameters) in future.
17154 * Make the ciphers, s_server and s_client programs check the return values
17155 when a new cipher list is set.
17159 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17160 ciphers. Before when the 56bit ciphers were enabled the sorting was
17163 The syntax for the cipher sorting has been extended to support sorting by
17164 cipher-strength (using the strength_bits hard coded in the tables).
17165 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17167 Fix a bug in the cipher-command parser: when supplying a cipher command
17168 string with an "undefined" symbol (neither command nor alphanumeric
17169 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17170 an error is flagged.
17172 Due to the strength-sorting extension, the code of the
17173 ssl_create_cipher_list() function was completely rearranged. I hope that
17174 the readability was also increased :-)
17176 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17178 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17179 for the first serial number and places 2 in the serial number file. This
17180 avoids problems when the root CA is created with serial number zero and
17181 the first user certificate has the same issuer name and serial number
17186 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17187 the new code. Add documentation for this stuff.
17191 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17192 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17193 structures and behave in an analogous way to the X509v3 functions:
17194 they shouldn't be called directly but wrapper functions should be used
17197 So we also now have some wrapper functions that call the X509at functions
17198 when passed certificate requests. (TO DO: similar things can be done with
17199 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17200 things. Some of these need some d2i or i2d and print functionality
17201 because they handle more complex structures.)
17205 * Add missing #ifndefs that caused missing symbols when building libssl
17206 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17207 NO_RSA in `ssl/s2*.c`.
17209 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17211 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17212 has a return value which indicates the quality of the random data
17213 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17214 error queue. New function RAND_pseudo_bytes() generates output that is
17215 guaranteed to be unique but not unpredictable. RAND_add is like
17216 RAND_seed, but takes an extra argument for an entropy estimate
17217 (RAND_seed always assumes full entropy).
17221 * Do more iterations of Rabin-Miller probable prime test (specifically,
17222 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17223 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17224 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17225 false-positive rate of at most 2^-80 for random input.
17229 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17233 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17234 in the 0.9.5 release), this returns the chain
17235 from an X509_CTX structure with a dup of the stack and all
17236 the X509 reference counts upped: so the stack will exist
17237 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17240 Also make SSL_SESSION_print() print out the verify return
17245 * Add manpage for the pkcs12 command. Also change the default
17246 behaviour so MAC iteration counts are used unless the new
17247 -nomaciter option is used. This improves file security and
17248 only older versions of MSIE (4.0 for example) need it.
17252 * Honor the no-xxx Configure options when creating .DEF files.
17256 * Add PKCS#10 attributes to field table: challengePassword,
17257 unstructuredName and unstructuredAddress. These are taken from
17258 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17259 international characters are used.
17261 More changes to X509_ATTRIBUTE code: allow the setting of types
17262 based on strings. Remove the 'loc' parameter when adding
17263 attributes because these will be a SET OF encoding which is sorted
17268 * Initial changes to the 'req' utility to allow request generation
17269 automation. This will allow an application to just generate a template
17270 file containing all the field values and have req construct the
17273 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17274 used all over the place including certificate requests and PKCS#7
17275 structures. They are currently handled manually where necessary with
17276 some primitive wrappers for PKCS#7. The new functions behave in a
17277 manner analogous to the X509 extension functions: they allow
17278 attributes to be looked up by NID and added.
17280 Later something similar to the X509V3 code would be desirable to
17281 automatically handle the encoding, decoding and printing of the
17282 more complex types. The string types like challengePassword can
17283 be handled by the string table functions.
17285 Also modified the multi byte string table handling. Now there is
17286 a 'global mask' which masks out certain types. The table itself
17287 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17288 is useful when for example there is only one permissible type
17289 (as in countryName) and using the mask might result in no valid
17294 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17295 SSL_get_peer_finished to allow applications to obtain the latest
17296 Finished messages sent to the peer or expected from the peer,
17297 respectively. (SSL_get_peer_finished is usually the Finished message
17298 actually received from the peer, otherwise the protocol will be aborted.)
17300 As the Finished message are message digests of the complete handshake
17301 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17302 be used for external authentication procedures when the authentication
17303 provided by SSL/TLS is not desired or is not enough.
17307 * Enhanced support for Alpha Linux is added. Now ./config checks if
17308 the host supports BWX extension and if Compaq C is present on the
17309 $PATH. Just exploiting of the BWX extension results in 20-30%
17310 performance kick for some algorithms, e.g. DES and RC4 to mention
17311 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17316 * Add support for MS "fast SGC". This is arguably a violation of the
17317 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17318 weak crypto and after checking the certificate is SGC a second one
17319 with strong crypto. MS SGC stops the first handshake after receiving
17320 the server certificate message and sends a second client hello. Since
17321 a server will typically do all the time consuming operations before
17322 expecting any further messages from the client (server key exchange
17323 is the most expensive) there is little difference between the two.
17325 To get OpenSSL to support MS SGC we have to permit a second client
17326 hello message after we have sent server done. In addition we have to
17327 reset the MAC if we do get this second client hello.
17331 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17332 if a DER encoded private key is RSA or DSA traditional format. Changed
17333 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17334 format DER encoded private key. Newer code should use PKCS#8 format which
17335 has the key type encoded in the ASN1 structure. Added DER private key
17336 support to pkcs8 application.
17340 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17341 ciphersuites has been selected (as required by the SSL 3/TLS 1
17342 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17343 is set, we interpret this as a request to violate the specification
17344 (the worst that can happen is a handshake failure, and 'correct'
17345 behaviour would result in a handshake failure anyway).
17349 * In SSL_CTX_add_session, take into account that there might be multiple
17350 SSL_SESSION structures with the same session ID (e.g. when two threads
17351 concurrently obtain them from an external cache).
17352 The internal cache can handle only one SSL_SESSION with a given ID,
17353 so if there's a conflict, we now throw out the old one to achieve
17358 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17359 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17360 some routines that use cipher OIDs: some ciphers do not have OIDs
17361 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17366 * Simplify the trust setting structure and code. Now we just have
17367 two sequences of OIDs for trusted and rejected settings. These will
17368 typically have values the same as the extended key usage extension
17369 and any application specific purposes.
17371 The trust checking code now has a default behaviour: it will just
17372 check for an object with the same NID as the passed id. Functions can
17373 be provided to override either the default behaviour or the behaviour
17374 for a given id. SSL client, server and email already have functions
17375 in place for compatibility: they check the NID and also return "trusted"
17376 if the certificate is self signed.
17380 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17381 traditional format into an EVP_PKEY structure.
17385 * Add a password callback function PEM_cb() which either prompts for
17386 a password if usr_data is NULL or otherwise assumes it is a null
17387 terminated password. Allow passwords to be passed on command line
17388 environment or config files in a few more utilities.
17392 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17393 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17394 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17395 Update documentation.
17399 * Support for ASN1 "NULL" type. This could be handled before by using
17400 ASN1_TYPE but there wasn't any function that would try to read a NULL
17401 and produce an error if it couldn't. For compatibility we also have
17402 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17403 don't allocate anything because they don't need to.
17407 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17410 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17412 * Rebuild of the memory allocation routines used by OpenSSL code and
17413 possibly others as well. The purpose is to make an interface that
17414 provide hooks so anyone can build a separate set of allocation and
17415 deallocation routines to be used by OpenSSL, for example memory
17416 pool implementations, or something else, which was previously hard
17417 since Malloc(), Realloc() and Free() were defined as macros having
17418 the values malloc, realloc and free, respectively (except for Win32
17419 compilations). The same is provided for memory debugging code.
17420 OpenSSL already comes with functionality to find memory leaks, but
17421 this gives people a chance to debug other memory problems.
17423 With these changes, a new set of functions and macros have appeared:
17425 CRYPTO_set_mem_debug_functions() [F]
17426 CRYPTO_get_mem_debug_functions() [F]
17427 CRYPTO_dbg_set_options() [F]
17428 CRYPTO_dbg_get_options() [F]
17429 CRYPTO_malloc_debug_init() [M]
17431 The memory debug functions are NULL by default, unless the library
17432 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17433 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17434 gives the standard debugging functions that come with OpenSSL) or
17435 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17436 provided by the library user) must be used. When the standard
17437 debugging functions are used, CRYPTO_dbg_set_options can be used to
17438 request additional information:
17439 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17440 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17442 Also, things like CRYPTO_set_mem_functions will always give the
17443 expected result (the new set of functions is used for allocation
17444 and deallocation) at all times, regardless of platform and compiler
17447 To finish it up, some functions that were never use in any other
17448 way than through macros have a new API and new semantic:
17450 CRYPTO_dbg_malloc()
17451 CRYPTO_dbg_realloc()
17454 All macros of value have retained their old syntax.
17456 *Richard Levitte and Bodo Moeller*
17458 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17459 ordering of SMIMECapabilities wasn't in "strength order" and there
17460 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17465 * Some ASN1 types with illegal zero length encoding (INTEGER,
17466 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17468 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17470 * Merge in my S/MIME library for OpenSSL. This provides a simple
17471 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17472 functionality to handle multipart/signed properly) and a utility
17473 called 'smime' to call all this stuff. This is based on code I
17474 originally wrote for Celo who have kindly allowed it to be
17475 included in OpenSSL.
17479 * Add variants des_set_key_checked and des_set_key_unchecked of
17480 des_set_key (aka des_key_sched). Global variable des_check_key
17481 decides which of these is called by des_set_key; this way
17482 des_check_key behaves as it always did, but applications and
17483 the library itself, which was buggy for des_check_key == 1,
17484 have a cleaner way to pick the version they need.
17488 * New function PKCS12_newpass() which changes the password of a
17493 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17494 dynamic mix. In both cases the ids can be used as an index into the
17495 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17496 functions so they accept a list of the field values and the
17497 application doesn't need to directly manipulate the X509_TRUST
17502 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17507 * Modify the way the V3 extension code looks up extensions. This now
17508 works in a similar way to the object code: we have some "standard"
17509 extensions in a static table which is searched with OBJ_bsearch()
17510 and the application can add dynamic ones if needed. The file
17511 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17512 updated whenever a new extension is added to the core code and kept
17513 in ext_nid order. There is a simple program 'tabtest.c' which checks
17514 this. New extensions are not added too often so this file can readily
17515 be maintained manually.
17517 There are two big advantages in doing things this way. The extensions
17518 can be looked up immediately and no longer need to be "added" using
17519 X509V3_add_standard_extensions(): this function now does nothing.
17520 Side note: I get *lots* of email saying the extension code doesn't
17521 work because people forget to call this function.
17522 Also no dynamic allocation is done unless new extensions are added:
17523 so if we don't add custom extensions there is no need to call
17524 X509V3_EXT_cleanup().
17528 * Modify enc utility's salting as follows: make salting the default. Add a
17529 magic header, so unsalted files fail gracefully instead of just decrypting
17530 to garbage. This is because not salting is a big security hole, so people
17531 should be discouraged from doing it.
17535 * Fixes and enhancements to the 'x509' utility. It allowed a message
17536 digest to be passed on the command line but it only used this
17537 parameter when signing a certificate. Modified so all relevant
17538 operations are affected by the digest parameter including the
17539 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17540 DSA key was used because it didn't fix the digest.
17544 * Initial certificate chain verify code. Currently tests the untrusted
17545 certificates for consistency with the verify purpose (which is set
17546 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17548 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17549 this is because it will reject chains with invalid extensions whereas
17550 every previous version of OpenSSL and SSLeay made no checks at all.
17552 Trust code: checks the root CA for the relevant trust settings. Trust
17553 settings have an initial value consistent with the verify purpose: e.g.
17554 if the verify purpose is for SSL client use it expects the CA to be
17555 trusted for SSL client use. However the default value can be changed to
17556 permit custom trust settings: one example of this would be to only trust
17557 certificates from a specific "secure" set of CAs.
17559 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17560 which should be used for version portability: especially since the
17561 verify structure is likely to change more often now.
17563 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17564 to set them. If not set then assume SSL clients will verify SSL servers
17567 Two new options to the verify program: -untrusted allows a set of
17568 untrusted certificates to be passed in and -purpose which sets the
17569 intended purpose of the certificate. If a purpose is set then the
17570 new chain verify code is used to check extension consistency.
17574 * Support for the authority information access extension.
17578 * Modify RSA and DSA PEM read routines to transparently handle
17579 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17580 public keys in a format compatible with certificate
17581 SubjectPublicKeyInfo structures. Unfortunately there were already
17582 functions called *_PublicKey_* which used various odd formats so
17583 these are retained for compatibility: however the DSA variants were
17584 never in a public release so they have been deleted. Changed dsa/rsa
17585 utilities to handle the new format: note no releases ever handled public
17586 keys so we should be OK.
17588 The primary motivation for this change is to avoid the same fiasco
17589 that dogs private keys: there are several incompatible private key
17590 formats some of which are standard and some OpenSSL specific and
17591 require various evil hacks to allow partial transparent handling and
17592 even then it doesn't work with DER formats. Given the option anything
17593 other than PKCS#8 should be dumped: but the other formats have to
17594 stay in the name of compatibility.
17596 With public keys and the benefit of hindsight one standard format
17597 is used which works with EVP_PKEY, RSA or DSA structures: though
17598 it clearly returns an error if you try to read the wrong kind of key.
17600 Added a -pubkey option to the 'x509' utility to output the public key.
17601 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17602 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17603 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17604 that do the same as the `EVP_PKEY_assign_*()` except they up the
17605 reference count of the added key (they don't "swallow" the
17610 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17611 CRLs would fail if the file contained no certificates or no CRLs:
17612 added a new function to read in both types and return the number
17613 read: this means that if none are read it will be an error. The
17614 DER versions of the certificate and CRL reader would always fail
17615 because it isn't possible to mix certificates and CRLs in DER format
17616 without choking one or the other routine. Changed this to just read
17617 a certificate: this is the best we can do. Also modified the code
17618 in `apps/verify.c` to take notice of return codes: it was previously
17619 attempting to read in certificates from NULL pointers and ignoring
17620 any errors: this is one reason why the cert and CRL reader seemed
17621 to work. It doesn't check return codes from the default certificate
17622 routines: these may well fail if the certificates aren't installed.
17626 * Code to support otherName option in GeneralName.
17630 * First update to verify code. Change the verify utility
17631 so it warns if it is passed a self signed certificate:
17632 for consistency with the normal behaviour. X509_verify
17633 has been modified to it will now verify a self signed
17634 certificate if *exactly* the same certificate appears
17635 in the store: it was previously impossible to trust a
17636 single self signed certificate. This means that:
17637 openssl verify ss.pem
17638 now gives a warning about a self signed certificate but
17639 openssl verify -CAfile ss.pem ss.pem
17644 * For servers, store verify_result in SSL_SESSION data structure
17645 (and add it to external session representation).
17646 This is needed when client certificate verifications fails,
17647 but an application-provided verification callback (set by
17648 SSL_CTX_set_cert_verify_callback) allows accepting the session
17649 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17650 but returns 1): When the session is reused, we have to set
17651 ssl->verify_result to the appropriate error code to avoid
17654 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17656 * Fix a bug in the new PKCS#7 code: it didn't consider the
17657 case in PKCS7_dataInit() where the signed PKCS7 structure
17658 didn't contain any existing data because it was being created.
17660 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17662 * Add a salt to the key derivation routines in enc.c. This
17663 forms the first 8 bytes of the encrypted file. Also add a
17664 -S option to allow a salt to be input on the command line.
17668 * New function X509_cmp(). Oddly enough there wasn't a function
17669 to compare two certificates. We do this by working out the SHA1
17670 hash and comparing that. X509_cmp() will be needed by the trust
17675 * SSL_get1_session() is like SSL_get_session(), but increments
17676 the reference count in the SSL_SESSION returned.
17678 *Geoff Thorpe <geoff@eu.c2.net>*
17680 * Fix for 'req': it was adding a null to request attributes.
17681 Also change the X509_LOOKUP and X509_INFO code to handle
17682 certificate auxiliary information.
17686 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17691 * Add the possibility to add extra information to the memory leak
17692 detecting output, to form tracebacks, showing from where each
17693 allocation was originated: CRYPTO_push_info("constant string") adds
17694 the string plus current file name and line number to a per-thread
17695 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17696 is like calling CYRPTO_pop_info() until the stack is empty.
17697 Also updated memory leak detection code to be multi-thread-safe.
17701 * Add options -text and -noout to pkcs7 utility and delete the
17702 encryption options which never did anything. Update docs.
17706 * Add options to some of the utilities to allow the pass phrase
17707 to be included on either the command line (not recommended on
17708 OSes like Unix) or read from the environment. Update the
17709 manpages and fix a few bugs.
17713 * Add a few manpages for some of the openssl commands.
17717 * Fix the -revoke option in ca. It was freeing up memory twice,
17718 leaking and not finding already revoked certificates.
17722 * Extensive changes to support certificate auxiliary information.
17723 This involves the use of X509_CERT_AUX structure and X509_AUX
17724 functions. An X509_AUX function such as PEM_read_X509_AUX()
17725 can still read in a certificate file in the usual way but it
17726 will also read in any additional "auxiliary information". By
17727 doing things this way a fair degree of compatibility can be
17728 retained: existing certificates can have this information added
17729 using the new 'x509' options.
17731 Current auxiliary information includes an "alias" and some trust
17732 settings. The trust settings will ultimately be used in enhanced
17733 certificate chain verification routines: currently a certificate
17734 can only be trusted if it is self signed and then it is trusted
17739 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17740 The problem was that one of the replacement routines had not been working
17741 since SSLeay releases. For now the offending routine has been replaced
17742 with non-optimised assembler. Even so, this now gives around 95%
17743 performance improvement for 1024 bit RSA signs.
17747 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17748 handling. Most clients have the effective key size in bits equal to
17749 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17750 A few however don't do this and instead use the size of the decrypted key
17751 to determine the RC2 key length and the AlgorithmIdentifier to determine
17752 the effective key length. In this case the effective key length can still
17753 be 40 bits but the key length can be 168 bits for example. This is fixed
17754 by manually forcing an RC2 key into the EVP_PKEY structure because the
17755 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17756 the key length and effective key length are equal.
17760 * Add a bunch of functions that should simplify the creation of
17761 X509_NAME structures. Now you should be able to do:
17762 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17763 and have it automatically work out the correct field type and fill in
17764 the structures. The more adventurous can try:
17765 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17766 and it will (hopefully) work out the correct multibyte encoding.
17770 * Change the 'req' utility to use the new field handling and multibyte
17771 copy routines. Before the DN field creation was handled in an ad hoc
17772 way in req, ca, and x509 which was rather broken and didn't support
17773 BMPStrings or UTF8Strings. Since some software doesn't implement
17774 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17775 using the dirstring_type option. See the new comment in the default
17776 openssl.cnf for more info.
17780 * Make crypto/rand/md_rand.c more robust:
17781 - Assure unique random numbers after fork().
17782 - Make sure that concurrent threads access the global counter and
17783 md serializably so that we never lose entropy in them
17784 or use exactly the same state in multiple threads.
17785 Access to the large state is not always serializable because
17786 the additional locking could be a performance killer, and
17787 md should be large enough anyway.
17791 * New file `apps/app_rand.c` with commonly needed functionality
17792 for handling the random seed file.
17794 Use the random seed file in some applications that previously did not:
17796 dsaparam -genkey (which also ignored its '-rand' option),
17799 x509 (when signing).
17800 Except on systems with /dev/urandom, it is crucial to have a random
17801 seed file at least for key creation, DSA signing, and for DH exchanges;
17802 for RSA signatures we could do without one.
17804 gendh and gendsa (unlike genrsa) used to read only the first byte
17805 of each file listed in the '-rand' option. The function as previously
17806 found in genrsa is now in app_rand.c and is used by all programs
17807 that support '-rand'.
17811 * In RAND_write_file, use mode 0600 for creating files;
17812 don't just chmod when it may be too late.
17816 * Report an error from X509_STORE_load_locations
17817 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17821 * New function ASN1_mbstring_copy() this copies a string in either
17822 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17823 into an ASN1_STRING type. A mask of permissible types is passed
17824 and it chooses the "minimal" type to use or an error if not type
17829 * Add function equivalents to the various macros in asn1.h. The old
17830 macros are retained with an `M_` prefix. Code inside the library can
17831 use the `M_` macros. External code (including the openssl utility)
17832 should *NOT* in order to be "shared library friendly".
17836 * Add various functions that can check a certificate's extensions
17837 to see if it usable for various purposes such as SSL client,
17838 server or S/MIME and CAs of these types. This is currently
17839 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17840 verification. Also added a -purpose flag to x509 utility to
17841 print out all the purposes.
17845 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17850 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17851 for, obtain and decode and extension and obtain its critical flag.
17852 This allows all the necessary extension code to be handled in a
17853 single function call.
17857 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17858 platforms. See crypto/rc4/rc4_enc.c for further details.
17862 * New -noout option to asn1parse. This causes no output to be produced
17863 its main use is when combined with -strparse and -out to extract data
17864 from a file (which may not be in ASN.1 format).
17868 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17869 when producing the local key id.
17871 *Richard Levitte <levitte@stacken.kth.se>*
17873 * New option -dhparam in s_server. This allows a DH parameter file to be
17874 stated explicitly. If it is not stated then it tries the first server
17875 certificate file. The previous behaviour hard coded the filename
17880 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17881 a public key to be input or output. For example:
17882 openssl rsa -in key.pem -pubout -out pubkey.pem
17883 Also added necessary DSA public key functions to handle this.
17887 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17888 in the message. This was handled by allowing
17889 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17891 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17893 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17894 to the end of the strings whereas this didn't. This would cause problems
17895 if strings read with d2i_ASN1_bytes() were later modified.
17897 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17899 * Fix for base64 decode bug. When a base64 bio reads only one line of
17900 data and it contains EOF it will end up returning an error. This is
17901 caused by input 46 bytes long. The cause is due to the way base64
17902 BIOs find the start of base64 encoded data. They do this by trying a
17903 trial decode on each line until they find one that works. When they
17904 do a flag is set and it starts again knowing it can pass all the
17905 data directly through the decoder. Unfortunately it doesn't reset
17906 the context it uses. This means that if EOF is reached an attempt
17907 is made to pass two EOFs through the context and this causes the
17908 resulting error. This can also cause other problems as well. As is
17909 usual with these problems it takes *ages* to find and the fix is
17910 trivial: move one line.
17912 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17914 * Ugly workaround to get s_client and s_server working under Windows. The
17915 old code wouldn't work because it needed to select() on sockets and the
17916 tty (for keypresses and to see if data could be written). Win32 only
17917 supports select() on sockets so we select() with a 1s timeout on the
17918 sockets and then see if any characters are waiting to be read, if none
17919 are present then we retry, we also assume we can always write data to
17920 the tty. This isn't nice because the code then blocks until we've
17921 received a complete line of data and it is effectively polling the
17922 keyboard at 1s intervals: however it's quite a bit better than not
17923 working at all :-) A dedicated Windows application might handle this
17924 with an event loop for example.
17928 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17929 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17930 will be called when RSA_sign() and RSA_verify() are used. This is useful
17931 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17932 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17933 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17934 This necessitated the support of an extra signature type NID_md5_sha1
17935 for SSL signatures and modifications to the SSL library to use it instead
17936 of calling RSA_public_decrypt() and RSA_private_encrypt().
17940 * Add new -verify -CAfile and -CApath options to the crl program, these
17941 will lookup a CRL issuers certificate and verify the signature in a
17942 similar way to the verify program. Tidy up the crl program so it
17943 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17944 less strict. It will now permit CRL extensions even if it is not
17945 a V2 CRL: this will allow it to tolerate some broken CRLs.
17949 * Initialize all non-automatic variables each time one of the openssl
17950 sub-programs is started (this is necessary as they may be started
17951 multiple times from the "OpenSSL>" prompt).
17953 *Lennart Bang, Bodo Moeller*
17955 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17956 removing all other RSA functionality (this is what NO_RSA does). This
17957 is so (for example) those in the US can disable those operations covered
17958 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17963 * Non-copying interface to BIO pairs.
17964 (still largely untested)
17968 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17969 ASCII string. This was handled independently in various places before.
17973 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17974 UTF8 strings a character at a time.
17978 * Use client_version from client hello to select the protocol
17979 (s23_srvr.c) and for RSA client key exchange verification
17980 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17984 * Add various utility functions to handle SPKACs, these were previously
17985 handled by poking round in the structure internals. Added new function
17986 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17987 print, verify and generate SPKACs. Based on an original idea from
17988 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17992 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17996 * Allow the config file extension section to be overwritten on the
17997 command line. Based on an original idea from Massimiliano Pala
17998 <madwolf@comune.modena.it>. The new option is called -extensions
17999 and can be applied to ca, req and x509. Also -reqexts to override
18000 the request extensions in req and -crlexts to override the crl extensions
18005 * Add new feature to the SPKAC handling in ca. Now you can include
18006 the same field multiple times by preceding it by "XXXX." for example:
18009 this is the same syntax as used in the req config file.
18013 * Allow certificate extensions to be added to certificate requests. These
18014 are specified in a 'req_extensions' option of the req section of the
18015 config file. They can be printed out with the -text option to req but
18016 are otherwise ignored at present.
18020 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
18021 data read consists of only the final block it would not decrypted because
18022 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
18023 A misplaced 'break' also meant the decrypted final block might not be
18024 copied until the next read.
18028 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
18029 a few extra parameters to the DH structure: these will be useful if
18030 for example we want the value of 'q' or implement X9.42 DH.
18034 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
18035 provides hooks that allow the default DSA functions or functions on a
18036 "per key" basis to be replaced. This allows hardware acceleration and
18037 hardware key storage to be handled without major modification to the
18038 library. Also added low-level modexp hooks and CRYPTO_EX structure and
18039 associated functions.
18043 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
18044 as "read only": it can't be written to and the buffer it points to will
18045 not be freed. Reading from a read only BIO is much more efficient than
18046 a normal memory BIO. This was added because there are several times when
18047 an area of memory needs to be read from a BIO. The previous method was
18048 to create a memory BIO and write the data to it, this results in two
18049 copies of the data and an O(n^2) reading algorithm. There is a new
18050 function BIO_new_mem_buf() which creates a read only memory BIO from
18051 an area of memory. Also modified the PKCS#7 routines to use read only
18056 * Bugfix: ssl23_get_client_hello did not work properly when called in
18057 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
18058 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
18059 but a retry condition occurred while trying to read the rest.
18063 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
18064 NID_pkcs7_encrypted by default: this was wrong since this should almost
18065 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
18066 the encrypted data type: this is a more sensible place to put it and it
18067 allows the PKCS#12 code to be tidied up that duplicated this
18072 * Changed obj_dat.pl script so it takes its input and output files on
18073 the command line. This should avoid shell escape redirection problems
18078 * Initial support for certificate extension requests, these are included
18079 in things like Xenroll certificate requests. Included functions to allow
18080 extensions to be obtained and added.
18084 * -crlf option to s_client and s_server for sending newlines as
18085 CRLF (as required by many protocols).
18089 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
18091 * Install libRSAglue.a when OpenSSL is built with RSAref.
18093 *Ralf S. Engelschall*
18095 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
18097 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
18099 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
18104 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
18105 DH parameters/keys (q is lost during that conversion, but the resulting
18106 DH parameters contain its length).
18108 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
18109 much faster than DH_generate_parameters (which creates parameters
18110 where `p = 2*q + 1`), and also the smaller q makes DH computations
18111 much more efficient (160-bit exponentiation instead of 1024-bit
18112 exponentiation); so this provides a convenient way to support DHE
18113 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
18114 utter importance to use
18115 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18117 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18118 when such DH parameters are used, because otherwise small subgroup
18119 attacks may become possible!
18123 * Avoid memory leak in i2d_DHparams.
18127 * Allow the -k option to be used more than once in the enc program:
18128 this allows the same encrypted message to be read by multiple recipients.
18132 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
18133 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
18134 it will always use the numerical form of the OID, even if it has a short
18139 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
18140 method only got called if p,q,dmp1,dmq1,iqmp components were present,
18141 otherwise bn_mod_exp was called. In the case of hardware keys for example
18142 no private key components need be present and it might store extra data
18143 in the RSA structure, which cannot be accessed from bn_mod_exp.
18144 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
18145 private key operations.
18149 * Added support for SPARC Linux.
18153 * pem_password_cb function type incompatibly changed from
18154 typedef int pem_password_cb(char *buf, int size, int rwflag);
18156 ....(char *buf, int size, int rwflag, void *userdata);
18157 so that applications can pass data to their callbacks:
18158 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18159 additional void * argument, which is just handed through whenever
18160 the password callback is called.
18162 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18164 New function SSL_CTX_set_default_passwd_cb_userdata.
18166 Compatibility note: As many C implementations push function arguments
18167 onto the stack in reverse order, the new library version is likely to
18168 interoperate with programs that have been compiled with the old
18169 pem_password_cb definition (PEM_whatever takes some data that
18170 happens to be on the stack as its last argument, and the callback
18171 just ignores this garbage); but there is no guarantee whatsoever that
18174 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18175 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18176 problems not only on Windows, but also on some Unix platforms.
18177 To avoid problematic command lines, these definitions are now in an
18178 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18179 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18183 * MIPS III/IV assembler module is reimplemented.
18187 * More DES library cleanups: remove references to srand/rand and
18188 delete an unused file.
18192 * Add support for the free Netwide assembler (NASM) under Win32,
18193 since not many people have MASM (ml) and it can be hard to obtain.
18194 This is currently experimental but it seems to work OK and pass all
18195 the tests. Check out INSTALL.W32 for info.
18199 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18200 without temporary keys kept an extra copy of the server key,
18201 and connections with temporary keys did not free everything in case
18206 * New function RSA_check_key and new openssl rsa option -check
18207 for verifying the consistency of RSA keys.
18209 *Ulf Moeller, Bodo Moeller*
18211 * Various changes to make Win32 compile work:
18212 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18213 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18214 comparison" warnings.
18215 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18219 * Add a debugging option to PKCS#5 v2 key generation function: when
18220 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18221 derived keys are printed to stderr.
18225 * Copy the flags in ASN1_STRING_dup().
18227 *Roman E. Pavlov <pre@mo.msk.ru>*
18229 * The x509 application mishandled signing requests containing DSA
18230 keys when the signing key was also DSA and the parameters didn't match.
18232 It was supposed to omit the parameters when they matched the signing key:
18233 the verifying software was then supposed to automatically use the CA's
18234 parameters if they were absent from the end user certificate.
18236 Omitting parameters is no longer recommended. The test was also
18237 the wrong way round! This was probably due to unusual behaviour in
18238 EVP_cmp_parameters() which returns 1 if the parameters match.
18239 This meant that parameters were omitted when they *didn't* match and
18240 the certificate was useless. Certificates signed with 'ca' didn't have
18243 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18245 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18246 The interface is as follows:
18247 Applications can use
18248 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18249 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18250 "off" is now the default.
18251 The library internally uses
18252 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18253 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18254 to disable memory-checking temporarily.
18256 Some inconsistent states that previously were possible (and were
18257 even the default) are now avoided.
18259 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18260 with each memory chunk allocated; this is occasionally more helpful
18261 than just having a counter.
18263 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18265 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18270 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18271 which largely parallels "options", but is for changing API behaviour,
18272 whereas "options" are about protocol behaviour.
18273 Initial "mode" flags are:
18275 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18276 a single record has been written.
18277 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18278 retries use the same buffer location.
18279 (But all of the contents must be
18284 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18287 * Fix problems with no-hmac etc.
18289 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18291 * New functions RSA_get_default_method(), RSA_set_method() and
18292 RSA_get_method(). These allows replacement of RSA_METHODs without having
18293 to mess around with the internals of an RSA structure.
18297 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18298 Also really enable memory leak checks in openssl.c and in some
18301 *Chad C. Mulligan, Bodo Moeller*
18303 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18304 up the length of negative integers. This has now been simplified to just
18305 store the length when it is first determined and use it later, rather
18306 than trying to keep track of where data is copied and updating it to
18308 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18310 * Add a new function PKCS7_signatureVerify. This allows the verification
18311 of a PKCS#7 signature but with the signing certificate passed to the
18312 function itself. This contrasts with PKCS7_dataVerify which assumes the
18313 certificate is present in the PKCS#7 structure. This isn't always the
18314 case: certificates can be omitted from a PKCS#7 structure and be
18315 distributed by "out of band" means (such as a certificate database).
18319 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18320 function prototypes in pem.h, also change util/mkdef.pl to add the
18321 necessary function names.
18325 * mk1mf.pl (used by Windows builds) did not properly read the
18326 options set by Configure in the top level Makefile, and Configure
18327 was not even able to write more than one option correctly.
18328 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18332 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18333 file to be loaded from a BIO or FILE pointer. The BIO version will
18334 for example allow memory BIOs to contain config info.
18338 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18339 Whoever hopes to achieve shared-library compatibility across versions
18340 must use this, not the compile-time macro.
18341 (Exercise 0.9.4: Which is the minimum library version required by
18343 Note: All this applies only to multi-threaded programs, others don't
18348 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18349 through a BIO pair triggered the default case, i.e.
18350 SSLerr(...,SSL_R_UNKNOWN_STATE).
18354 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18355 can use the SSL library even if none of the specific BIOs is
18360 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18361 for the encoded length.
18363 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18365 * Add initial documentation of the X509V3 functions.
18369 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18370 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18371 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18372 secure PKCS#8 private key format with a high iteration count.
18376 * Fix determination of Perl interpreter: A perl or perl5
18377 *directory* in $PATH was also accepted as the interpreter.
18379 *Ralf S. Engelschall*
18381 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18382 wrong with it but it was very old and did things like calling
18383 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18384 unusual formatting.
18388 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18389 to use the new extension code.
18393 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18394 with macros. This should make it easier to change their form, add extra
18395 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18400 * Add to configuration table a new entry that can specify an alternative
18401 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18402 according to Mark Crispin <MRC@Panda.COM>.
18406 * DES CBC did not update the IV. Weird.
18410 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18411 Changing the behaviour of the former might break existing programs --
18412 where IV updating is needed, des_ncbc_encrypt can be used.
18415 * When bntest is run from "make test" it drives bc to check its
18416 calculations, as well as internally checking them. If an internal check
18417 fails, it needs to cause bc to give a non-zero result or make test carries
18418 on without noticing the failure. Fixed.
18422 * DES library cleanups.
18426 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18427 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18428 ciphers. NOTE: although the key derivation function has been verified
18429 against some published test vectors it has not been extensively tested
18430 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18435 * Instead of "mkdir -p", which is not fully portable, use new
18436 Perl script "util/mkdir-p.pl".
18440 * Rewrite the way password based encryption (PBE) is handled. It used to
18441 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18442 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18443 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18444 the 'parameter' field of the AlgorithmIdentifier is passed to the
18445 underlying key generation function so it must do its own ASN1 parsing.
18446 This has also changed the EVP_PBE_CipherInit() function which now has a
18447 'parameter' argument instead of literal salt and iteration count values
18448 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18452 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18453 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18454 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18455 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18456 value was just used as a "magic string" and not used directly its
18457 value doesn't matter.
18461 * Introduce some semblance of const correctness to BN. Shame C doesn't
18466 * "linux-sparc64" configuration (ultrapenguin).
18468 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18469 "linux-sparc" configuration.
18471 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18473 * config now generates no-xxx options for missing ciphers.
18477 * Support the EBCDIC character set (work in progress).
18478 File ebcdic.c not yet included because it has a different license.
18480 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18482 * Support BS2000/OSD-POSIX.
18484 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18486 * Make callbacks for key generation use `void *` instead of `char *`.
18490 * Make S/MIME samples compile (not yet tested).
18494 * Additional typesafe stacks.
18498 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18502 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18504 * New configuration variant "sco5-gcc".
18506 * Updated some demos.
18508 *Sean O Riordain, Wade Scholine*
18510 * Add missing BIO_free at exit of pkcs12 application.
18514 * Fix memory leak in conf.c.
18518 * Updates for Win32 to assembler version of MD5.
18522 * Set #! path to perl in `apps/der_chop` to where we found it
18523 instead of using a fixed path.
18527 * SHA library changes for irix64-mips4-cc.
18531 * Improvements for VMS support.
18535 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18537 * Bignum library bug fix. IRIX 6 passes "make test" now!
18538 This also avoids the problems with SC4.2 and unpatched SC5.
18540 *Andy Polyakov <appro@fy.chalmers.se>*
18542 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18543 These are required because of the typesafe stack would otherwise break
18544 existing code. If old code used a structure member which used to be STACK
18545 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18546 sk_num or sk_value it would produce an error because the num, data members
18547 are not present in STACK_OF. Now it just produces a warning. sk_set
18548 replaces the old method of assigning a value to sk_value
18549 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18550 that does this will no longer work (and should use sk_set instead) but
18551 this could be regarded as a "questionable" behaviour anyway.
18555 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18556 correctly handle encrypted S/MIME data.
18560 * Change type of various DES function arguments from des_cblock
18561 (which means, in function argument declarations, pointer to char)
18562 to des_cblock * (meaning pointer to array with 8 char elements),
18563 which allows the compiler to do more typechecking; it was like
18564 that back in SSLeay, but with lots of ugly casts.
18566 Introduce new type const_des_cblock.
18570 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18571 problems: find RecipientInfo structure that matches recipient certificate
18572 and initialise the ASN1 structures properly based on passed cipher.
18576 * Belatedly make the BN tests actually check the results.
18580 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18581 to and from BNs: it was completely broken. New compilation option
18582 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18583 key elements as negative integers.
18587 * Reorganize and speed up MD5.
18589 *Andy Polyakov <appro@fy.chalmers.se>*
18593 *Richard Levitte <richard@levitte.org>*
18595 * New option -out to asn1parse to allow the parsed structure to be
18596 output to a file. This is most useful when combined with the -strparse
18597 option to examine the output of things like OCTET STRINGS.
18601 * Make SSL library a little more fool-proof by not requiring any longer
18602 that `SSL_set_{accept,connect}_state` be called before
18603 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18604 in many applications because usually everything *appeared* to work as
18605 intended anyway -- now it really works as intended).
18609 * Move openssl.cnf out of lib/.
18613 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18614 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18615 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18617 *Ralf S. Engelschall*
18619 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18620 handle PKCS#7 enveloped data properly.
18622 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18624 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18625 copying pointers. The cert_st handling is changed by this in
18626 various ways (and thus what used to be known as ctx->default_cert
18627 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18628 any longer when s->cert does not give us what we need).
18629 ssl_cert_instantiate becomes obsolete by this change.
18630 As soon as we've got the new code right (possibly it already is?),
18631 we have solved a couple of bugs of the earlier code where s->cert
18632 was used as if it could not have been shared with other SSL structures.
18634 Note that using the SSL API in certain dirty ways now will result
18635 in different behaviour than observed with earlier library versions:
18636 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18637 does not influence s as it used to.
18639 In order to clean up things more thoroughly, inside SSL_SESSION
18640 we don't use CERT any longer, but a new structure SESS_CERT
18641 that holds per-session data (if available); currently, this is
18642 the peer's certificate chain and, for clients, the server's certificate
18643 and temporary key. CERT holds only those values that can have
18644 meaningful defaults in an SSL_CTX.
18648 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18649 from the internal representation. Various PKCS#7 fixes: remove some
18650 evil casts and set the enc_dig_alg field properly based on the signing
18655 * Allow PKCS#12 password to be set from the command line or the
18656 environment. Let 'ca' get its config file name from the environment
18657 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18662 * Allow certificate policies extension to use an IA5STRING for the
18663 organization field. This is contrary to the PKIX definition but
18664 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18669 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18670 without disallowing inline assembler and the like for non-pedantic builds.
18674 * Support Borland C++ builder.
18676 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18682 * SHA-1 cleanups and performance enhancements.
18684 *Andy Polyakov <appro@fy.chalmers.se>*
18686 * Sparc v8plus assembler for the bignum library.
18688 *Andy Polyakov <appro@fy.chalmers.se>*
18690 * Accept any -xxx and +xxx compiler options in Configure.
18694 * Update HPUX configuration.
18698 * Add missing `sk_<type>_unshift()` function to safestack.h
18700 *Ralf S. Engelschall*
18702 * New function SSL_CTX_use_certificate_chain_file that sets the
18703 "extra_cert"s in addition to the certificate. (This makes sense
18704 only for "PEM" format files, as chains as a whole are not
18709 * Support verify_depth from the SSL API.
18710 x509_vfy.c had what can be considered an off-by-one-error:
18711 Its depth (which was not part of the external interface)
18712 was actually counting the number of certificates in a chain;
18713 now it really counts the depth.
18717 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18718 instead of X509err, which often resulted in confusing error
18719 messages since the error codes are not globally unique
18720 (e.g. an alleged error in ssl3_accept when a certificate
18721 didn't match the private key).
18723 * New function SSL_CTX_set_session_id_context that allows to set a default
18724 value (so that you don't need SSL_set_session_id_context for each
18725 connection using the SSL_CTX).
18729 * OAEP decoding bug fix.
18733 * Support INSTALL_PREFIX for package builders, as proposed by
18738 * New Configure options "threads" and "no-threads". For systems
18739 where the proper compiler options are known (currently Solaris
18740 and Linux), "threads" is the default.
18744 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18748 * Install various scripts to $(OPENSSLDIR)/misc, not to
18749 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18750 such as /usr/local/bin.
18754 * "make linux-shared" to build shared libraries.
18756 *Niels Poppe <niels@netbox.org>*
18758 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18762 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18763 extension adding in x509 utility.
18767 * Remove NOPROTO sections and error code comments.
18771 * Partial rewrite of the DEF file generator to now parse the ANSI
18776 * New Configure options --prefix=DIR and --openssldir=DIR.
18780 * Complete rewrite of the error code script(s). It is all now handled
18781 by one script at the top level which handles error code gathering,
18782 header rewriting and C source file generation. It should be much better
18783 than the old method: it now uses a modified version of Ulf's parser to
18784 read the ANSI prototypes in all header files (thus the old K&R definitions
18785 aren't needed for error creation any more) and do a better job of
18786 translating function codes into names. The old 'ASN1 error code embedded
18787 in a comment' is no longer necessary and it doesn't use .err files which
18788 have now been deleted. Also the error code call doesn't have to appear all
18789 on one line (which resulted in some large lines...).
18793 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18797 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18798 0 (which usually indicates a closed connection), but continue reading.
18802 * Fix some race conditions.
18806 * Add support for CRL distribution points extension. Add Certificate
18807 Policies and CRL distribution points documentation.
18811 * Move the autogenerated header file parts to crypto/opensslconf.h.
18815 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18816 8 of keying material. Merlin has also confirmed interop with this fix
18817 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18819 *Merlin Hughes <merlin@baltimore.ie>*
18821 * Fix lots of warnings.
18823 *Richard Levitte <levitte@stacken.kth.se>*
18825 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18826 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18828 *Richard Levitte <levitte@stacken.kth.se>*
18830 * Fix problems with sizeof(long) == 8.
18832 *Andy Polyakov <appro@fy.chalmers.se>*
18834 * Change functions to ANSI C.
18838 * Fix typos in error codes.
18840 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18842 * Remove defunct assembler files from Configure.
18846 * SPARC v8 assembler BIGNUM implementation.
18848 *Andy Polyakov <appro@fy.chalmers.se>*
18850 * Support for Certificate Policies extension: both print and set.
18851 Various additions to support the r2i method this uses.
18855 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18856 return a const string when you are expecting an allocated buffer.
18860 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18861 types DirectoryString and DisplayText.
18865 * Add code to allow r2i extensions to access the configuration database,
18866 add an LHASH database driver and add several ctx helper functions.
18870 * Fix an evil bug in bn_expand2() which caused various BN functions to
18871 fail when they extended the size of a BIGNUM.
18875 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18876 support typesafe stack.
18880 * Fix typo in SSL_[gs]et_options().
18882 *Nils Frostberg <nils@medcom.se>*
18884 * Delete various functions and files that belonged to the (now obsolete)
18885 old X509V3 handling code.
18889 * New Configure option "rsaref".
18893 * Don't auto-generate pem.h.
18897 * Introduce type-safe ASN.1 SETs.
18901 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18903 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18905 * Introduce type-safe STACKs. This will almost certainly break lots of code
18906 that links with OpenSSL (well at least cause lots of warnings), but fear
18907 not: the conversion is trivial, and it eliminates loads of evil casts. A
18908 few STACKed things have been converted already. Feel free to convert more.
18909 In the fullness of time, I'll do away with the STACK type altogether.
18913 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18914 specified in `<certfile>` by updating the entry in the index.txt file.
18915 This way one no longer has to edit the index.txt file manually for
18916 revoking a certificate. The -revoke option does the gory details now.
18918 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18920 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18921 `-text` option at all and this way the `-noout -text` combination was
18922 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18924 *Ralf S. Engelschall*
18926 * Make sure a corresponding plain text error message exists for the
18927 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18928 verify callback function determined that a certificate was revoked.
18930 *Ralf S. Engelschall*
18932 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18933 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18934 all available ciphers including rc5, which was forgotten until now.
18935 In order to let the testing shell script know which algorithms
18936 are available, a new (up to now undocumented) command
18937 `openssl list-cipher-commands` is used.
18941 * Bugfix: s_client occasionally would sleep in select() when
18942 it should have checked SSL_pending() first.
18946 * New functions DSA_do_sign and DSA_do_verify to provide access to
18947 the raw DSA values prior to ASN.1 encoding.
18951 * Tweaks to Configure
18953 *Niels Poppe <niels@netbox.org>*
18955 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18960 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18964 * New config option to avoid instructions that are illegal on the 80386.
18965 The default code is faster, but requires at least a 486.
18969 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18970 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18971 same as SSL2_VERSION anyway.
18975 * New "-showcerts" option for s_client.
18979 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18980 application. Various cleanups and fixes.
18984 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18985 modify error routines to work internally. Add error codes and PBE init
18986 to library startup routines.
18990 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18991 packing functions to asn1 and evp. Changed function names and error
18992 codes along the way.
18996 * PKCS12 integration: and so it begins... First of several patches to
18997 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18998 objects to objects.h
19002 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
19003 and display support for Thawte strong extranet extension.
19007 * Add LinuxPPC support.
19009 *Jeff Dubrule <igor@pobox.org>*
19011 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
19012 bn_div_words in alpha.s.
19014 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
19016 * Make sure the RSA OAEP test is skipped under -DRSAref because
19017 OAEP isn't supported when OpenSSL is built with RSAref.
19019 *Ulf Moeller <ulf@fitug.de>*
19021 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
19022 so they no longer are missing under -DNOPROTO.
19024 *Soren S. Jorvang <soren@t.dk>*
19026 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
19028 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
19029 doesn't work when the session is reused. Coming soon!
19033 * Fix a security hole, that allows sessions to be reused in the wrong
19034 context thus bypassing client cert protection! All software that uses
19035 client certs and session caches in multiple contexts NEEDS PATCHING to
19036 allow session reuse! A fuller solution is in the works.
19038 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
19040 * Some more source tree cleanups (removed obsolete files
19041 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
19042 permission on "config" script to be executable) and a fix for the INSTALL
19045 *Ulf Moeller <ulf@fitug.de>*
19047 * Remove some legacy and erroneous uses of malloc, free instead of
19050 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
19052 * Make rsa_oaep_test return non-zero on error.
19054 *Ulf Moeller <ulf@fitug.de>*
19056 * Add support for native Solaris shared libraries. Configure
19057 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
19058 if someone would make that last step automatic.
19060 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
19062 * ctx_size was not built with the right compiler during "make links". Fixed.
19066 * Change the meaning of 'ALL' in the cipher list. It now means "everything
19067 except NULL ciphers". This means the default cipher list will no longer
19068 enable NULL ciphers. They need to be specifically enabled e.g. with
19069 the string "DEFAULT:eNULL".
19073 * Fix to RSA private encryption routines: if p < q then it would
19074 occasionally produce an invalid result. This will only happen with
19075 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
19079 * Be less restrictive and allow also `perl util/perlpath.pl
19080 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
19081 because this way one can also use an interpreter named `perl5` (which is
19082 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
19083 installed as `perl`).
19085 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19087 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
19089 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19091 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
19092 advapi32.lib to Win32 build and change the pem test comparison
19093 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
19094 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
19095 and crypto/des/ede_cbcm_enc.c.
19099 * DES quad checksum was broken on big-endian architectures. Fixed.
19103 * Comment out two functions in bio.h that aren't implemented. Fix up the
19104 Win32 test batch file so it (might) work again. The Win32 test batch file
19105 is horrible: I feel ill....
19109 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
19110 in e_os.h. Audit of header files to check ANSI and non ANSI
19111 sections: 10 functions were absent from non ANSI section and not exported
19112 from Windows DLLs. Fixed up libeay.num for new functions.
19116 * Make `openssl version` output lines consistent.
19118 *Ralf S. Engelschall*
19120 * Fix Win32 symbol export lists for BIO functions: Added
19121 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
19122 to ms/libeay{16,32}.def.
19124 *Ralf S. Engelschall*
19126 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
19127 fine under Unix and passes some trivial tests I've now added. But the
19128 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
19129 added to make sure no one expects that this stuff really works in the
19130 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
19131 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
19134 *Ralf S. Engelschall*
19136 * Fix the generation of two part addresses in perl.
19138 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
19140 * Add config entry for Linux on MIPS.
19142 *John Tobey <jtobey@channel1.com>*
19144 * Make links whenever Configure is run, unless we are on Windoze.
19148 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
19149 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19154 * Add a useful kludge to allow package maintainers to specify compiler and
19155 other platforms details on the command line without having to patch the
19156 Configure script every time: One now can use
19157 `perl Configure <id>:<details>`,
19158 i.e. platform ids are allowed to have details appended
19159 to them (separated by colons). This is treated as there would be a static
19160 pre-configured entry in Configure's %table under key `<id>` with value
19161 `<details>` and `perl Configure <id>` is called. So, when you want to
19162 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19163 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19164 now, which overrides the FreeBSD-elf entry on-the-fly.
19166 *Ralf S. Engelschall*
19168 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19172 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19173 on the `perl Configure ...` command line. This way one can compile
19174 OpenSSL libraries with Position Independent Code (PIC) which is needed
19175 for linking it into DSOs.
19177 *Ralf S. Engelschall*
19179 * Remarkably, export ciphers were totally broken and no-one had noticed!
19184 * Cleaned up the LICENSE document: The official contact for any license
19185 questions now is the OpenSSL core team under openssl-core@openssl.org.
19186 And add a paragraph about the dual-license situation to make sure people
19187 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19188 to the OpenSSL toolkit.
19190 *Ralf S. Engelschall*
19192 * General source tree makefile cleanups: Made `making xxx in yyy...`
19193 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19194 Additionally cleaned up the `make links` target: Remove unnecessary
19195 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19196 to speed processing and no longer clutter the display with confusing
19197 stuff. Instead only the actually done links are displayed.
19199 *Ralf S. Engelschall*
19201 * Permit null encryption ciphersuites, used for authentication only. It used
19202 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19203 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19208 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19209 signed attributes when verifying signatures (this would break them),
19210 the detached data encoding was wrong and public keys obtained using
19211 X509_get_pubkey() weren't freed.
19215 * Add text documentation for the BUFFER functions. Also added a work around
19216 to a Win95 console bug. This was triggered by the password read stuff: the
19217 last character typed gets carried over to the next fread(). If you were
19218 generating a new cert request using 'req' for example then the last
19219 character of the passphrase would be CR which would then enter the first
19224 * Added the new 'Includes OpenSSL Cryptography Software' button as
19225 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19226 button and can be used by applications based on OpenSSL to show the
19227 relationship to the OpenSSL project.
19229 *Ralf S. Engelschall*
19231 * Remove confusing variables in function signatures in files
19232 ssl/ssl_lib.c and ssl/ssl.h.
19234 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19236 * Don't install bss_file.c under PREFIX/include/
19238 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19240 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19241 functions that return function pointers and has support for NT specific
19242 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19243 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19244 unsigned to signed types: this was killing the Win32 compile.
19248 * Add new certificate file to stack functions,
19249 SSL_add_dir_cert_subjects_to_stack() and
19250 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19251 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19252 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19253 This means that Apache-SSL and similar packages don't have to mess around
19254 to add as many CAs as they want to the preferred list.
19258 * Experiment with doxygen documentation. Currently only partially applied to
19260 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19261 openssl.doxy as the configuration file.
19265 * Get rid of remaining C++-style comments which strict C compilers hate.
19267 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19269 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19270 compiled in by default: it has problems with large keys.
19274 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19275 DH private keys and/or callback functions which directly correspond to
19276 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19277 is needed for applications which have to configure certificates on a
19278 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19280 For the RSA certificate situation is makes no difference, but
19281 for the DSA certificate situation this fixes the "no shared cipher"
19282 problem where the OpenSSL cipher selection procedure failed because the
19283 temporary keys were not overtaken from the context and the API provided
19284 no way to reconfigure them.
19285 The new functions now let applications reconfigure the stuff and they
19286 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19287 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19288 non-public-API function ssl_cert_instantiate() is used as a helper
19289 function and also to reduce code redundancy inside ssl_rsa.c.
19291 *Ralf S. Engelschall*
19293 * Move s_server -dcert and -dkey options out of the undocumented feature
19294 area because they are useful for the DSA situation and should be
19295 recognized by the users.
19297 *Ralf S. Engelschall*
19299 * Fix the cipher decision scheme for export ciphers: the export bits are
19300 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19301 SSL_EXP_MASK. So, the original variable has to be used instead of the
19302 already masked variable.
19304 *Richard Levitte <levitte@stacken.kth.se>*
19306 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19308 *Richard Levitte <levitte@stacken.kth.se>*
19310 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19311 from `int` to `unsigned int` because it is a length and initialized by
19312 EVP_DigestFinal() which expects an `unsigned int *`.
19314 *Richard Levitte <levitte@stacken.kth.se>*
19316 * Don't hard-code path to Perl interpreter on shebang line of Configure
19317 script. Instead use the usual Shell->Perl transition trick.
19319 *Ralf S. Engelschall*
19321 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19322 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19323 -noout -modulus` as it's already the case for `openssl rsa -noout
19324 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19325 currently the public key is printed (a decision which was already done by
19326 `openssl dsa -modulus` in the past) which serves a similar purpose.
19327 Additionally the NO_RSA no longer completely removes the whole -modulus
19328 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19331 *Ralf S. Engelschall*
19333 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19334 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19336 *Arne Ansper <arne@ats.cyber.ee>*
19338 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19339 to be added. Now both 'req' and 'ca' can use new objects defined in the
19344 * Add cool BIO that does syslog (or event log on NT).
19346 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19348 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19349 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19350 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19351 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19355 * Add preliminary config info for new extension code.
19359 * Make RSA_NO_PADDING really use no padding.
19361 *Ulf Moeller <ulf@fitug.de>*
19363 * Generate errors when private/public key check is done.
19367 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19368 for some CRL extensions and new objects added.
19372 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19373 key usage extension and fuller support for authority key id.
19377 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19378 padding method for RSA, which is recommended for new applications in PKCS
19379 #1 v2.0 (RFC 2437, October 1998).
19380 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19381 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19382 against Bleichbacher's attack on RSA.
19383 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19386 * Updates to the new SSL compression code
19388 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19390 * Fix so that the version number in the master secret, when passed
19391 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19392 (because the server will not accept higher), that the version number
19393 is 0x03,0x01, not 0x03,0x00
19395 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19397 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19398 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19399 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19403 * Support for RAW extensions where an arbitrary extension can be
19404 created by including its DER encoding. See `apps/openssl.cnf` for
19409 * Make sure latest Perl versions don't interpret some generated C array
19410 code as Perl array code in the crypto/err/err_genc.pl script.
19412 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19414 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19415 not many people have the assembler. Various Win32 compilation fixes and
19416 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19417 build instructions.
19421 * Modify configure script 'Configure' to automatically create crypto/date.h
19422 file under Win32 and also build pem.h from pem.org. New script
19423 util/mkfiles.pl to create the MINFO file on environments that can't do a
19424 'make files': perl util/mkfiles.pl >MINFO should work.
19428 * Major rework of DES function declarations, in the pursuit of correctness
19429 and purity. As a result, many evil casts evaporated, and some weirdness,
19430 too. You may find this causes warnings in your code. Zapping your evil
19431 casts will probably fix them. Mostly.
19435 * Fix for a typo in asn1.h. Bug fix to object creation script
19436 obj_dat.pl. It considered a zero in an object definition to mean
19437 "end of object": none of the objects in objects.h have any zeros
19438 so it wasn't spotted.
19440 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19442 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19443 Masking (CBCM). In the absence of test vectors, the best I have been able
19444 to do is check that the decrypt undoes the encrypt, so far. Send me test
19445 vectors if you have them.
19449 * Correct calculation of key length for export ciphers (too much space was
19450 allocated for null ciphers). This has not been tested!
19454 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19455 message is now correct (it understands "crypto" and "ssl" on its
19456 command line). There is also now an "update" option. This will update
19457 the util/ssleay.num and util/libeay.num files with any new functions.
19459 perl util/mkdef.pl crypto ssl update
19460 it will update them.
19464 * Overhauled the Perl interface:
19465 - ported BN stuff to OpenSSL's different BN library
19466 - made the perl/ source tree CVS-aware
19467 - renamed the package from SSLeay to OpenSSL (the files still contain
19468 their history because I've copied them in the repository)
19469 - removed obsolete files (the test scripts will be replaced
19470 by better Test::Harness variants in the future)
19472 *Ralf S. Engelschall*
19474 * First cut for a very conservative source tree cleanup:
19475 1. merge various obsolete readme texts into doc/ssleay.txt
19476 where we collect the old documents and readme texts.
19477 2. remove the first part of files where I'm already sure that we no
19478 longer need them because of three reasons: either they are just temporary
19479 files which were left by Eric or they are preserved original files where
19480 I've verified that the diff is also available in the CVS via "cvs diff
19481 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19482 the crypto/md/ stuff).
19484 *Ralf S. Engelschall*
19486 * More extension code. Incomplete support for subject and issuer alt
19487 name, issuer and authority key id. Change the i2v function parameters
19488 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19489 what that's for :-) Fix to ASN1 macro which messed up
19490 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19494 * Preliminary support for ENUMERATED type. This is largely copied from the
19499 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19501 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19503 * Make sure `make rehash` target really finds the `openssl` program.
19505 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19507 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19508 like to hear about it if this slows down other processors.
19512 * Add CygWin32 platform information to Configure script.
19514 *Alan Batie <batie@aahz.jf.intel.com>*
19516 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19518 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19520 * New program nseq to manipulate netscape certificate sequences
19524 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19529 * Fixes to BN code. Previously the default was to define BN_RECURSION
19530 but the BN code had some problems that would cause failures when
19531 doing certificate verification and some other functions.
19533 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19535 * Add ASN1 and PEM code to support netscape certificate sequences.
19539 * Add ASN1 and PEM code to support netscape certificate sequences.
19543 * Add several PKIX and private extended key usage OIDs.
19547 * Modify the 'ca' program to handle the new extension code. Modify
19548 openssl.cnf for new extension format, add comments.
19552 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19553 and add a sample to openssl.cnf so req -x509 now adds appropriate
19558 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19559 error code, add initial support to X509_print() and x509 application.
19563 * Takes a deep breath and start adding X509 V3 extension support code. Add
19564 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19565 stuff is currently isolated and isn't even compiled yet.
19569 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19570 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19571 Removed the versions check from X509 routines when loading extensions:
19572 this allows certain broken certificates that don't set the version
19573 properly to be processed.
19577 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19578 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19579 can still be regenerated with "make depend".
19583 * Spelling mistake in C version of CAST-128.
19585 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19587 * Changes to the error generation code. The perl script err-code.pl
19588 now reads in the old error codes and retains the old numbers, only
19589 adding new ones if necessary. It also only changes the .err files if new
19590 codes are added. The makefiles have been modified to only insert errors
19591 when needed (to avoid needlessly modifying header files). This is done
19592 by only inserting errors if the .err file is newer than the auto generated
19593 C file. To rebuild all the error codes from scratch (the old behaviour)
19594 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19595 or delete all the .err files.
19599 * CAST-128 was incorrectly implemented for short keys. The C version has
19600 been fixed, but is untested. The assembler versions are also fixed, but
19601 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19602 to regenerate it if needed.
19603 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19604 Hagino <itojun@kame.net>*
19606 * File was opened incorrectly in randfile.c.
19608 *Ulf Möller <ulf@fitug.de>*
19610 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19611 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19612 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19613 al: it's just almost always a UTCTime. Note this patch adds new error
19614 codes so do a "make errors" if there are problems.
19618 * Correct Linux 1 recognition in config.
19620 *Ulf Möller <ulf@fitug.de>*
19622 * Remove pointless MD5 hash when using DSA keys in ca.
19624 *Anonymous <nobody@replay.com>*
19626 * Generate an error if given an empty string as a cert directory. Also
19627 generate an error if handed NULL (previously returned 0 to indicate an
19628 error, but didn't set one).
19630 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19632 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19636 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19637 parameters. This was causing a warning which killed off the Win32 compile.
19641 * Remove C++ style comments from crypto/bn/bn_local.h.
19643 *Neil Costigan <neil.costigan@celocom.com>*
19645 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19646 based on a text string, looking up short and long names and finally
19647 "dot" format. The "dot" format stuff didn't work. Added new function
19648 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19649 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19650 OID is not part of the table.
19654 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19655 X509_LOOKUP_by_alias().
19659 * Sort openssl functions by name.
19663 * Get the `gendsa` command working and add it to the `list` command. Remove
19664 encryption from sample DSA keys (in case anyone is interested the password
19669 * Make *all* `*_free` functions accept a NULL pointer.
19671 *Frans Heymans <fheymans@isaserver.be>*
19673 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19676 *Anonymous <nobody@replay.com>*
19678 * s_server should send the CAfile as acceptable CAs, not its own cert.
19680 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19682 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19684 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19686 * Temp key "for export" tests were wrong in s3_srvr.c.
19688 *Anonymous <nobody@replay.com>*
19690 * Add prototype for temp key callback functions
19691 SSL_CTX_set_tmp_{rsa,dh}_callback().
19695 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19696 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19700 * X509_name_add_entry() freed the wrong thing after an error.
19702 *Arne Ansper <arne@ats.cyber.ee>*
19704 * rsa_eay.c would attempt to free a NULL context.
19706 *Arne Ansper <arne@ats.cyber.ee>*
19708 * BIO_s_socket() had a broken should_retry() on Windoze.
19710 *Arne Ansper <arne@ats.cyber.ee>*
19712 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19714 *Arne Ansper <arne@ats.cyber.ee>*
19716 * Make sure the already existing X509_STORE->depth variable is initialized
19717 in X509_STORE_new(), but document the fact that this variable is still
19718 unused in the certificate verification process.
19720 *Ralf S. Engelschall*
19722 * Fix the various library and `apps/` files to free up pkeys obtained from
19723 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19727 * Fix reference counting in X509_PUBKEY_get(). This makes
19728 demos/maurice/example2.c work, amongst others, probably.
19730 *Steve Henson and Ben Laurie*
19732 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19733 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19734 are no longer created. This way we have a single and consistent command
19735 line interface `openssl <command>`, similar to `cvs <command>`.
19737 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19739 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19740 BIT STRING wrapper always have zero unused bits.
19744 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19748 * Make the top-level INSTALL documentation easier to understand.
19752 * Makefiles updated to exit if an error occurs in a sub-directory
19753 make (including if user presses ^C) [Paul Sutton]
19755 * Make Montgomery context stuff explicit in RSA data structure.
19759 * Fix build order of pem and err to allow for generated pem.h.
19763 * Fix renumbering bug in X509_NAME_delete_entry().
19767 * Enhanced the err-ins.pl script so it makes the error library number
19768 global and can add a library name. This is needed for external ASN1 and
19769 other error libraries.
19773 * Fixed sk_insert which never worked properly.
19777 * Fix ASN1 macros so they can handle indefinite length constructed
19778 EXPLICIT tags. Some non standard certificates use these: they can now
19783 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19784 into a single doc/ssleay.txt bundle. This way the information is still
19785 preserved but no longer messes up this directory. Now it's new room for
19786 the new set of documentation files.
19788 *Ralf S. Engelschall*
19790 * SETs were incorrectly DER encoded. This was a major pain, because they
19791 shared code with SEQUENCEs, which aren't coded the same. This means that
19792 almost everything to do with SETs or SEQUENCEs has either changed name or
19793 number of arguments.
19795 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19797 * Fix test data to work with the above.
19801 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19802 was already fixed by Eric for 0.9.1 it seems.
19804 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19806 * Autodetect FreeBSD3.
19810 * Fix various bugs in Configure. This affects the following platforms:
19814 unixware-2.0-pentium
19819 * Eliminate generated files from CVS. Reorder tests to regenerate files
19820 before they are needed.
19824 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19828 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19830 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19831 changed SSLeay to OpenSSL in version strings.
19833 *Ralf S. Engelschall*
19835 * Some fixups to the top-level documents.
19839 * Fixed the nasty bug where rsaref.h was not found under compile-time
19840 because the symlink to include/ was missing.
19842 *Ralf S. Engelschall*
19844 * Incorporated the popular no-RSA/DSA-only patches
19845 which allow to compile an RSA-free SSLeay.
19847 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19849 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19850 when "ssleay" is still not found.
19852 *Ralf S. Engelschall*
19854 * Added more platforms to Configure: Cray T3E, HPUX 11,
19856 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19858 * Updated the README file.
19860 *Ralf S. Engelschall*
19862 * Added various .cvsignore files in the CVS repository subdirs
19863 to make a "cvs update" really silent.
19865 *Ralf S. Engelschall*
19867 * Recompiled the error-definition header files and added
19868 missing symbols to the Win32 linker tables.
19870 *Ralf S. Engelschall*
19872 * Cleaned up the top-level documents;
19873 o new files: CHANGES and LICENSE
19874 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19875 o merged COPYRIGHT into LICENSE
19876 o removed obsolete TODO file
19877 o renamed MICROSOFT to INSTALL.W32
19879 *Ralf S. Engelschall*
19881 * Removed dummy files from the 0.9.1b source tree:
19882 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19883 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19884 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19885 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19886 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19888 *Ralf S. Engelschall*
19890 * Added various platform portability fixes.
19894 * The Genesis of the OpenSSL rpject:
19895 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19896 Young and Tim J. Hudson created while they were working for C2Net until
19899 *The OpenSSL Project*
19901 ### Changes between 0.9.0b and 0.9.1b [not released]
19903 * Updated a few CA certificates under certs/
19907 * Changed some BIGNUM api stuff.
19911 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19912 DGUX x86, Linux Alpha, etc.
19916 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19917 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19922 * Add -strparse option to asn1pars program which parses nested
19925 *Dr Stephen Henson <shenson@bigfoot.com>*
19927 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19931 * DSA fix for "ca" program.
19935 * Added "-genkey" option to "dsaparam" program.
19939 * Added RIPE MD160 (rmd160) message digest.
19943 * Added -a (all) option to "ssleay version" command.
19947 * Added PLATFORM define which is the id given to Configure.
19951 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19955 * Extended the ASN.1 parser routines.
19959 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19963 * Added a BN_CTX to the BN library.
19967 * Fixed the weak key values in DES library
19971 * Changed API in EVP library for cipher aliases.
19975 * Added support for RC2/64bit cipher.
19979 * Converted the lhash library to the crypto/mem.c functions.
19983 * Added more recognized ASN.1 object ids.
19987 * Added more RSA padding checks for SSL/TLS.
19991 * Added BIO proxy/filter functionality.
19995 * Added extra_certs to SSL_CTX which can be used
19996 send extra CA certificates to the client in the CA cert chain sending
19997 process. It can be configured with SSL_CTX_add_extra_chain_cert().
20001 * Now Fortezza is denied in the authentication phase because
20002 this is key exchange mechanism is not supported by SSLeay at all.
20006 * Additional PKCS1 checks.
20010 * Support the string "TLSv1" for all TLS v1 ciphers.
20014 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
20015 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
20019 * Fixed a few memory leaks.
20023 * Fixed various code and comment typos.
20027 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
20028 bytes sent in the client random.
20030 *Edward Bishop <ebishop@spyglass.com>*
20034 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
20035 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
20036 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
20037 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
20038 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
20039 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
20040 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
20041 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
20042 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
20043 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
20044 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
20045 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
20046 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
20047 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
20048 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
20049 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
20050 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
20051 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
20052 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
20053 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
20054 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
20055 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
20056 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
20057 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
20058 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
20059 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
20060 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
20061 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
20062 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
20063 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
20064 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
20065 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
20066 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
20067 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
20068 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
20069 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
20070 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
20071 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
20072 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
20073 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
20074 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
20075 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
20076 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
20077 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
20078 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
20079 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
20080 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
20081 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
20082 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
20083 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
20084 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
20085 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
20086 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
20087 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
20088 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
20089 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
20090 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
20091 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
20092 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
20093 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
20094 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
20095 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
20096 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
20097 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
20098 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
20099 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
20100 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
20101 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
20102 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
20103 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
20104 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
20105 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
20106 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
20107 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
20108 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
20109 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
20110 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
20111 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
20112 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
20113 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
20114 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
20115 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
20116 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
20117 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
20118 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
20119 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
20120 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
20121 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
20122 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
20123 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
20124 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
20125 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
20126 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
20127 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
20128 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
20129 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
20130 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
20131 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
20132 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
20133 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
20134 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
20135 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
20136 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
20137 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
20138 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
20139 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
20140 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
20141 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
20142 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
20143 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
20144 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
20145 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20146 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20147 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20148 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20149 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20150 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20151 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20152 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20153 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20154 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20155 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20156 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20157 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20158 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20159 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20160 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20161 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20162 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20163 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20164 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20165 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20166 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20167 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20168 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20169 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20170 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20171 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20172 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20173 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20174 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20175 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20176 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20177 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20178 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20179 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20180 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20181 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20182 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20183 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20184 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20185 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20186 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20187 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20188 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20189 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20190 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20191 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20192 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20193 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20194 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20195 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20196 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20197 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20198 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20199 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20200 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20201 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20202 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20203 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20204 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20205 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20206 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20207 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20208 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20209 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20210 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20211 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20212 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20213 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655