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 * Fix excessive time spent checking DH q parameter value.
291 The function DH_check() performs various checks on DH parameters. After
292 fixing CVE-2023-3446 it was discovered that a large q parameter value can
293 also trigger an overly long computation during some of these checks.
294 A correct q value, if present, cannot be larger than the modulus p
295 parameter, thus it is unnecessary to perform these checks if q is larger
298 If DH_check() is called with such q parameter value,
299 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally
300 intensive checks are skipped.
306 * Fix DH_check() excessive time with over sized modulus.
308 The function DH_check() performs various checks on DH parameters. One of
309 those checks confirms that the modulus ("p" parameter) is not too large.
310 Trying to use a very large modulus is slow and OpenSSL will not normally use
311 a modulus which is over 10,000 bits in length.
313 However the DH_check() function checks numerous aspects of the key or
314 parameters that have been supplied. Some of those checks use the supplied
315 modulus value even if it has already been found to be too large.
317 A new limit has been added to DH_check of 32,768 bits. Supplying a
318 key/parameters with a modulus over this size will simply cause DH_check() to
325 * Do not ignore empty associated data entries with AES-SIV.
327 The AES-SIV algorithm allows for authentication of multiple associated
328 data entries along with the encryption. To authenticate empty data the
329 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
330 with NULL pointer as the output buffer and 0 as the input buffer length.
331 The AES-SIV implementation in OpenSSL just returns success for such call
332 instead of performing the associated data authentication operation.
333 The empty data thus will not be authenticated. ([CVE-2023-2975])
335 Thanks to Juerg Wullschleger (Google) for discovering the issue.
337 The fix changes the authentication tag value and the ciphertext for
338 applications that use empty associated data entries with AES-SIV.
339 To decrypt data encrypted with previous versions of OpenSSL the application
340 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
345 * When building with the `enable-fips` option and using the resulting
346 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
347 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
348 not operate with truncated digests (FIPS 140-3 IG G.R).
352 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
354 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
355 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
357 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
358 numeric text form. For gigantic sub-identifiers, this would take a very
359 long time, the time complexity being O(n^2) where n is the size of that
360 sub-identifier. ([CVE-2023-2650])
362 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
363 IDENTIFIER to canonical numeric text form if the size of that OBJECT
364 IDENTIFIER is 586 bytes or less, and fail otherwise.
366 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
367 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
368 most 128 sub-identifiers, and that the maximum value that each sub-
369 identifier may have is 2^32-1 (4294967295 decimal).
371 For each byte of every sub-identifier, only the 7 lower bits are part of
372 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
373 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
378 * Multiple algorithm implementation fixes for ARM BE platforms.
382 * Added a -pedantic option to fipsinstall that adjusts the various
383 settings to ensure strict FIPS compliance rather than backwards
388 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
389 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
390 trigger a crash of an application using AES-XTS decryption if the memory
391 just after the buffer being decrypted is not mapped.
392 Thanks to Anton Romanov (Amazon) for discovering the issue.
397 * Add FIPS provider configuration option to disallow the use of
398 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
399 The option '-no_drbg_truncated_digests' can optionally be
400 supplied to 'openssl fipsinstall'.
404 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
405 that it does not enable policy checking. Thanks to David Benjamin for
406 discovering this issue.
411 * Fixed an issue where invalid certificate policies in leaf certificates are
412 silently ignored by OpenSSL and other certificate policy checks are skipped
413 for that certificate. A malicious CA could use this to deliberately assert
414 invalid certificate policies in order to circumvent policy checking on the
415 certificate altogether.
420 * Limited the number of nodes created in a policy tree to mitigate
421 against CVE-2023-0464. The default limit is set to 1000 nodes, which
422 should be sufficient for most installations. If required, the limit
423 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
424 time define to a desired maximum number of nodes or zero to allow
430 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
432 * Add FIPS provider configuration option to enforce the
433 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
434 The option '-ems-check' can optionally be supplied to
435 'openssl fipsinstall'.
439 * The FIPS provider includes a few non-approved algorithms for
440 backward compatibility purposes and the "fips=yes" property query
441 must be used for all algorithm fetches to ensure FIPS compliance.
443 The algorithms that are included but not approved are Triple DES ECB,
444 Triple DES CBC and EdDSA.
448 * Added support for KMAC in KBKDF.
452 * RNDR and RNDRRS support in provider functions to provide
453 random number generation for Arm CPUs (aarch64).
457 * s_client and s_server apps now explicitly say when the TLS version
458 does not include the renegotiation mechanism. This avoids confusion
459 between that scenario versus when the TLS version includes secure
460 renegotiation but the peer lacks support for it.
464 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
466 *Tomasz Kantecki, Andrey Matyukov*
468 * The various OBJ_* functions have been made thread safe.
472 * Parallel dual-prime 1536/2048-bit modular exponentiation for
473 AVX512_IFMA capable processors.
475 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
477 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
478 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
479 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
480 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
481 `OPENSSL_NO_DEPRECATED_3_1`.
483 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
484 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
485 definitions for these functions regardless of whether
486 `OPENSSL_NO_DEPRECATED_3_1` is defined.
488 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
489 functions regardless of whether they are using them. It is recommended that
490 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
494 * When generating safe-prime DH parameters set the recommended private key
495 length equivalent to minimum key lengths as in RFC 7919.
499 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
500 maximum size that is smaller or equal to the digest length to comply with
501 FIPS 186-4 section 5. This is implemented by a new option
502 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
503 `rsa_pss_saltlen` parameter, which is now the default. Signature
504 verification is not affected by this change and continues to work as before.
511 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
512 listed here are only a brief description.
513 The migration guide contains more detailed information related to new features,
514 breaking changes, and mappings for the large list of deprecated functions.
516 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
518 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
520 * Fixed NULL dereference during PKCS7 data verification.
522 A NULL pointer can be dereferenced when signatures are being
523 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
524 algorithm used for the signature is known to the OpenSSL library but
525 the implementation of the hash algorithm is not available the digest
526 initialization will fail. There is a missing check for the return
527 value from the initialization function which later leads to invalid
528 usage of the digest API most likely leading to a crash.
531 PKCS7 data is processed by the SMIME library calls and also by the
532 time stamp (TS) library calls. The TLS implementation in OpenSSL does
533 not call these functions however third party applications would be
534 affected if they call these functions to verify signatures on untrusted
539 * Fixed X.400 address type confusion in X.509 GeneralName.
541 There is a type confusion vulnerability relating to X.400 address processing
542 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
543 but the public structure definition for GENERAL_NAME incorrectly specified
544 the type of the x400Address field as ASN1_TYPE. This field is subsequently
545 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
548 When CRL checking is enabled (i.e. the application sets the
549 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
550 pass arbitrary pointers to a memcmp call, enabling them to read memory
551 contents or enact a denial of service.
556 * Fixed NULL dereference validating DSA public key.
558 An invalid pointer dereference on read can be triggered when an
559 application tries to check a malformed DSA public key by the
560 EVP_PKEY_public_check() function. This will most likely lead
561 to an application crash. This function can be called on public
562 keys supplied from untrusted sources which could allow an attacker
563 to cause a denial of service attack.
565 The TLS implementation in OpenSSL does not call this function
566 but applications might call the function if there are additional
567 security requirements imposed by standards such as FIPS 140-3.
570 *Shane Lontis, Tomáš Mráz*
572 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
574 An invalid pointer dereference on read can be triggered when an
575 application tries to load malformed PKCS7 data with the
576 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
578 The result of the dereference is an application crash which could
579 lead to a denial of service attack. The TLS implementation in OpenSSL
580 does not call this function however third party applications might
581 call these functions on untrusted data.
586 * Fixed Use-after-free following BIO_new_NDEF.
588 The public API function BIO_new_NDEF is a helper function used for
589 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
590 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
591 be called directly by end user applications.
593 The function receives a BIO from the caller, prepends a new BIO_f_asn1
594 filter BIO onto the front of it to form a BIO chain, and then returns
595 the new head of the BIO chain to the caller. Under certain conditions,
596 for example if a CMS recipient public key is invalid, the new filter BIO
597 is freed and the function returns a NULL result indicating a failure.
598 However, in this case, the BIO chain is not properly cleaned up and the
599 BIO passed by the caller still retains internal pointers to the previously
600 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
601 then a use-after-free will occur. This will most likely result in a crash.
604 *Viktor Dukhovni, Matt Caswell*
606 * Fixed Double free after calling PEM_read_bio_ex.
608 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
609 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
610 data. If the function succeeds then the "name_out", "header" and "data"
611 arguments are populated with pointers to buffers containing the relevant
612 decoded data. The caller is responsible for freeing those buffers. It is
613 possible to construct a PEM file that results in 0 bytes of payload data.
614 In this case PEM_read_bio_ex() will return a failure code but will populate
615 the header argument with a pointer to a buffer that has already been freed.
616 If the caller also frees this buffer then a double free will occur. This
617 will most likely lead to a crash.
619 The functions PEM_read_bio() and PEM_read() are simple wrappers around
620 PEM_read_bio_ex() and therefore these functions are also directly affected.
622 These functions are also called indirectly by a number of other OpenSSL
623 functions including PEM_X509_INFO_read_bio_ex() and
624 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
625 internal uses of these functions are not vulnerable because the caller does
626 not free the header argument if PEM_read_bio_ex() returns a failure code.
629 *Kurt Roeckx, Matt Caswell*
631 * Fixed Timing Oracle in RSA Decryption.
633 A timing based side channel exists in the OpenSSL RSA Decryption
634 implementation which could be sufficient to recover a plaintext across
635 a network in a Bleichenbacher style attack. To achieve a successful
636 decryption an attacker would have to be able to send a very large number
637 of trial messages for decryption. The vulnerability affects all RSA padding
638 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
641 *Dmitry Belyavsky, Hubert Kario*
643 * Fixed X.509 Name Constraints Read Buffer Overflow.
645 A read buffer overrun can be triggered in X.509 certificate verification,
646 specifically in name constraint checking. The read buffer overrun might
647 result in a crash which could lead to a denial of service attack.
648 In a TLS client, this can be triggered by connecting to a malicious
649 server. In a TLS server, this can be triggered if the server requests
650 client authentication and a malicious client connects.
655 * Fixed X.509 Policy Constraints Double Locking security issue.
657 If an X.509 certificate contains a malformed policy constraint and
658 policy processing is enabled, then a write lock will be taken twice
659 recursively. On some operating systems (most widely: Windows) this
660 results in a denial of service when the affected process hangs. Policy
661 processing being enabled on a publicly facing server is not considered
662 to be a common setup.
667 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
668 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
669 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
670 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
671 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
672 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
673 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
674 for legacy EC and SM2 keys is also changed similarly to honor the
675 equivalent conversion format flag as specified in the underlying
676 `EC_KEY` object being exported to a provider, when this function is
677 called through `EVP_PKEY_export()`.
681 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
683 * Fixed two buffer overflows in punycode decoding functions.
685 A buffer overrun can be triggered in X.509 certificate verification,
686 specifically in name constraint checking. Note that this occurs after
687 certificate chain signature verification and requires either a CA to
688 have signed the malicious certificate or for the application to continue
689 certificate verification despite failure to construct a path to a trusted
692 In a TLS client, this can be triggered by connecting to a malicious
693 server. In a TLS server, this can be triggered if the server requests
694 client authentication and a malicious client connects.
696 An attacker can craft a malicious email address to overflow
697 an arbitrary number of bytes containing the `.` character (decimal 46)
698 on the stack. This buffer overflow could result in a crash (causing a
702 An attacker can craft a malicious email address to overflow four
703 attacker-controlled bytes on the stack. This buffer overflow could
704 result in a crash (causing a denial of service) or potentially remote code
705 execution depending on stack layout for any given platform/compiler.
710 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
711 parameters in OpenSSL code.
712 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
713 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
714 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
715 Using these invalid names may cause algorithms to use slower methods
716 that ignore the CRT parameters.
720 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
725 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
726 data to be signed before signing the certificate.
730 * Added RIPEMD160 to the default provider.
734 * Ensured that the key share group sent or accepted for the key exchange
735 is allowed for the protocol version.
739 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
741 * OpenSSL supports creating a custom cipher via the legacy
742 EVP_CIPHER_meth_new() function and associated function calls. This function
743 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
744 to use the new provider mechanism in order to implement custom ciphers.
746 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
747 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
748 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
749 and decryption initialisation functions). Instead of using the custom cipher
750 directly it incorrectly tries to fetch an equivalent cipher from the
751 available providers. An equivalent cipher is found based on the NID passed to
752 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
753 given cipher. However it is possible for an application to incorrectly pass
754 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
755 is used in this way the OpenSSL encryption/decryption initialisation function
756 will match the NULL cipher as being equivalent and will fetch this from the
757 available providers. This will succeed if the default provider has been
758 loaded (or if a third party provider has been loaded that offers this
759 cipher). Using the NULL cipher means that the plaintext is emitted as the
762 Applications are only affected by this issue if they call
763 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
764 encryption/decryption initialisation function. Applications that only use
765 SSL/TLS are not impacted by this issue.
770 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
775 * Fixed the linux-mips64 Configure target which was missing the
776 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
781 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
786 * Correctly handle a retransmitted ClientHello in DTLS
790 * Fixed detection of ktls support in cross-compile environment on Linux
794 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
799 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
800 report correct results in some cases
804 * Fix UWP builds by defining VirtualLock
808 * For known safe primes use the minimum key length according to RFC 7919.
809 Longer private key sizes unnecessarily raise the cycles needed to compute the
810 shared secret without any increase of the real security. This fixes a
811 regression from 1.1.1 where these shorter keys were generated for the known
816 * Added the loongarch64 target
820 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
821 only passed to the FIPS provider and not to the default or legacy provider.
825 * Fixed reported performance degradation on aarch64. Restored the
826 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
827 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
828 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
829 The new algorithm is still used for 32 bit targets.
833 * Added a missing header for memcmp that caused compilation failure on some
838 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
840 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
841 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
842 This issue makes the RSA implementation with 2048 bit private keys
843 incorrect on such machines and memory corruption will happen during
844 the computation. As a consequence of the memory corruption an attacker
845 may be able to trigger a remote code execution on the machine performing
848 SSL/TLS servers or other servers using 2048 bit RSA private keys running
849 on machines supporting AVX512IFMA instructions of the X86_64 architecture
850 are affected by this issue.
855 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
856 implementation would not encrypt the entirety of the data under some
857 circumstances. This could reveal sixteen bytes of data that was
858 preexisting in the memory that wasn't written. In the special case of
859 "in place" encryption, sixteen bytes of the plaintext would be revealed.
861 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
862 they are both unaffected.
865 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
867 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
869 * In addition to the c_rehash shell command injection identified in
870 CVE-2022-1292, further bugs where the c_rehash script does not
871 properly sanitise shell metacharacters to prevent command injection have been
874 When the CVE-2022-1292 was fixed it was not discovered that there
875 are other places in the script where the file names of certificates
876 being hashed were possibly passed to a command executed through the shell.
878 This script is distributed by some operating systems in a manner where
879 it is automatically executed. On such operating systems, an attacker
880 could execute arbitrary commands with the privileges of the script.
882 Use of the c_rehash script is considered obsolete and should be replaced
883 by the OpenSSL rehash command line tool.
886 *Daniel Fiala, Tomáš Mráz*
888 * Case insensitive string comparison no longer uses locales. It has instead
889 been directly implemented.
893 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
895 * Case insensitive string comparison is reimplemented via new locale-agnostic
896 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
897 comparison. The previous implementation had problems when the Turkish locale
902 * Fixed a bug in the c_rehash script which was not properly sanitising shell
903 metacharacters to prevent command injection. This script is distributed by
904 some operating systems in a manner where it is automatically executed. On
905 such operating systems, an attacker could execute arbitrary commands with the
906 privileges of the script.
908 Use of the c_rehash script is considered obsolete and should be replaced
909 by the OpenSSL rehash command line tool.
914 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
915 certificate on an OCSP response. The bug caused the function in the case
916 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
917 response (meaning a successful verification) even in the case where the
918 response signing certificate fails to verify.
920 It is anticipated that most users of `OCSP_basic_verify` will not use the
921 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
922 a negative value (indicating a fatal error) in the case of a certificate
923 verification failure. The normal expected return value in this case would be
926 This issue also impacts the command line OpenSSL "ocsp" application. When
927 verifying an ocsp response with the "-no_cert_checks" option the command line
928 application will report that the verification is successful even though it
929 has in fact failed. In this case the incorrect successful response will also
930 be accompanied by error messages showing the failure and contradicting the
931 apparently successful result.
936 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
937 AAD data as the MAC key. This made the MAC key trivially predictable.
939 An attacker could exploit this issue by performing a man-in-the-middle attack
940 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
941 that the modified data would still pass the MAC integrity check.
943 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
944 endpoint will always be rejected by the recipient and the connection will
945 fail at that point. Many application protocols require data to be sent from
946 the client to the server first. Therefore, in such a case, only an OpenSSL
947 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
949 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
950 sent in both directions. In this case both clients and servers could be
951 affected, regardless of the application protocol.
953 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
954 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
955 the handshake when using this ciphersuite.
957 The confidentiality of data is not impacted by this issue, i.e. an attacker
958 cannot decrypt data that has been encrypted using this ciphersuite - they can
961 In order for this attack to work both endpoints must legitimately negotiate
962 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
963 OpenSSL 3.0, and is not available within the default provider or the default
964 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
965 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
966 following must have occurred:
968 1) OpenSSL must have been compiled with the (non-default) compile time option
969 enable-weak-ssl-ciphers
971 2) OpenSSL must have had the legacy provider explicitly loaded (either
972 through application code or via configuration)
974 3) The ciphersuite must have been explicitly added to the ciphersuite list
976 4) The libssl security level must have been set to 0 (default is 1)
978 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
980 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
981 others that both endpoints have in common
986 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
987 occupied by the removed hash table entries.
989 This function is used when decoding certificates or keys. If a long lived
990 process periodically decodes certificates or keys its memory usage will
991 expand without bounds and the process might be terminated by the operating
992 system causing a denial of service. Also traversing the empty hash table
993 entries will take increasingly more time.
995 Typically such long lived processes might be TLS clients or TLS servers
996 configured to accept client certificate authentication.
999 *Hugo Landau, Aliaksei Levin*
1001 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
1002 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
1003 statistics are no longer supported. For compatibility, these statistics are
1004 still listed in the output but are now always reported as zero.
1008 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
1010 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
1011 for non-prime moduli.
1013 Internally this function is used when parsing certificates that contain
1014 elliptic curve public keys in compressed form or explicit elliptic curve
1015 parameters with a base point encoded in compressed form.
1017 It is possible to trigger the infinite loop by crafting a certificate that
1018 has invalid explicit curve parameters.
1020 Since certificate parsing happens prior to verification of the certificate
1021 signature, any process that parses an externally supplied certificate may thus
1022 be subject to a denial of service attack. The infinite loop can also be
1023 reached when parsing crafted private keys as they can contain explicit
1024 elliptic curve parameters.
1026 Thus vulnerable situations include:
1028 - TLS clients consuming server certificates
1029 - TLS servers consuming client certificates
1030 - Hosting providers taking certificates or private keys from customers
1031 - Certificate authorities parsing certification requests from subscribers
1032 - Anything else which parses ASN.1 elliptic curve parameters
1034 Also any other applications that use the BN_mod_sqrt() where the attacker
1035 can control the parameter values are vulnerable to this DoS issue.
1040 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
1041 to the list of ciphersuites providing Perfect Forward Secrecy as
1042 required by SECLEVEL >= 3.
1044 *Dmitry Belyavskiy, Nicola Tuveri*
1046 * Made the AES constant time code for no-asm configurations
1047 optional due to the resulting 95% performance degradation.
1048 The AES constant time code can be enabled, for no assembly
1049 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1053 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
1058 * The negative return value handling of the certificate verification callback
1059 was reverted. The replacement is to set the verification retry state with
1060 the SSL_set_retry_verify() function.
1064 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
1066 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
1067 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
1068 verify a certificate supplied by a server. That function may return a
1069 negative return value to indicate an internal error (for example out of
1070 memory). Such a negative return value is mishandled by OpenSSL and will cause
1071 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
1072 success and a subsequent call to SSL_get_error() to return the value
1073 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
1074 returned by OpenSSL if the application has previously called
1075 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
1076 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
1077 totally unexpected and applications may not behave correctly as a result. The
1078 exact behaviour will depend on the application but it could result in
1079 crashes, infinite loops or other similar incorrect responses.
1081 This issue is made more serious in combination with a separate bug in OpenSSL
1082 3.0 that will cause X509_verify_cert() to indicate an internal error when
1083 processing a certificate chain. This will occur where a certificate does not
1084 include the Subject Alternative Name extension but where a Certificate
1085 Authority has enforced name constraints. This issue can occur even with valid
1091 * Corrected a few file name and file reference bugs in the build,
1092 installation and setup scripts, which lead to installation verification
1093 failures. Slightly enhanced the installation verification script.
1097 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
1102 * Fixed PVK encoder to properly query for the passphrase.
1106 * Multiple fixes in the OSSL_HTTP API functions.
1110 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
1111 OSSL_PARAM_INTEGER data type and return error on negative numbers
1112 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
1113 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
1117 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
1121 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
1125 * Multiple threading fixes.
1129 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
1133 * Allow fetching an operation from the provider that owns an unexportable key
1134 as a fallback if that is still allowed by the property query.
1138 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
1140 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
1145 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
1146 S390X capability vector to zero. This simplifies testing of different code
1147 paths on S390X architecture.
1151 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
1152 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
1153 SP 800-38D". The communication will fail at this point.
1157 * The EC_GROUP_clear_free() function is deprecated as there is nothing
1158 confidential in EC_GROUP data.
1162 * The byte order mark (BOM) character is ignored if encountered at the
1163 beginning of a PEM-formatted file.
1167 * Added CMS support for the Russian GOST algorithms.
1171 * Due to move of the implementation of cryptographic operations
1172 to the providers, validation of various operation parameters can
1173 be postponed until the actual operation is executed where previously
1174 it happened immediately when an operation parameter was set.
1176 For example when setting an unsupported curve with
1177 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
1178 fail but later keygen operations with the EVP_PKEY_CTX will fail.
1180 *OpenSSL team members and many third party contributors*
1182 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
1183 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
1184 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1185 instead to retrieve these algorithms from a provider.
1189 * On build targets where the multilib postfix is set in the build
1190 configuration the libdir directory was changing based on whether
1191 the lib directory with the multilib postfix exists on the system
1192 or not. This unpredictable behavior was removed and eventual
1193 multilib postfix is now always added to the default libdir. Use
1194 `--libdir=lib` to override the libdir if adding the postfix is
1199 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1200 no longer interoperable with OpenSSL 1.1.1.
1204 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1205 function codes, this function can only cause problems for calling
1210 * Add a configurable flag to output date formats as ISO 8601. Does not
1211 change the default date format.
1215 * Version of MSVC earlier than 1300 could get link warnings, which could
1216 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1217 Support for this flag has been removed.
1221 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1222 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1223 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1224 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1225 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1229 * The signatures of the functions to get and set options on SSL and
1230 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1231 Some source code changes may be required.
1235 * The public definitions of conf_method_st and conf_st have been
1236 deprecated. They will be made opaque in a future release.
1238 *Rich Salz and Tomáš Mráz*
1240 * Client-initiated renegotiation is disabled by default. To allow it, use
1241 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1242 flag, or the "ClientRenegotiation" config parameter as appropriate.
1246 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1247 or modify relative pathname inclusion.
1251 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1252 validated. Please consult the README-FIPS and
1253 README-PROVIDERS files, as well as the migration guide.
1255 *OpenSSL team members and many third party contributors*
1257 * For the key types DH and DHX the allowed settable parameters are now different.
1261 * The openssl commands that read keys, certificates, and CRLs now
1262 automatically detect the PEM or DER format of the input files.
1264 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1266 * Added enhanced PKCS#12 APIs which accept a library context.
1270 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1274 * Added support for Kernel TLS (KTLS).
1276 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1278 * Support for RFC 5746 secure renegotiation is now required by default for
1279 SSL or TLS connections to succeed.
1283 * The signature of the `copy` functional parameter of the
1284 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1285 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1286 the signature of the `pub_decode` functional parameter of the
1287 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1288 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1292 * The error return values from some control calls (ctrl) have changed.
1296 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1300 * Many functions in the EVP_ namespace that are getters of values from
1301 implementations or contexts were renamed to include get or get0 in their
1302 names. Old names are provided as macro aliases for compatibility and
1307 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1308 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1309 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1314 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1317 * The output from the command line applications may have minor
1322 * The output from numerous "printing" may have minor changes.
1326 * Windows thread synchronization uses read/write primitives (SRWLock) when
1327 supported by the OS, otherwise CriticalSection continues to be used.
1331 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1332 work on read only BIO source/sinks that do not support these functions.
1333 This allows piping or redirection of a file BIO using stdin to be buffered
1334 into memory. This is used internally in OSSL_DECODER_from_bio().
1338 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1339 this function would return one of the values OSSL_STORE_INFO_NAME,
1340 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1341 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1342 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1343 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1344 using this function should be amended to handle the changed return value.
1348 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1349 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1350 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1351 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1352 contains more than one certificate identifier: This means that all
1353 certificates referenced there MUST be part of the validation chain.
1357 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1358 RC5, DESX and DES have been moved to the legacy provider.
1362 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1363 RIPEMD-160 have been moved to the legacy provider.
1367 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1372 * The deprecated functions EVP_PKEY_get0_RSA(),
1373 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1374 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1375 well as the similarly named "get1" functions behave differently in
1380 * A number of functions handling low-level keys or engines were deprecated
1381 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1382 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1383 EVP_PKEY_get0_siphash().
1387 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1388 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1389 will need to load the legacy crypto provider. This includes these PBE
1390 algorithms which use this KDF:
1391 - NID_pbeWithMD2AndDES_CBC
1392 - NID_pbeWithMD5AndDES_CBC
1393 - NID_pbeWithSHA1AndRC2_CBC
1394 - NID_pbeWithMD2AndRC2_CBC
1395 - NID_pbeWithMD5AndRC2_CBC
1396 - NID_pbeWithSHA1AndDES_CBC
1400 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1401 BIO_debug_callback() functions.
1405 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1406 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1410 * The RAND_METHOD APIs have been deprecated.
1414 * The SRP APIs have been deprecated.
1418 * Add a compile time option to prevent the caching of provider fetched
1419 algorithms. This is enabled by including the no-cached-fetch option
1420 at configuration time.
1424 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1425 count of PKCS12_DEFAULT_ITER.
1427 *Tomáš Mráz and Sahana Prasad*
1429 * The openssl speed command does not use low-level API calls anymore.
1433 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1436 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1438 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1442 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1443 providers may supply their own group implementations (using either the "key
1444 exchange" or the "key encapsulation" methods) which will automatically be
1445 detected and used by libssl.
1447 *Matt Caswell, Nicola Tuveri*
1449 * The undocumented function X509_certificate_type() has been deprecated;
1453 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1457 * Removed RSA padding mode for SSLv23 (which was only used for
1458 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1459 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1464 * Deprecated the obsolete X9.31 RSA key generation related functions.
1466 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1467 is not allowed to return a value > 1, this is no more taken as failure.
1469 *Viktor Dukhovni and David von Oheimb*
1471 * Deprecated the obsolete X9.31 RSA key generation related functions
1472 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1473 BN_X931_generate_prime_ex().
1477 * The default key generation method for the regular 2-prime RSA keys was
1478 changed to the FIPS 186-4 B.3.6 method.
1482 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1486 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1490 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1491 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1493 *Rich Salz, Richard Levitte, and David von Oheimb*
1495 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1499 * Deprecated `OCSP_parse_url()`.
1503 * Validation of SM2 keys has been separated from the validation of regular EC
1508 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1509 switches: a validation failure triggers an early exit, returning a failure
1510 exit status to the parent process.
1514 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1515 to ignore unknown ciphers.
1519 * The `-cipher-commands` and `-digest-commands` options
1520 of the command line utility `list` have been deprecated.
1521 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1525 * Added convenience functions for generating asymmetric key pairs:
1526 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1527 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1531 * All of the low-level EC_KEY functions have been deprecated.
1533 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1535 * Deprecated all the libcrypto and libssl error string loading
1540 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1541 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1546 * The `-crypt` option to the `passwd` command line tool has been removed.
1550 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1555 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1559 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1560 EVP_PKEY_get1_tls_encodedpoint().
1564 * The security callback, which can be customised by application code, supports
1565 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1566 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1570 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1571 interface. Their functionality remains unchanged.
1575 * Added new option for 'openssl list', '-providers', which will display the
1576 list of loaded providers, their names, version and status. It optionally
1577 displays their gettable parameters.
1581 * Removed EVP_PKEY_set_alias_type().
1585 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1586 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1590 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1591 parameters are still checked for type safety at compile time via helper
1596 * Remove the RAND_DRBG API
1598 *Paul Dale and Matthias St. Pierre*
1600 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1601 as well as actual hostnames.
1605 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1606 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1607 conversely, silently ignore DTLS protocol version bounds when configuring
1608 TLS-based contexts. The commands can be repeated to set bounds of both
1609 types. The same applies with the corresponding "min_protocol" and
1610 "max_protocol" command-line switches, in case some application uses both TLS
1613 SSL_CTX instances that are created for a fixed protocol version (e.g.
1614 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1615 attempts to apply bounds to these protocol versions would result in an
1616 error. Now only the "version-flexible" SSL_CTX instances are subject to
1617 limits in configuration files in command-line options.
1621 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1626 * Reworked the recorded ERR codes to make better space for system errors.
1627 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1628 given code is a system error (true) or an OpenSSL error (false).
1632 * Reworked the test perl framework to better allow parallel testing.
1634 *Nicola Tuveri and David von Oheimb*
1636 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1637 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1641 * 'Configure' has been changed to figure out the configuration target if
1642 none is given on the command line. Consequently, the 'config' script is
1643 now only a mere wrapper. All documentation is changed to only mention
1646 *Rich Salz and Richard Levitte*
1648 * Added a library context `OSSL_LIB_CTX` that applications as well as
1649 other libraries can use to form a separate context within which
1650 libcrypto operations are performed.
1654 * Added various `_ex` functions to the OpenSSL API that support using
1655 a non-default `OSSL_LIB_CTX`.
1659 * Handshake now fails if Extended Master Secret extension is dropped
1664 * Dropped interactive mode from the `openssl` program.
1668 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1670 *David von Oheimb and Shane Lontis*
1672 * Deprecated `EC_METHOD_get_field_type()`.
1676 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1677 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1678 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1682 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1686 * Add CAdES-BES signature verification support, mostly derived
1687 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1689 *Filipe Raimundo da Silva*
1691 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1695 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1696 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1700 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1704 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1705 EC_KEY_precompute_mult().
1709 * Deprecated EC_POINTs_mul().
1713 * Removed FIPS_mode() and FIPS_mode_set().
1717 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1721 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1722 EC_POINT_get_Jprojective_coordinates_GFp().
1726 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1727 arrays to be more easily constructed via a series of utility functions.
1728 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1729 the various push functions and finally convert to a passable OSSL_PARAM
1730 array using OSSL_PARAM_BLD_to_param().
1734 * The security strength of SHA1 and MD5 based signatures in TLS has been
1739 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1740 contain a provider side internal key.
1744 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1748 * Project text documents not yet having a proper file name extension
1749 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1750 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1754 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1755 have been converted to Markdown with the goal to produce documents
1756 which not only look pretty when viewed online in the browser, but
1757 remain well readable inside a plain text editor.
1759 To achieve this goal, a 'minimalistic' Markdown style has been applied
1760 which avoids formatting elements that interfere too much with the
1761 reading flow in the text file. For example, it
1763 * avoids [ATX headings][] and uses [setext headings][] instead
1764 (which works for `<h1>` and `<h2>` headings only).
1765 * avoids [inline links][] and uses [reference links][] instead.
1766 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1768 [ATX headings]: https://github.github.com/gfm/#atx-headings
1769 [setext headings]: https://github.github.com/gfm/#setext-headings
1770 [inline links]: https://github.github.com/gfm/#inline-link
1771 [reference links]: https://github.github.com/gfm/#reference-link
1772 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1773 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1775 *Matthias St. Pierre*
1777 * The test suite is changed to preserve results of each test recipe.
1778 A new directory test-runs/ with subdirectories named like the
1779 test recipes are created in the build tree for this purpose.
1783 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1784 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1785 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1787 *David von Oheimb, Martin Peylo*
1789 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1790 It supports arbitrary request and response content types, GET redirection,
1791 TLS, connections via HTTP(S) proxies, connections and exchange via
1792 user-defined BIOs (allowing implicit connections), persistent connections,
1793 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1794 The legacy OCSP-focused (and only partly documented) API
1795 is retained for backward compatibility, while most of it is deprecated.
1799 * Added `util/check-format.pl`, a tool for checking adherence to the
1800 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1801 The checks performed are incomplete and yield some false positives.
1802 Still the tool should be useful for detecting most typical glitches.
1806 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1807 If domain name resolution yields multiple IP addresses all of them are tried
1808 after `connect()` failures.
1812 * All of the low-level RSA functions have been deprecated.
1816 * X509 certificates signed using SHA1 are no longer allowed at security
1821 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1822 modified to use PKEY APIs. These commands are now in maintenance mode
1823 and no new features will be added to them.
1827 * The command line utility rsautl has been deprecated.
1831 * The command line utilities genrsa and rsa have been modified to use PKEY
1832 APIs. They now write PKCS#8 keys by default. These commands are now in
1833 maintenance mode and no new features will be added to them.
1837 * All of the low-level DH functions have been deprecated.
1839 *Paul Dale and Matt Caswell*
1841 * All of the low-level DSA functions have been deprecated.
1845 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1846 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1850 * Deprecated low-level ECDH and ECDSA functions.
1854 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1858 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1859 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1860 a new formulation to include all the things it can be used for,
1861 as well as words of caution.
1865 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1869 * All of the low-level HMAC functions have been deprecated.
1871 *Paul Dale and David von Oheimb*
1873 * Over two thousand fixes were made to the documentation, including:
1874 - Common options (such as -rand/-writerand, TLS version control, etc)
1875 were refactored and point to newly-enhanced descriptions in openssl.pod.
1876 - Added style conformance for all options (with help from Richard Levitte),
1877 documented all reported missing options, added a CI build to check
1878 that all options are documented and that no unimplemented options
1880 - Documented some internals, such as all use of environment variables.
1881 - Addressed all internal broken L<> references.
1885 * All of the low-level CMAC functions have been deprecated.
1889 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1890 functions have been deprecated.
1892 *Paul Dale and David von Oheimb*
1894 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1895 set of functions. The documentation mentioned negative values for some
1896 errors, but this was never the case, so the mention of negative values
1899 Code that followed the documentation and thereby check with something
1900 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1904 * All of the low-level cipher functions have been deprecated.
1906 *Matt Caswell and Paul Dale*
1908 * Removed include/openssl/opensslconf.h.in and replaced it with
1909 include/openssl/configuration.h.in, which differs in not including
1910 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1911 was added to include both.
1913 This allows internal hacks where one might need to modify the set
1914 of configured macros, for example this if deprecated symbols are
1915 still supposed to be available internally:
1917 #include <openssl/configuration.h>
1919 #undef OPENSSL_NO_DEPRECATED
1920 #define OPENSSL_SUPPRESS_DEPRECATED
1922 #include <openssl/macros.h>
1924 This should not be used by applications that use the exported
1925 symbols, as that will lead to linking errors.
1929 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1930 used in exponentiation with 512-bit moduli. No EC algorithms are
1931 affected. Analysis suggests that attacks against 2-prime RSA1024,
1932 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1933 difficult to perform and are not believed likely. Attacks against DH512
1934 are considered just feasible. However, for an attack the target would
1935 have to reuse the DH512 private key, which is not recommended anyway.
1936 Also applications directly using the low-level API BN_mod_exp may be
1937 affected if they use BN_FLG_CONSTTIME.
1942 * Most memory-debug features have been deprecated, and the functionality
1943 replaced with no-ops.
1947 * Added documentation for the STACK API.
1951 * Introduced a new method type and API, OSSL_ENCODER, to represent
1952 generic encoders. These do the same sort of job that PEM writers
1953 and d2i functions do, but with support for methods supplied by
1954 providers, and the possibility for providers to support other
1959 * Introduced a new method type and API, OSSL_DECODER, to represent
1960 generic decoders. These do the same sort of job that PEM readers
1961 and i2d functions do, but with support for methods supplied by
1962 providers, and the possibility for providers to support other
1967 * Added a .pragma directive to the syntax of configuration files, to
1968 allow varying behavior in a supported and predictable manner.
1969 Currently added pragma:
1973 This allows dollar signs to be a keyword character unless it's
1974 followed by a opening brace or parenthesis. This is useful for
1975 platforms where dollar signs are commonly used in names, such as
1976 volume names and system directory names on VMS.
1980 * Added functionality to create an EVP_PKEY from user data.
1984 * Change the interpretation of the '--api' configuration option to
1985 mean that this is a desired API compatibility level with no
1986 further meaning. The previous interpretation, that this would
1987 also mean to remove all deprecated symbols up to and including
1988 the given version, no requires that 'no-deprecated' is also used
1989 in the configuration.
1991 When building applications, the desired API compatibility level
1992 can be set with the OPENSSL_API_COMPAT macro like before. For
1993 API compatibility version below 3.0, the old style numerical
1994 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1995 For version 3.0 and on, the value is expected to be the decimal
1996 value calculated from the major and minor version like this:
1998 MAJOR * 10000 + MINOR * 100
2002 -DOPENSSL_API_COMPAT=30000 For 3.0
2003 -DOPENSSL_API_COMPAT=30200 For 3.2
2005 To hide declarations that are deprecated up to and including the
2006 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
2007 given when building the application as well.
2011 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
2012 access to certificate and CRL stores via URIs and OSSL_STORE
2015 This adds the following functions:
2017 - X509_LOOKUP_store()
2018 - X509_STORE_load_file()
2019 - X509_STORE_load_path()
2020 - X509_STORE_load_store()
2021 - SSL_add_store_cert_subjects_to_stack()
2022 - SSL_CTX_set_default_verify_store()
2023 - SSL_CTX_load_verify_file()
2024 - SSL_CTX_load_verify_dir()
2025 - SSL_CTX_load_verify_store()
2029 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2030 The presence of this system service is determined at run-time.
2034 * Added functionality to create an EVP_PKEY context based on data
2035 for methods from providers. This takes an algorithm name and a
2036 property query string and simply stores them, with the intent
2037 that any operation that uses this context will use those strings
2038 to fetch the needed methods implicitly, thereby making the port
2039 of application written for pre-3.0 OpenSSL easier.
2043 * The undocumented function NCONF_WIN32() has been deprecated; for
2044 conversion details see the HISTORY section of doc/man5/config.pod
2048 * Introduced the new functions EVP_DigestSignInit_ex() and
2049 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
2050 EVP_DigestVerifyUpdate() have been converted to functions. See the man
2051 pages for further details.
2055 * Over two thousand fixes were made to the documentation, including:
2056 adding missing command flags, better style conformance, documentation
2059 *Rich Salz, Richard Levitte*
2061 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
2062 X25519, X448, Ed25519 and Ed448.
2066 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2071 * Deprecated the public definition of `ERR_STATE` as well as the function
2072 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
2077 * Added ERR functionality to give callers access to the stored function
2078 names that have replaced the older function code based functions.
2080 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
2081 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
2082 ERR_peek_error_all() and ERR_peek_last_error_all().
2084 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
2085 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
2086 ERR_func_error_string().
2090 * Extended testing to be verbose for failing tests only. The make variables
2091 VERBOSE_FAILURE or VF can be used to enable this:
2093 $ make VF=1 test # Unix
2094 $ mms /macro=(VF=1) test ! OpenVMS
2095 $ nmake VF=1 test # Windows
2099 * Added the `-copy_extensions` option to the `x509` command for use with
2100 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
2101 all extensions in the request are copied to the certificate or vice versa.
2103 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
2105 * Added the `-copy_extensions` option to the `req` command for use with
2106 `-x509`. When given with the `copy` or `copyall` argument,
2107 all extensions in the certification request are copied to the certificate.
2111 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
2112 they generate are by default RFC 5280 compliant in the following sense:
2113 There is a subjectKeyIdentifier extension with a hash value of the public key
2114 and for not self-signed certs there is an authorityKeyIdentifier extension
2115 with a keyIdentifier field or issuer information identifying the signing key.
2116 This is done unless some configuration overrides the new default behavior,
2117 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
2121 * Added several checks to `X509_verify_cert()` according to requirements in
2122 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
2123 (which may be done by using the CLI option `-x509_strict`):
2124 * The basicConstraints of CA certificates must be marked critical.
2125 * CA certificates must explicitly include the keyUsage extension.
2126 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
2127 * The issuer name of any certificate must not be empty.
2128 * The subject name of CA certs, certs with keyUsage crlSign,
2129 and certs without subjectAlternativeName must not be empty.
2130 * If a subjectAlternativeName extension is given it must not be empty.
2131 * The signatureAlgorithm field and the cert signature must be consistent.
2132 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
2133 must not be marked critical.
2134 * The authorityKeyIdentifier must be given for X.509v3 certs
2135 unless they are self-signed.
2136 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
2140 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
2141 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
2145 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2146 used even when parsing explicit parameters, when loading a encoded key
2147 or calling `EC_GROUP_new_from_ecpkparameters()`/
2148 `EC_GROUP_new_from_ecparameters()`.
2149 This prevents bypass of security hardening and performance gains,
2150 especially for curves with specialized EC_METHODs.
2151 By default, if a key encoded with explicit parameters is loaded and later
2152 encoded, the output is still encoded with explicit parameters, even if
2153 internally a "named" EC_GROUP is used for computation.
2157 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2158 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2159 NULL. After this change, only the cofactor parameter can be NULL. It also
2160 does some minimal sanity checks on the passed order.
2165 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2166 An attack is simple, if the first CMS_recipientInfo is valid but the
2167 second CMS_recipientInfo is chosen ciphertext. If the second
2168 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2169 encryption key will be replaced by garbage, and the message cannot be
2170 decoded, but if the RSA decryption fails, the correct encryption key is
2171 used and the recipient will not notice the attack.
2172 As a work around for this potential attack the length of the decrypted
2173 key must be equal to the cipher default key length, in case the
2174 certificate is not given and all recipientInfo are tried out.
2175 The old behaviour can be re-enabled in the CMS code by setting the
2176 CMS_DEBUG_DECRYPT flag.
2180 * Early start up entropy quality from the DEVRANDOM seed source has been
2181 improved for older Linux systems. The RAND subsystem will wait for
2182 /dev/random to be producing output before seeding from /dev/urandom.
2183 The seeded state is stored for future library initialisations using
2184 a system global shared memory segment. The shared memory identifier
2185 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2186 the desired value. The default identifier is 114.
2190 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2191 when primes for RSA keys are computed.
2192 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2193 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2194 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2195 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2196 This avoids possible fingerprinting of newly generated RSA modules.
2200 * Correct the extended master secret constant on EBCDIC systems. Without this
2201 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2202 negotiate EMS will fail. Unfortunately this also means that TLS connections
2203 between EBCDIC systems with this fix, and EBCDIC systems without this
2204 fix will fail if they negotiate EMS.
2208 * Changed the library initialisation so that the config file is now loaded
2209 by default. This was already the case for libssl. It now occurs for both
2210 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2211 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2215 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2216 where the former acts as a replacement for `ERR_put_error()`, and the
2217 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2218 `ERR_raise_data()` adds more flexibility by taking a format string and
2219 an arbitrary number of arguments following it, to be processed with
2224 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2225 to check if a named provider is loaded and available. When called, it
2226 will also activate all fallback providers if such are still present.
2230 * Enforce a minimum DH modulus size of 512 bits.
2234 * Changed DH parameters to generate the order q subgroup instead of 2q.
2235 Previously generated DH parameters are still accepted by DH_check
2236 but DH_generate_key works around that by clearing bit 0 of the
2237 private key for those. This avoids leaking bit 0 of the private key.
2241 * Significantly reduce secure memory usage by the randomness pools.
2245 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2250 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2251 algorithms. An implementation of a key exchange algorithm can be obtained
2252 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2253 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2254 the older EVP_PKEY_derive_init() function. See the man pages for the new
2255 functions for further details.
2259 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2263 * Removed the function names from error messages and deprecated the
2268 * Removed NextStep support and the macro OPENSSL_UNISTD
2272 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2273 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2274 Also removed "export var as function" capability; we do not export
2275 variables, only functions.
2279 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2280 an error and 1 indicating success. In previous versions of OpenSSL this
2281 was a void type. If a key was set longer than the maximum possible this
2286 * Support SM2 signing and verification schemes with X509 certificate.
2290 * Use SHA256 as the default digest for TS query in the `ts` app.
2294 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2298 * Default cipher lists/suites are now available via a function, the
2299 #defines are deprecated.
2303 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2304 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2305 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2309 * Join the directories crypto/x509 and crypto/x509v3
2313 * Added command 'openssl kdf' that uses the EVP_KDF API.
2317 * Added command 'openssl mac' that uses the EVP_MAC API.
2321 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2322 as default directories. Also added the command 'openssl info'
2323 for scripting purposes.
2327 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2332 * Add prediction resistance to the DRBG reseeding process.
2336 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2337 mandated by IEEE Std 1619-2018.
2341 * Added newline escaping functionality to a filename when using openssl dgst.
2342 This output format is to replicate the output format found in the `*sum`
2343 checksum programs. This aims to preserve backward compatibility.
2345 *Matt Eaton, Richard Levitte, and Paul Dale*
2347 * Removed the heartbeat message in DTLS feature, as it has very
2348 little usage and doesn't seem to fulfill a valuable purpose.
2349 The configuration option is now deprecated.
2353 * Changed the output of 'openssl {digestname} < file' to display the
2354 digest name in its output.
2358 * Added a new generic trace API which provides support for enabling
2359 instrumentation through trace output.
2361 *Richard Levitte & Matthias St. Pierre*
2363 * Added build tests for C++. These are generated files that only do one
2364 thing, to include one public OpenSSL head file each. This tests that
2365 the public header files can be usefully included in a C++ application.
2367 This test isn't enabled by default. It can be enabled with the option
2368 'enable-buildtest-c++'.
2372 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2376 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2380 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2384 * Added KMAC to EVP_MAC.
2388 * Added property based algorithm implementation selection framework to
2393 * Added SCA hardening for modular field inversion in EC_GROUP through
2394 a new dedicated field_inv() pointer in EC_METHOD.
2395 This also addresses a leakage affecting conversions from projective
2396 to affine coordinates.
2398 *Billy Bob Brumley, Nicola Tuveri*
2400 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2401 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2402 those algorithms that were already supported through the EVP_PKEY API
2403 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2404 and scrypt are now wrappers that call EVP_KDF.
2408 * Build devcrypto engine as a dynamic engine.
2410 *Eneas U de Queiroz*
2412 * Add keyed BLAKE2 to EVP_MAC.
2416 * Fix a bug in the computation of the endpoint-pair shared secret used
2417 by DTLS over SCTP. This breaks interoperability with older versions
2418 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2419 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2420 interoperability with such broken implementations. However, enabling
2421 this switch breaks interoperability with correct implementations.
2423 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2424 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2428 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2432 * Changed the license to the Apache License v2.0.
2436 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2438 - Major releases (indicated by incrementing the MAJOR release number)
2439 may introduce incompatible API/ABI changes.
2440 - Minor releases (indicated by incrementing the MINOR release number)
2441 may introduce new features but retain API/ABI compatibility.
2442 - Patch releases (indicated by incrementing the PATCH number)
2443 are intended for bug fixes and other improvements of existing
2444 features only (like improving performance or adding documentation)
2445 and retain API/ABI compatibility.
2449 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2453 * Remove the 'dist' target and add a tarball building script. The
2454 'dist' target has fallen out of use, and it shouldn't be
2455 necessary to configure just to create a source distribution.
2459 * Recreate the OS390-Unix config target. It no longer relies on a
2460 special script like it did for OpenSSL pre-1.1.0.
2464 * Instead of having the source directories listed in Configure, add
2465 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2470 * Add GMAC to EVP_MAC.
2474 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2478 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2479 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2480 to facilitate the continued use of MACs through raw private keys in
2481 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2485 * Deprecate ECDH_KDF_X9_62().
2489 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2490 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2491 are retained for backwards compatibility.
2495 * AES-XTS mode now enforces that its two keys are different to mitigate
2496 the attacked described in "Efficient Instantiations of Tweakable
2497 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2498 Details of this attack can be obtained from:
2499 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2503 * Rename the object files, i.e. give them other names than in previous
2504 versions. Their names now include the name of the final product, as
2505 well as its type mnemonic (bin, lib, shlib).
2509 * Added new option for 'openssl list', '-objects', which will display the
2510 list of built in objects, i.e. OIDs with names.
2514 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2515 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2520 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2521 improves application performance by removing data copies and providing
2522 applications with zero-copy system calls such as sendfile and splice.
2526 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2530 * `PKCS12_parse` now maintains the order of the parsed certificates
2531 when outputting them via `*ca` (rather than reversing it).
2535 * Deprecated pthread fork support methods.
2539 * Added support for FFDHE key exchange in TLS 1.3.
2543 * Added a new concept for OpenSSL plugability: providers. This
2544 functionality is designed to replace the ENGINE API and ENGINE
2545 implementations, and to be much more dynamic, allowing provider
2546 authors to introduce new algorithms among other things, as long as
2547 there's an API that supports the algorithm type.
2549 With this concept comes a new core API for interaction between
2550 libcrypto and provider implementations. Public libcrypto functions
2551 that want to use providers do so through this core API.
2553 The main documentation for this core API is found in
2554 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2555 refer to other manuals describing the API specific for supported
2556 algorithm types (also called operations).
2563 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2565 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2567 * Avoid loading of a dynamic engine twice.
2571 * Prioritise DANE TLSA issuer certs over peer certs
2575 * Fixed random API for MacOS prior to 10.12
2577 These MacOS versions don't support the CommonCrypto APIs
2581 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2583 * Fixed an SM2 Decryption Buffer Overflow.
2585 In order to decrypt SM2 encrypted data an application is expected to
2586 call the API function EVP_PKEY_decrypt(). Typically an application will
2587 call this function twice. The first time, on entry, the "out" parameter
2588 can be NULL and, on exit, the "outlen" parameter is populated with the
2589 buffer size required to hold the decrypted plaintext. The application
2590 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2591 again, but this time passing a non-NULL value for the "out" parameter.
2593 A bug in the implementation of the SM2 decryption code means that the
2594 calculation of the buffer size required to hold the plaintext returned
2595 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2596 size required by the second call. This can lead to a buffer overflow
2597 when EVP_PKEY_decrypt() is called by the application a second time with
2598 a buffer that is too small.
2600 A malicious attacker who is able present SM2 content for decryption to
2601 an application could cause attacker chosen data to overflow the buffer
2602 by up to a maximum of 62 bytes altering the contents of other data held
2603 after the buffer, possibly changing application behaviour or causing
2604 the application to crash. The location of the buffer is application
2605 dependent but is typically heap allocated.
2610 * Fixed various read buffer overruns processing ASN.1 strings
2612 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2613 structure which contains a buffer holding the string data and a field
2614 holding the buffer length. This contrasts with normal C strings which
2615 are represented as a buffer for the string data which is terminated
2616 with a NUL (0) byte.
2618 Although not a strict requirement, ASN.1 strings that are parsed using
2619 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2620 well as any string whose value has been set with the ASN1_STRING_set()
2621 function will additionally NUL terminate the byte array in the
2622 ASN1_STRING structure.
2624 However, it is possible for applications to directly construct valid
2625 ASN1_STRING structures which do not NUL terminate the byte array by
2626 directly setting the "data" and "length" fields in the ASN1_STRING
2627 array. This can also happen by using the ASN1_STRING_set0() function.
2629 Numerous OpenSSL functions that print ASN.1 data have been found to
2630 assume that the ASN1_STRING byte array will be NUL terminated, even
2631 though this is not guaranteed for strings that have been directly
2632 constructed. Where an application requests an ASN.1 structure to be
2633 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2634 been directly constructed by the application without NUL terminating
2635 the "data" field, then a read buffer overrun can occur.
2637 The same thing can also occur during name constraints processing
2638 of certificates (for example if a certificate has been directly
2639 constructed by the application instead of loading it via the OpenSSL
2640 parsing functions, and the certificate contains non NUL terminated
2641 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2642 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2644 If a malicious actor can cause an application to directly construct an
2645 ASN1_STRING and then process it through one of the affected OpenSSL
2646 functions then this issue could be hit. This might result in a crash
2647 (causing a Denial of Service attack). It could also result in the
2648 disclosure of private memory contents (such as private keys, or
2649 sensitive plaintext).
2654 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2656 * Fixed a problem with verifying a certificate chain when using the
2657 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2658 the certificates present in a certificate chain. It is not set by default.
2660 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2661 the chain that have explicitly encoded elliptic curve parameters was added
2662 as an additional strict check.
2664 An error in the implementation of this check meant that the result of a
2665 previous check to confirm that certificates in the chain are valid CA
2666 certificates was overwritten. This effectively bypasses the check
2667 that non-CA certificates must not be able to issue other certificates.
2669 If a "purpose" has been configured then there is a subsequent opportunity
2670 for checks that the certificate is a valid CA. All of the named "purpose"
2671 values implemented in libcrypto perform this check. Therefore, where
2672 a purpose is set the certificate chain will still be rejected even when the
2673 strict flag has been used. A purpose is set by default in libssl client and
2674 server certificate verification routines, but it can be overridden or
2675 removed by an application.
2677 In order to be affected, an application must explicitly set the
2678 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2679 for the certificate verification or, in the case of TLS client or server
2680 applications, override the default purpose.
2685 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2686 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2687 renegotiation ClientHello omits the signature_algorithms extension (where it
2688 was present in the initial ClientHello), but includes a
2689 signature_algorithms_cert extension then a NULL pointer dereference will
2690 result, leading to a crash and a denial of service attack.
2692 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2693 (which is the default configuration). OpenSSL TLS clients are not impacted by
2697 *Peter Kästle and Samuel Sapalski*
2699 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2701 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2702 create a unique hash value based on the issuer and serial number data
2703 contained within an X509 certificate. However, it was failing to correctly
2704 handle any errors that may occur while parsing the issuer field (which might
2705 occur if the issuer field is maliciously constructed). This may subsequently
2706 result in a NULL pointer deref and a crash leading to a potential denial of
2712 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2713 padding mode to correctly check for rollback attacks. This is considered a
2714 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2719 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2720 functions. Previously they could overflow the output length argument in some
2721 cases where the input length is close to the maximum permissible length for
2722 an integer on the platform. In such cases the return value from the function
2723 call would be 1 (indicating success), but the output length value would be
2724 negative. This could cause applications to behave incorrectly or crash.
2729 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2730 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2731 could be exploited in a side channel attack to recover the password. Since
2732 the attack is local host only this is outside of the current OpenSSL
2733 threat model and therefore no CVE is assigned.
2735 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2740 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2742 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2743 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2744 If an attacker can control both items being compared then this could lead
2745 to a possible denial of service attack. OpenSSL itself uses the
2746 GENERAL_NAME_cmp function for two purposes:
2747 1) Comparing CRL distribution point names between an available CRL and a
2748 CRL distribution point embedded in an X509 certificate
2749 2) When verifying that a timestamp response token signer matches the
2750 timestamp authority name (exposed via the API functions
2751 TS_RESP_verify_response and TS_RESP_verify_token)
2756 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2758 * Certificates with explicit curve parameters are now disallowed in
2759 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2763 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2764 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2765 conversely, silently ignore DTLS protocol version bounds when configuring
2766 TLS-based contexts. The commands can be repeated to set bounds of both
2767 types. The same applies with the corresponding "min_protocol" and
2768 "max_protocol" command-line switches, in case some application uses both TLS
2771 SSL_CTX instances that are created for a fixed protocol version (e.g.
2772 TLSv1_server_method()) also silently ignore version bounds. Previously
2773 attempts to apply bounds to these protocol versions would result in an
2774 error. Now only the "version-flexible" SSL_CTX instances are subject to
2775 limits in configuration files in command-line options.
2779 * Handshake now fails if Extended Master Secret extension is dropped
2784 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2786 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2788 * Fixed segmentation fault in SSL_check_chain()
2789 Server or client applications that call the SSL_check_chain() function
2790 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2791 dereference as a result of incorrect handling of the
2792 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2793 or unrecognised signature algorithm is received from the peer. This could
2794 be exploited by a malicious peer in a Denial of Service attack.
2799 * Added AES consttime code for no-asm configurations
2800 an optional constant time support for AES was added
2801 when building openssl for no-asm.
2802 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2803 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2804 At this time this feature is by default disabled.
2805 It will be enabled by default in 3.0.
2809 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2811 * Revert the change of EOF detection while reading in libssl to avoid
2812 regressions in applications depending on the current way of reporting
2813 the EOF. As the existing method is not fully accurate the change to
2814 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2815 branch and will be present in the 3.0 release.
2819 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2820 when primes for RSA keys are computed.
2821 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2822 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2823 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2824 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2825 This avoids possible fingerprinting of newly generated RSA modules.
2829 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2831 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2832 while reading in libssl then we would report an error back to the
2833 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2834 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2835 therefore give a hint as to what went wrong.
2839 * Check that ed25519 and ed448 are allowed by the security level. Previously
2840 signature algorithms not using an MD were not being checked that they were
2841 allowed by the security level.
2845 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2846 was not quite right. The behaviour was not consistent between resumption
2847 and normal handshakes, and also not quite consistent with historical
2848 behaviour. The behaviour in various scenarios has been clarified and
2849 it has been updated to make it match historical behaviour as closely as
2854 * *[VMS only]* The header files that the VMS compilers include automatically,
2855 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2856 that the C++ compiler doesn't understand. This is a shortcoming in the
2857 compiler, but can be worked around with `__cplusplus` guards.
2859 C++ applications that use OpenSSL libraries must be compiled using the
2860 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2861 functions. Otherwise, only functions with symbols of less than 31
2862 characters can be used, as the linker will not be able to successfully
2863 resolve symbols with longer names.
2867 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2868 The presence of this system service is determined at run-time.
2872 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2877 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2879 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2880 number generator (RNG). This was intended to include protection in the
2881 event of a fork() system call in order to ensure that the parent and child
2882 processes did not share the same RNG state. However, this protection was not
2883 being used in the default case.
2885 A partial mitigation for this issue is that the output from a high
2886 precision timer is mixed into the RNG state so the likelihood of a parent
2887 and child process sharing state is significantly reduced.
2889 If an application already calls OPENSSL_init_crypto() explicitly using
2890 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2893 *Matthias St. Pierre*
2895 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2896 used even when parsing explicit parameters, when loading a encoded key
2897 or calling `EC_GROUP_new_from_ecpkparameters()`/
2898 `EC_GROUP_new_from_ecparameters()`.
2899 This prevents bypass of security hardening and performance gains,
2900 especially for curves with specialized EC_METHODs.
2901 By default, if a key encoded with explicit parameters is loaded and later
2902 encoded, the output is still encoded with explicit parameters, even if
2903 internally a "named" EC_GROUP is used for computation.
2907 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2908 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2909 NULL. After this change, only the cofactor parameter can be NULL. It also
2910 does some minimal sanity checks on the passed order.
2915 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2916 An attack is simple, if the first CMS_recipientInfo is valid but the
2917 second CMS_recipientInfo is chosen ciphertext. If the second
2918 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2919 encryption key will be replaced by garbage, and the message cannot be
2920 decoded, but if the RSA decryption fails, the correct encryption key is
2921 used and the recipient will not notice the attack.
2922 As a work around for this potential attack the length of the decrypted
2923 key must be equal to the cipher default key length, in case the
2924 certificate is not given and all recipientInfo are tried out.
2925 The old behaviour can be re-enabled in the CMS code by setting the
2926 CMS_DEBUG_DECRYPT flag.
2931 * Early start up entropy quality from the DEVRANDOM seed source has been
2932 improved for older Linux systems. The RAND subsystem will wait for
2933 /dev/random to be producing output before seeding from /dev/urandom.
2934 The seeded state is stored for future library initialisations using
2935 a system global shared memory segment. The shared memory identifier
2936 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2937 the desired value. The default identifier is 114.
2941 * Correct the extended master secret constant on EBCDIC systems. Without this
2942 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2943 negotiate EMS will fail. Unfortunately this also means that TLS connections
2944 between EBCDIC systems with this fix, and EBCDIC systems without this
2945 fix will fail if they negotiate EMS.
2949 * Use Windows installation paths in the mingw builds
2951 Mingw isn't a POSIX environment per se, which means that Windows
2952 paths should be used for installation.
2957 * Changed DH_check to accept parameters with order q and 2q subgroups.
2958 With order 2q subgroups the bit 0 of the private key is not secret
2959 but DH_generate_key works around that by clearing bit 0 of the
2960 private key for those. This avoids leaking bit 0 of the private key.
2964 * Significantly reduce secure memory usage by the randomness pools.
2968 * Revert the DEVRANDOM_WAIT feature for Linux systems
2970 The DEVRANDOM_WAIT feature added a select() call to wait for the
2971 /dev/random device to become readable before reading from the
2972 /dev/urandom device.
2974 It turned out that this change had negative side effects on
2975 performance which were not acceptable. After some discussion it
2976 was decided to revert this feature and leave it up to the OS
2977 resp. the platform maintainer to ensure a proper initialization
2978 during early boot time.
2980 *Matthias St. Pierre*
2982 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2984 * Add build tests for C++. These are generated files that only do one
2985 thing, to include one public OpenSSL head file each. This tests that
2986 the public header files can be usefully included in a C++ application.
2988 This test isn't enabled by default. It can be enabled with the option
2989 'enable-buildtest-c++'.
2993 * Enable SHA3 pre-hashing for ECDSA and DSA.
2997 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2998 This changes the size when using the `genpkey` command when no size is given.
2999 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3000 generation commands to use 2048 bits by default.
3004 * Reorganize the manual pages to consistently have RETURN VALUES,
3005 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
3006 util/fix-doc-nits accordingly.
3008 *Paul Yang, Joshua Lock*
3010 * Add the missing accessor EVP_PKEY_get0_engine()
3014 * Have commands like `s_client` and `s_server` output the signature scheme
3015 along with other cipher suite parameters when debugging.
3019 * Make OPENSSL_config() error agnostic again.
3023 * Do the error handling in RSA decryption constant time.
3027 * Prevent over long nonces in ChaCha20-Poly1305.
3029 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3030 for every encryption operation. RFC 7539 specifies that the nonce value
3031 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3032 and front pads the nonce with 0 bytes if it is less than 12
3033 bytes. However it also incorrectly allows a nonce to be set of up to 16
3034 bytes. In this case only the last 12 bytes are significant and any
3035 additional leading bytes are ignored.
3037 It is a requirement of using this cipher that nonce values are
3038 unique. Messages encrypted using a reused nonce value are susceptible to
3039 serious confidentiality and integrity attacks. If an application changes
3040 the default nonce length to be longer than 12 bytes and then makes a
3041 change to the leading bytes of the nonce expecting the new value to be a
3042 new unique nonce then such an application could inadvertently encrypt
3043 messages with a reused nonce.
3045 Additionally the ignored bytes in a long nonce are not covered by the
3046 integrity guarantee of this cipher. Any application that relies on the
3047 integrity of these ignored leading bytes of a long nonce may be further
3048 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3049 is safe because no such use sets such a long nonce value. However user
3050 applications that use this cipher directly and set a non-default nonce
3051 length to be longer than 12 bytes may be vulnerable.
3053 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3059 * Add DEVRANDOM_WAIT feature for Linux systems
3061 On older Linux systems where the getrandom() system call is not available,
3062 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
3063 Contrary to getrandom(), the /dev/urandom device will not block during
3064 early boot when the kernel CSPRNG has not been seeded yet.
3066 To mitigate this known weakness, use select() to wait for /dev/random to
3067 become readable before reading from /dev/urandom.
3069 * Ensure that SM2 only uses SM3 as digest algorithm
3073 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
3075 * Change the info callback signals for the start and end of a post-handshake
3076 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
3077 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
3078 confused by this and assume that a TLSv1.2 renegotiation has started. This
3079 can break KeyUpdate handling. Instead we no longer signal the start and end
3080 of a post handshake message exchange (although the messages themselves are
3081 still signalled). This could break some applications that were expecting
3082 the old signals. However without this KeyUpdate is not usable for many
3087 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
3089 * Timing vulnerability in DSA signature generation
3091 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3092 timing side channel attack. An attacker could use variations in the signing
3093 algorithm to recover the private key.
3095 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3100 * Timing vulnerability in ECDSA signature generation
3102 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3103 timing side channel attack. An attacker could use variations in the signing
3104 algorithm to recover the private key.
3106 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3111 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
3112 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
3113 of two gigabytes and the error handling improved.
3115 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
3116 categorized as a normal bug, not a security issue, because the DRBG reseeds
3117 automatically and is fully functional even without additional randomness
3118 provided by the application.
3120 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
3122 * Add a new ClientHello callback. Provides a callback interface that gives
3123 the application the ability to adjust the nascent SSL object at the
3124 earliest stage of ClientHello processing, immediately after extensions have
3125 been collected but before they have been processed. In particular, this
3126 callback can adjust the supported TLS versions in response to the contents
3131 * Add SM2 base algorithm support.
3135 * s390x assembly pack: add (improved) hardware-support for the following
3136 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
3137 aes-cfb/cfb8, aes-ecb.
3141 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3142 parameter is no longer accepted, as it leads to a corrupt table. NULL
3143 pem_str is reserved for alias entries only.
3147 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3148 step for prime curves. The new implementation is based on formulae from
3149 differential addition-and-doubling in homogeneous projective coordinates
3150 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
3151 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
3152 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
3153 to work in projective coordinates.
3155 *Billy Bob Brumley, Nicola Tuveri*
3157 * Change generating and checking of primes so that the error rate of not
3158 being prime depends on the intended use based on the size of the input.
3159 For larger primes this will result in more rounds of Miller-Rabin.
3160 The maximal error rate for primes with more than 1080 bits is lowered
3163 *Kurt Roeckx, Annie Yousar*
3165 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3169 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
3170 moving between systems, and to avoid confusion when a Windows build is
3171 done with mingw vs with MSVC. For POSIX installs, there's still a
3172 symlink or copy named 'tsget' to avoid that confusion as well.
3176 * Revert blinding in ECDSA sign and instead make problematic addition
3177 length-invariant. Switch even to fixed-length Montgomery multiplication.
3181 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3182 step for binary curves. The new implementation is based on formulae from
3183 differential addition-and-doubling in mixed Lopez-Dahab projective
3184 coordinates, modified to independently blind the operands.
3186 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3188 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3189 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3190 EC_METHODs to implement their own specialized "ladder step", to take
3191 advantage of more favorable coordinate systems or more efficient
3192 differential addition-and-doubling algorithms.
3194 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3196 * Modified the random device based seed sources to keep the relevant
3197 file descriptors open rather than reopening them on each access.
3198 This allows such sources to operate in a chroot() jail without
3199 the associated device nodes being available. This behaviour can be
3200 controlled using RAND_keep_random_devices_open().
3204 * Numerous side-channel attack mitigations have been applied. This may have
3205 performance impacts for some algorithms for the benefit of improved
3206 security. Specific changes are noted in this change log by their respective
3211 * AIX shared library support overhaul. Switch to AIX "natural" way of
3212 handling shared libraries, which means collecting shared objects of
3213 different versions and bitnesses in one common archive. This allows to
3214 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3215 doesn't affect the way 3rd party applications are linked, only how
3216 multi-version installation is managed.
3220 * Make ec_group_do_inverse_ord() more robust and available to other
3221 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3222 mitigations are applied to the fallback BN_mod_inverse().
3223 When using this function rather than BN_mod_inverse() directly, new
3224 EC cryptosystem implementations are then safer-by-default.
3228 * Add coordinate blinding for EC_POINT and implement projective
3229 coordinate blinding for generic prime curves as a countermeasure to
3230 chosen point SCA attacks.
3232 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3234 * Add blinding to ECDSA and DSA signatures to protect against side channel
3235 attacks discovered by Keegan Ryan (NCC Group).
3239 * Enforce checking in the `pkeyutl` command to ensure that the input
3240 length does not exceed the maximum supported digest length when performing
3241 a sign, verify or verifyrecover operation.
3245 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3246 I/O in combination with something like select() or poll() will hang. This
3247 can be turned off again using SSL_CTX_clear_mode().
3248 Many applications do not properly handle non-application data records, and
3249 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3250 around the problems in those applications, but can also break some.
3251 It's recommended to read the manpages about SSL_read(), SSL_write(),
3252 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3253 SSL_CTX_set_read_ahead() again.
3257 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3258 now allow empty (zero character) pass phrases.
3262 * Apply blinding to binary field modular inversion and remove patent
3263 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3267 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3268 binary and prime elliptic curves.
3272 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3273 constant time fixed point multiplication.
3277 * Revise elliptic curve scalar multiplication with timing attack
3278 defenses: ec_wNAF_mul redirects to a constant time implementation
3279 when computing fixed point and variable point multiplication (which
3280 in OpenSSL are mostly used with secret scalars in keygen, sign,
3281 ECDH derive operations).
3282 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3285 * Updated CONTRIBUTING
3289 * Updated DRBG / RAND to request nonce and additional low entropy
3290 randomness from the system.
3292 *Matthias St. Pierre*
3294 * Updated 'openssl rehash' to use OpenSSL consistent default.
3298 * Moved the load of the ssl_conf module to libcrypto, which helps
3299 loading engines that libssl uses before libssl is initialised.
3303 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3307 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3309 *Ingo Schwarze, Rich Salz*
3311 * Added output of accepting IP address and port for 'openssl s_server'
3315 * Added a new API for TLSv1.3 ciphersuites:
3316 SSL_CTX_set_ciphersuites()
3317 SSL_set_ciphersuites()
3321 * Memory allocation failures consistently add an error to the error
3326 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3327 in libcrypto when run as setuid/setgid.
3331 * Load any config file by default when libssl is used.
3335 * Added new public header file <openssl/rand_drbg.h> and documentation
3336 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3338 *Matthias St. Pierre*
3340 * QNX support removed (cannot find contributors to get their approval
3341 for the license change).
3345 * TLSv1.3 replay protection for early data has been implemented. See the
3346 SSL_read_early_data() man page for further details.
3350 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3351 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3352 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3353 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3354 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3355 configuration has been separated out. See the ciphers man page or the
3356 SSL_CTX_set_ciphersuites() man page for more information.
3360 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3361 in responder mode now supports the new "-multi" option, which
3362 spawns the specified number of child processes to handle OCSP
3363 requests. The "-timeout" option now also limits the OCSP
3364 responder's patience to wait to receive the full client request
3365 on a newly accepted connection. Child processes are respawned
3366 as needed, and the CA index file is automatically reloaded
3367 when changed. This makes it possible to run the "ocsp" responder
3368 as a long-running service, making the OpenSSL CA somewhat more
3369 feature-complete. In this mode, most diagnostic messages logged
3370 after entering the event loop are logged via syslog(3) rather than
3375 * Added support for X448 and Ed448. Heavily based on original work by
3380 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3381 objects loaded. This adds the functions OSSL_STORE_expect() and
3382 OSSL_STORE_find() as well as needed tools to construct searches and
3383 get the search data out of them.
3387 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3388 version of OpenSSL should review their configuration settings to ensure
3389 that they are still appropriate for TLSv1.3. For further information see:
3390 <https://wiki.openssl.org/index.php/TLS1.3>
3394 * Grand redesign of the OpenSSL random generator
3396 The default RAND method now utilizes an AES-CTR DRBG according to
3397 NIST standard SP 800-90Ar1. The new random generator is essentially
3398 a port of the default random generator from the OpenSSL FIPS 2.0
3399 object module. It is a hybrid deterministic random bit generator
3400 using an AES-CTR bit stream and which seeds and reseeds itself
3401 automatically using trusted system entropy sources.
3403 Some of its new features are:
3404 - Support for multiple DRBG instances with seed chaining.
3405 - The default RAND method makes use of a DRBG.
3406 - There is a public and private DRBG instance.
3407 - The DRBG instances are fork-safe.
3408 - Keep all global DRBG instances on the secure heap if it is enabled.
3409 - The public and private DRBG instance are per thread for lock free
3412 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3414 * Changed Configure so it only says what it does and doesn't dump
3415 so much data. Instead, ./configdata.pm should be used as a script
3416 to display all sorts of configuration data.
3420 * Added processing of "make variables" to Configure.
3424 * Added SHA512/224 and SHA512/256 algorithm support.
3428 * The last traces of Netware support, first removed in 1.1.0, have
3433 * Get rid of Makefile.shared, and in the process, make the processing
3434 of certain files (rc.obj, or the .def/.map/.opt files produced from
3435 the ordinal files) more visible and hopefully easier to trace and
3436 debug (or make silent).
3440 * Make it possible to have environment variable assignments as
3441 arguments to config / Configure.
3445 * Add multi-prime RSA (RFC 8017) support.
3449 * Add SM3 implemented according to GB/T 32905-2016
3450 *Jack Lloyd <jack.lloyd@ribose.com>,*
3451 *Ronald Tse <ronald.tse@ribose.com>,*
3452 *Erick Borsboom <erick.borsboom@ribose.com>*
3454 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3455 as documented in RFC6066.
3456 Based on a patch from Tomasz Moń
3458 *Filipe Raimundo da Silva*
3460 * Add SM4 implemented according to GB/T 32907-2016.
3461 *Jack Lloyd <jack.lloyd@ribose.com>,*
3462 *Ronald Tse <ronald.tse@ribose.com>,*
3463 *Erick Borsboom <erick.borsboom@ribose.com>*
3465 * Reimplement -newreq-nodes and ERR_error_string_n; the
3466 original author does not agree with the license change.
3470 * Add ARIA AEAD TLS support.
3474 * Some macro definitions to support VS6 have been removed. Visual
3475 Studio 6 has not worked since 1.1.0
3479 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3480 without clearing the errors.
3484 * Add "atfork" functions. If building on a system that without
3485 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3486 requirements. The RAND facility now uses/requires this.
3494 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3495 not possible to disable entirely. However, it's still possible to
3496 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3499 To disable, configure with 'no-ui-console'. 'no-ui' is still
3500 possible to use as an alias. Check at compile time with the
3501 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3502 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3506 * Add a STORE module, which implements a uniform and URI based reader of
3507 stores that can contain keys, certificates, CRLs and numerous other
3508 objects. The main API is loosely based on a few stdio functions,
3509 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3510 OSSL_STORE_error and OSSL_STORE_close.
3511 The implementation uses backends called "loaders" to implement arbitrary
3512 URI schemes. There is one built in "loader" for the 'file' scheme.
3516 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3517 then adjusted to work on FreeBSD 8.4 as well.
3518 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3519 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3523 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3524 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3525 error code calls like this:
3527 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3529 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3530 that can be encoded in C. For the foreseeable future, this will only
3533 *Richard Levitte and Tim Hudson*
3535 * Removed BSD cryptodev engine.
3539 * Add a build target 'build_all_generated', to build all generated files
3540 and only that. This can be used to prepare everything that requires
3541 things like perl for a system that lacks perl and then move everything
3542 to that system and do the rest of the build there.
3546 * In the UI interface, make it possible to duplicate the user data. This
3547 can be used by engines that need to retain the data for a longer time
3548 than just the call where this user data is passed.
3552 * Ignore the '-named_curve auto' value for compatibility of applications
3555 *Tomáš Mráz <tmraz@fedoraproject.org>*
3557 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3558 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3559 alerts across multiple records (some of which could be empty). In practice
3560 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3561 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3562 support this at all. Supporting it adds significant complexity to the
3563 record layer, and its removal is unlikely to cause interoperability
3568 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3569 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3570 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3575 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3576 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3578 *Richard Levitte, Andy Polyakov*
3580 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3585 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3586 platform rather than 'mingw'.
3590 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3591 success if they are asked to add an object which already exists
3592 in the store. This change cascades to other functions which load
3593 certificates and CRLs.
3597 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3598 facilitate stack unwinding even from assembly subroutines.
3602 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3603 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3607 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3608 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3609 which is the minimum version we support.
3613 * Certificate time validation (X509_cmp_time) enforces stricter
3614 compliance with RFC 5280. Fractional seconds and timezone offsets
3615 are no longer allowed.
3619 * Add support for ARIA
3623 * s_client will now send the Server Name Indication (SNI) extension by
3624 default unless the new "-noservername" option is used. The server name is
3625 based on the host provided to the "-connect" option unless overridden by
3626 using "-servername".
3630 * Add support for SipHash
3634 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3635 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3636 prevent issues where no progress is being made and the peer continually
3637 sends unrecognised record types, using up resources processing them.
3641 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3642 using the algorithm defined in
3643 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3647 * Heartbeat support has been removed; the ABI is changed for now.
3649 *Richard Levitte, Rich Salz*
3651 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3655 * The RSA "null" method, which was partially supported to avoid patent
3656 issues, has been replaced to always returns NULL.
3663 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3665 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3666 used even when parsing explicit parameters, when loading a encoded key
3667 or calling `EC_GROUP_new_from_ecpkparameters()`/
3668 `EC_GROUP_new_from_ecparameters()`.
3669 This prevents bypass of security hardening and performance gains,
3670 especially for curves with specialized EC_METHODs.
3671 By default, if a key encoded with explicit parameters is loaded and later
3672 encoded, the output is still encoded with explicit parameters, even if
3673 internally a "named" EC_GROUP is used for computation.
3677 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3678 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3679 NULL. After this change, only the cofactor parameter can be NULL. It also
3680 does some minimal sanity checks on the passed order.
3685 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3686 An attack is simple, if the first CMS_recipientInfo is valid but the
3687 second CMS_recipientInfo is chosen ciphertext. If the second
3688 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3689 encryption key will be replaced by garbage, and the message cannot be
3690 decoded, but if the RSA decryption fails, the correct encryption key is
3691 used and the recipient will not notice the attack.
3692 As a work around for this potential attack the length of the decrypted
3693 key must be equal to the cipher default key length, in case the
3694 certificate is not given and all recipientInfo are tried out.
3695 The old behaviour can be re-enabled in the CMS code by setting the
3696 CMS_DEBUG_DECRYPT flag.
3701 * Use Windows installation paths in the mingw builds
3703 Mingw isn't a POSIX environment per se, which means that Windows
3704 paths should be used for installation.
3709 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3711 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3712 This changes the size when using the `genpkey` command when no size is given.
3713 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3714 generation commands to use 2048 bits by default.
3718 * Prevent over long nonces in ChaCha20-Poly1305.
3720 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3721 for every encryption operation. RFC 7539 specifies that the nonce value
3722 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3723 and front pads the nonce with 0 bytes if it is less than 12
3724 bytes. However it also incorrectly allows a nonce to be set of up to 16
3725 bytes. In this case only the last 12 bytes are significant and any
3726 additional leading bytes are ignored.
3728 It is a requirement of using this cipher that nonce values are
3729 unique. Messages encrypted using a reused nonce value are susceptible to
3730 serious confidentiality and integrity attacks. If an application changes
3731 the default nonce length to be longer than 12 bytes and then makes a
3732 change to the leading bytes of the nonce expecting the new value to be a
3733 new unique nonce then such an application could inadvertently encrypt
3734 messages with a reused nonce.
3736 Additionally the ignored bytes in a long nonce are not covered by the
3737 integrity guarantee of this cipher. Any application that relies on the
3738 integrity of these ignored leading bytes of a long nonce may be further
3739 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3740 is safe because no such use sets such a long nonce value. However user
3741 applications that use this cipher directly and set a non-default nonce
3742 length to be longer than 12 bytes may be vulnerable.
3744 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3750 * Added SCA hardening for modular field inversion in EC_GROUP through
3751 a new dedicated field_inv() pointer in EC_METHOD.
3752 This also addresses a leakage affecting conversions from projective
3753 to affine coordinates.
3755 *Billy Bob Brumley, Nicola Tuveri*
3757 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3758 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3762 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3766 * Remove the 'dist' target and add a tarball building script. The
3767 'dist' target has fallen out of use, and it shouldn't be
3768 necessary to configure just to create a source distribution.
3772 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3774 * Timing vulnerability in DSA signature generation
3776 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3777 timing side channel attack. An attacker could use variations in the signing
3778 algorithm to recover the private key.
3780 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3785 * Timing vulnerability in ECDSA signature generation
3787 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3788 timing side channel attack. An attacker could use variations in the signing
3789 algorithm to recover the private key.
3791 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3796 * Add coordinate blinding for EC_POINT and implement projective
3797 coordinate blinding for generic prime curves as a countermeasure to
3798 chosen point SCA attacks.
3800 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3802 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3804 * Client DoS due to large DH parameter
3806 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3807 malicious server can send a very large prime value to the client. This will
3808 cause the client to spend an unreasonably long period of time generating a
3809 key for this prime resulting in a hang until the client has finished. This
3810 could be exploited in a Denial Of Service attack.
3812 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3817 * Cache timing vulnerability in RSA Key Generation
3819 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3820 a cache timing side channel attack. An attacker with sufficient access to
3821 mount cache timing attacks during the RSA key generation process could
3822 recover the private key.
3824 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3825 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3830 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3831 parameter is no longer accepted, as it leads to a corrupt table. NULL
3832 pem_str is reserved for alias entries only.
3836 * Revert blinding in ECDSA sign and instead make problematic addition
3837 length-invariant. Switch even to fixed-length Montgomery multiplication.
3841 * Change generating and checking of primes so that the error rate of not
3842 being prime depends on the intended use based on the size of the input.
3843 For larger primes this will result in more rounds of Miller-Rabin.
3844 The maximal error rate for primes with more than 1080 bits is lowered
3847 *Kurt Roeckx, Annie Yousar*
3849 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3853 * Add blinding to ECDSA and DSA signatures to protect against side channel
3854 attacks discovered by Keegan Ryan (NCC Group).
3858 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3859 now allow empty (zero character) pass phrases.
3863 * Certificate time validation (X509_cmp_time) enforces stricter
3864 compliance with RFC 5280. Fractional seconds and timezone offsets
3865 are no longer allowed.
3869 * Fixed a text canonicalisation bug in CMS
3871 Where a CMS detached signature is used with text content the text goes
3872 through a canonicalisation process first prior to signing or verifying a
3873 signature. This process strips trailing space at the end of lines, converts
3874 line terminators to CRLF and removes additional trailing line terminators
3875 at the end of a file. A bug in the canonicalisation process meant that
3876 some characters, such as form-feed, were incorrectly treated as whitespace
3877 and removed. This is contrary to the specification (RFC5485). This fix
3878 could mean that detached text data signed with an earlier version of
3879 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3880 signed with a fixed OpenSSL may fail to verify with an earlier version of
3881 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3882 and use the "-binary" flag (for the "cms" command line application) or set
3883 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3887 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3889 * Constructed ASN.1 types with a recursive definition could exceed the stack
3891 Constructed ASN.1 types with a recursive definition (such as can be found
3892 in PKCS7) could eventually exceed the stack given malicious input with
3893 excessive recursion. This could result in a Denial Of Service attack. There
3894 are no such structures used within SSL/TLS that come from untrusted sources
3895 so this is considered safe.
3897 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3903 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3905 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3906 effectively reduced to only comparing the least significant bit of each
3907 byte. This allows an attacker to forge messages that would be considered as
3908 authenticated in an amount of tries lower than that guaranteed by the
3909 security claims of the scheme. The module can only be compiled by the
3910 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3912 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3918 * Add a build target 'build_all_generated', to build all generated files
3919 and only that. This can be used to prepare everything that requires
3920 things like perl for a system that lacks perl and then move everything
3921 to that system and do the rest of the build there.
3925 * Backport SSL_OP_NO_RENGOTIATION
3927 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3928 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3929 changes this is no longer possible in 1.1.0. Therefore, the new
3930 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3931 1.1.0 to provide equivalent functionality.
3933 Note that if an application built against 1.1.0h headers (or above) is run
3934 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3935 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3939 * Removed the OS390-Unix config target. It relied on a script that doesn't
3944 * rsaz_1024_mul_avx2 overflow bug on x86_64
3946 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3947 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3948 Analysis suggests that attacks against RSA and DSA as a result of this
3949 defect would be very difficult to perform and are not believed likely.
3950 Attacks against DH1024 are considered just feasible, because most of the
3951 work necessary to deduce information about a private key may be performed
3952 offline. The amount of resources required for such an attack would be
3953 significant. However, for an attack on TLS to be meaningful, the server
3954 would have to share the DH1024 private key among multiple clients, which is
3955 no longer an option since CVE-2016-0701.
3957 This only affects processors that support the AVX2 but not ADX extensions
3958 like Intel Haswell (4th generation).
3960 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3961 was originally found via the OSS-Fuzz project.
3966 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3968 * bn_sqrx8x_internal carry bug on x86_64
3970 There is a carry propagating bug in the x86_64 Montgomery squaring
3971 procedure. No EC algorithms are affected. Analysis suggests that attacks
3972 against RSA and DSA as a result of this defect would be very difficult to
3973 perform and are not believed likely. Attacks against DH are considered just
3974 feasible (although very difficult) because most of the work necessary to
3975 deduce information about a private key may be performed offline. The amount
3976 of resources required for such an attack would be very significant and
3977 likely only accessible to a limited number of attackers. An attacker would
3978 additionally need online access to an unpatched system using the target
3979 private key in a scenario with persistent DH parameters and a private
3980 key that is shared between multiple clients.
3982 This only affects processors that support the BMI1, BMI2 and ADX extensions
3983 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3985 This issue was reported to OpenSSL by the OSS-Fuzz project.
3990 * Malformed X.509 IPAddressFamily could cause OOB read
3992 If an X.509 certificate has a malformed IPAddressFamily extension,
3993 OpenSSL could do a one-byte buffer overread. The most likely result
3994 would be an erroneous display of the certificate in text format.
3996 This issue was reported to OpenSSL by the OSS-Fuzz project.
4001 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
4003 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4004 platform rather than 'mingw'.
4008 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
4009 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
4010 which is the minimum version we support.
4014 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
4016 * Encrypt-Then-Mac renegotiation crash
4018 During a renegotiation handshake if the Encrypt-Then-Mac extension is
4019 negotiated where it was not in the original handshake (or vice-versa) then
4020 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
4021 and servers are affected.
4023 This issue was reported to OpenSSL by Joe Orton (Red Hat).
4028 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
4030 * Truncated packet could crash via OOB read
4032 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4033 cipher is being used, then a truncated packet can cause that host to
4034 perform an out-of-bounds read, usually resulting in a crash.
4036 This issue was reported to OpenSSL by Robert Święcki of Google.
4041 * Bad (EC)DHE parameters cause a client crash
4043 If a malicious server supplies bad parameters for a DHE or ECDHE key
4044 exchange then this can result in the client attempting to dereference a
4045 NULL pointer leading to a client crash. This could be exploited in a Denial
4048 This issue was reported to OpenSSL by Guido Vranken.
4053 * BN_mod_exp may produce incorrect results on x86_64
4055 There is a carry propagating bug in the x86_64 Montgomery squaring
4056 procedure. No EC algorithms are affected. Analysis suggests that attacks
4057 against RSA and DSA as a result of this defect would be very difficult to
4058 perform and are not believed likely. Attacks against DH are considered just
4059 feasible (although very difficult) because most of the work necessary to
4060 deduce information about a private key may be performed offline. The amount
4061 of resources required for such an attack would be very significant and
4062 likely only accessible to a limited number of attackers. An attacker would
4063 additionally need online access to an unpatched system using the target
4064 private key in a scenario with persistent DH parameters and a private
4065 key that is shared between multiple clients. For example this can occur by
4066 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4067 similar to CVE-2015-3193 but must be treated as a separate problem.
4069 This issue was reported to OpenSSL by the OSS-Fuzz project.
4074 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
4076 * ChaCha20/Poly1305 heap-buffer-overflow
4078 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
4079 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
4080 crash. This issue is not considered to be exploitable beyond a DoS.
4082 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
4087 * CMS Null dereference
4089 Applications parsing invalid CMS structures can crash with a NULL pointer
4090 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
4091 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
4092 structure callback if an attempt is made to free certain invalid encodings.
4093 Only CHOICE structures using a callback which do not handle NULL value are
4096 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
4101 * Montgomery multiplication may produce incorrect results
4103 There is a carry propagating bug in the Broadwell-specific Montgomery
4104 multiplication procedure that handles input lengths divisible by, but
4105 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4106 and DH private keys are impossible. This is because the subroutine in
4107 question is not used in operations with the private key itself and an input
4108 of the attacker's direct choice. Otherwise the bug can manifest itself as
4109 transient authentication and key negotiation failures or reproducible
4110 erroneous outcome of public-key operations with specially crafted input.
4111 Among EC algorithms only Brainpool P-512 curves are affected and one
4112 presumably can attack ECDH key negotiation. Impact was not analyzed in
4113 detail, because pre-requisites for attack are considered unlikely. Namely
4114 multiple clients have to choose the curve in question and the server has to
4115 share the private key among them, neither of which is default behaviour.
4116 Even then only clients that chose the curve will be affected.
4118 This issue was publicly reported as transient failures and was not
4119 initially recognized as a security issue. Thanks to Richard Morgan for
4120 providing reproducible case.
4125 * Removed automatic addition of RPATH in shared libraries and executables,
4126 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
4130 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
4132 * Fix Use After Free for large message sizes
4134 The patch applied to address CVE-2016-6307 resulted in an issue where if a
4135 message larger than approx 16k is received then the underlying buffer to
4136 store the incoming message is reallocated and moved. Unfortunately a
4137 dangling pointer to the old location is left which results in an attempt to
4138 write to the previously freed location. This is likely to result in a
4139 crash, however it could potentially lead to execution of arbitrary code.
4141 This issue only affects OpenSSL 1.1.0a.
4143 This issue was reported to OpenSSL by Robert Święcki.
4148 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
4150 * OCSP Status Request extension unbounded memory growth
4152 A malicious client can send an excessively large OCSP Status Request
4153 extension. If that client continually requests renegotiation, sending a
4154 large OCSP Status Request extension each time, then there will be unbounded
4155 memory growth on the server. This will eventually lead to a Denial Of
4156 Service attack through memory exhaustion. Servers with a default
4157 configuration are vulnerable even if they do not support OCSP. Builds using
4158 the "no-ocsp" build time option are not affected.
4160 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4165 * SSL_peek() hang on empty record
4167 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
4168 sends an empty record. This could be exploited by a malicious peer in a
4169 Denial Of Service attack.
4171 This issue was reported to OpenSSL by Alex Gaynor.
4176 * Excessive allocation of memory in tls_get_message_header() and
4177 dtls1_preprocess_fragment()
4179 A (D)TLS message includes 3 bytes for its length in the header for the
4180 message. This would allow for messages up to 16Mb in length. Messages of
4181 this length are excessive and OpenSSL includes a check to ensure that a
4182 peer is sending reasonably sized messages in order to avoid too much memory
4183 being consumed to service a connection. A flaw in the logic of version
4184 1.1.0 means that memory for the message is allocated too early, prior to
4185 the excessive message length check. Due to way memory is allocated in
4186 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4187 to service a connection. This could lead to a Denial of Service through
4188 memory exhaustion. However, the excessive message length check still takes
4189 place, and this would cause the connection to immediately fail. Assuming
4190 that the application calls SSL_free() on the failed connection in a timely
4191 manner then the 21Mb of allocated memory will then be immediately freed
4192 again. Therefore, the excessive memory allocation will be transitory in
4193 nature. This then means that there is only a security impact if:
4195 1) The application does not call SSL_free() in a timely manner in the event
4196 that the connection fails
4198 2) The application is working in a constrained environment where there is
4199 very little free memory
4201 3) The attacker initiates multiple connection attempts such that there are
4202 multiple connections in a state where memory has been allocated for the
4203 connection; SSL_free() has not yet been called; and there is insufficient
4204 memory to service the multiple requests.
4206 Except in the instance of (1) above any Denial Of Service is likely to be
4207 transitory because as soon as the connection fails the memory is
4208 subsequently freed again in the SSL_free() call. However there is an
4209 increased risk during this period of application crashes due to the lack of
4210 memory - which would then mean a more serious Denial of Service.
4212 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4213 (CVE-2016-6307 and CVE-2016-6308)
4217 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4218 had to be removed. Primary reason is that vendor assembler can't
4219 assemble our modules with -KPIC flag. As result it, assembly
4220 support, was not even available as option. But its lack means
4221 lack of side-channel resistant code, which is incompatible with
4222 security by todays standards. Fortunately gcc is readily available
4223 prepackaged option, which we firmly point at...
4227 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4229 * Windows command-line tool supports UTF-8 opt-in option for arguments
4230 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4231 (to any value) allows Windows user to access PKCS#12 file generated
4232 with Windows CryptoAPI and protected with non-ASCII password, as well
4233 as files generated under UTF-8 locale on Linux also protected with
4238 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4239 have been disabled by default and removed from DEFAULT, just like RC4.
4240 See the RC4 item below to re-enable both.
4244 * The method for finding the storage location for the Windows RAND seed file
4245 has changed. First we check %RANDFILE%. If that is not set then we check
4246 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4247 all else fails we fall back to C:\.
4251 * The EVP_EncryptUpdate() function has had its return type changed from void
4252 to int. A return of 0 indicates and error while a return of 1 indicates
4257 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4258 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4259 off the constant time implementation for RSA, DSA and DH have been made
4260 no-ops and deprecated.
4264 * Windows RAND implementation was simplified to only get entropy by
4265 calling CryptGenRandom(). Various other RAND-related tickets
4268 *Joseph Wylie Yandle, Rich Salz*
4270 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4271 and `OPENSSL_LH_`, respectively. The old names are available
4272 with API compatibility. They new names are now completely documented.
4276 * Unify TYPE_up_ref(obj) methods signature.
4277 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4278 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4279 int (instead of void) like all others TYPE_up_ref() methods.
4280 So now these methods also check the return value of CRYPTO_atomic_add(),
4281 and the validity of object reference counter.
4283 *fdasilvayy@gmail.com*
4285 * With Windows Visual Studio builds, the .pdb files are installed
4286 alongside the installed libraries and executables. For a static
4287 library installation, ossl_static.pdb is the associate compiler
4288 generated .pdb file to be used when linking programs.
4292 * Remove openssl.spec. Packaging files belong with the packagers.
4296 * Automatic Darwin/OSX configuration has had a refresh, it will now
4297 recognise x86_64 architectures automatically. You can still decide
4298 to build for a different bitness with the environment variable
4299 KERNEL_BITS (can be 32 or 64), for example:
4301 KERNEL_BITS=32 ./config
4305 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4306 256 bit AES and HMAC with SHA256.
4310 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4314 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4318 * To enable users to have their own config files and build file templates,
4319 Configure looks in the directory indicated by the environment variable
4320 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4321 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4322 name and is used as is.
4326 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4327 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4328 X509_CERT_FILE_CTX was removed.
4332 * "shared" builds are now the default. To create only static libraries use
4333 the "no-shared" Configure option.
4337 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4338 All of these option have not worked for some while and are fundamental
4343 * Make various cleanup routines no-ops and mark them as deprecated. Most
4344 global cleanup functions are no longer required because they are handled
4345 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4346 Explicitly de-initing can cause problems (e.g. where a library that uses
4347 OpenSSL de-inits, but an application is still using it). The affected
4348 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4349 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4350 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4351 COMP_zlib_cleanup().
4355 * --strict-warnings no longer enables runtime debugging options
4356 such as REF_DEBUG. Instead, debug options are automatically
4357 enabled with '--debug' builds.
4359 *Andy Polyakov, Emilia Käsper*
4361 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4362 have been moved out of the public header files. New functions for managing
4363 these have been added.
4367 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4368 objects have been moved out of the public header files. New
4369 functions for managing these have been added.
4373 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4374 have been moved out of the public header files. New functions for managing
4375 these have been added.
4379 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4380 moved out of the public header files. New functions for managing these
4385 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4389 * Removed the mk1mf build scripts.
4393 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4394 it is always safe to #include a header now.
4398 * Removed the aged BC-32 config and all its supporting scripts
4402 * Removed support for Ultrix, Netware, and OS/2.
4406 * Add support for HKDF.
4408 *Alessandro Ghedini*
4410 * Add support for blake2b and blake2s
4414 * Added support for "pipelining". Ciphers that have the
4415 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4416 encryptions/decryptions simultaneously. There are currently no built-in
4417 ciphers with this property but the expectation is that engines will be able
4418 to offer it to significantly improve throughput. Support has been extended
4419 into libssl so that multiple records for a single connection can be
4420 processed in one go (for >=TLS 1.1).
4424 * Added the AFALG engine. This is an async capable engine which is able to
4425 offload work to the Linux kernel. In this initial version it only supports
4426 AES128-CBC. The kernel must be version 4.1.0 or greater.
4430 * OpenSSL now uses a new threading API. It is no longer necessary to
4431 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4432 are two supported threading models: pthreads and windows threads. It is
4433 also possible to configure OpenSSL at compile time for "no-threads". The
4434 old threading API should no longer be used. The functions have been
4435 replaced with "no-op" compatibility macros.
4437 *Alessandro Ghedini, Matt Caswell*
4439 * Modify behavior of ALPN to invoke callback after SNI/servername
4440 callback, such that updates to the SSL_CTX affect ALPN.
4444 * Add SSL_CIPHER queries for authentication and key-exchange.
4448 * Changes to the DEFAULT cipherlist:
4449 - Prefer (EC)DHE handshakes over plain RSA.
4450 - Prefer AEAD ciphers over legacy ciphers.
4451 - Prefer ECDSA over RSA when both certificates are available.
4452 - Prefer TLSv1.2 ciphers/PRF.
4453 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4458 * Change the ECC default curve list to be this, in order: x25519,
4459 secp256r1, secp521r1, secp384r1.
4463 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4464 disabled by default. They can be re-enabled using the
4465 enable-weak-ssl-ciphers option to Configure.
4469 * If the server has ALPN configured, but supports no protocols that the
4470 client advertises, send a fatal "no_application_protocol" alert.
4471 This behaviour is SHALL in RFC 7301, though it isn't universally
4472 implemented by other servers.
4476 * Add X25519 support.
4477 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4478 for public and private key encoding using the format documented in
4479 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4480 key generation and key derivation.
4482 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4487 * Deprecate SRP_VBASE_get_by_user.
4488 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4489 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4490 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4491 seed, even if the seed is configured.
4493 Users should use SRP_VBASE_get1_by_user instead. Note that in
4494 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4495 also that even though configuring the SRP seed attempts to hide
4496 invalid usernames by continuing the handshake with fake
4497 credentials, this behaviour is not constant time and no strong
4498 guarantees are made that the handshake is indistinguishable from
4499 that of a valid user.
4503 * Configuration change; it's now possible to build dynamic engines
4504 without having to build shared libraries and vice versa. This
4505 only applies to the engines in `engines/`, those in `crypto/engine/`
4506 will always be built into libcrypto (i.e. "static").
4508 Building dynamic engines is enabled by default; to disable, use
4509 the configuration option "disable-dynamic-engine".
4511 The only requirements for building dynamic engines are the
4512 presence of the DSO module and building with position independent
4513 code, so they will also automatically be disabled if configuring
4514 with "disable-dso" or "disable-pic".
4516 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4517 are also taken away from openssl/opensslconf.h, as they are
4522 * Configuration change; if there is a known flag to compile
4523 position independent code, it will always be applied on the
4524 libcrypto and libssl object files, and never on the application
4525 object files. This means other libraries that use routines from
4526 libcrypto / libssl can be made into shared libraries regardless
4527 of how OpenSSL was configured.
4529 If this isn't desirable, the configuration options "disable-pic"
4530 or "no-pic" can be used to disable the use of PIC. This will
4531 also disable building shared libraries and dynamic engines.
4535 * Removed JPAKE code. It was experimental and has no wide use.
4539 * The INSTALL_PREFIX Makefile variable has been renamed to
4540 DESTDIR. That makes for less confusion on what this variable
4541 is for. Also, the configuration option --install_prefix is
4546 * Heartbeat for TLS has been removed and is disabled by default
4547 for DTLS; configure with enable-heartbeats. Code that uses the
4548 old #define's might need to be updated.
4550 *Emilia Käsper, Rich Salz*
4552 * Rename REF_CHECK to REF_DEBUG.
4556 * New "unified" build system
4558 The "unified" build system is aimed to be a common system for all
4559 platforms we support. With it comes new support for VMS.
4561 This system builds supports building in a different directory tree
4562 than the source tree. It produces one Makefile (for unix family
4563 or lookalikes), or one descrip.mms (for VMS).
4565 The source of information to make the Makefile / descrip.mms is
4566 small files called 'build.info', holding the necessary
4567 information for each directory with source to compile, and a
4568 template in Configurations, like unix-Makefile.tmpl or
4571 With this change, the library names were also renamed on Windows
4572 and on VMS. They now have names that are closer to the standard
4573 on Unix, and include the major version number, and in certain
4574 cases, the architecture they are built for. See "Notes on shared
4575 libraries" in INSTALL.
4577 We rely heavily on the perl module Text::Template.
4581 * Added support for auto-initialisation and de-initialisation of the library.
4582 OpenSSL no longer requires explicit init or deinit routines to be called,
4583 except in certain circumstances. See the OPENSSL_init_crypto() and
4584 OPENSSL_init_ssl() man pages for further information.
4588 * The arguments to the DTLSv1_listen function have changed. Specifically the
4589 "peer" argument is now expected to be a BIO_ADDR object.
4591 * Rewrite of BIO networking library. The BIO library lacked consistent
4592 support of IPv6, and adding it required some more extensive
4593 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4594 which hold all types of addresses and chains of address information.
4595 It also introduces a new API, with functions like BIO_socket,
4596 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4597 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4598 have been adapted accordingly.
4602 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4607 * CRIME protection: disable compression by default, even if OpenSSL is
4608 compiled with zlib enabled. Applications can still enable compression
4609 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4610 using the SSL_CONF library to configure compression.
4614 * The signature of the session callback configured with
4615 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4616 was explicitly marked as `const unsigned char*` instead of
4621 * Always DPURIFY. Remove the use of uninitialized memory in the
4622 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4626 * Removed many obsolete configuration items, including
4627 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4628 MD2_CHAR, MD2_INT, MD2_LONG
4630 IDEA_SHORT, IDEA_LONG
4631 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4633 *Rich Salz, with advice from Andy Polyakov*
4635 * Many BN internals have been moved to an internal header file.
4637 *Rich Salz with help from Andy Polyakov*
4639 * Configuration and writing out the results from it has changed.
4640 Files such as Makefile include/openssl/opensslconf.h and are now
4641 produced through general templates, such as Makefile.in and
4642 crypto/opensslconf.h.in and some help from the perl module
4645 Also, the center of configuration information is no longer
4646 Makefile. Instead, Configure produces a perl module in
4647 configdata.pm which holds most of the config data (in the hash
4648 table %config), the target data that comes from the target
4649 configuration in one of the `Configurations/*.conf` files (in
4654 * To clarify their intended purposes, the Configure options
4655 --prefix and --openssldir change their semantics, and become more
4656 straightforward and less interdependent.
4658 --prefix shall be used exclusively to give the location INSTALLTOP
4659 where programs, scripts, libraries, include files and manuals are
4660 going to be installed. The default is now /usr/local.
4662 --openssldir shall be used exclusively to give the default
4663 location OPENSSLDIR where certificates, private keys, CRLs are
4664 managed. This is also where the default openssl.cnf gets
4666 If the directory given with this option is a relative path, the
4667 values of both the --prefix value and the --openssldir value will
4668 be combined to become OPENSSLDIR.
4669 The default for --openssldir is INSTALLTOP/ssl.
4671 Anyone who uses --openssldir to specify where OpenSSL is to be
4672 installed MUST change to use --prefix instead.
4676 * The GOST engine was out of date and therefore it has been removed. An up
4677 to date GOST engine is now being maintained in an external repository.
4678 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4679 support for GOST ciphersuites (these are only activated if a GOST engine
4684 * EGD is no longer supported by default; use enable-egd when
4687 *Ben Kaduk and Rich Salz*
4689 * The distribution now has Makefile.in files, which are used to
4690 create Makefile's when Configure is run. *Configure must be run
4691 before trying to build now.*
4695 * The return value for SSL_CIPHER_description() for error conditions
4700 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4702 Obtaining and performing DNSSEC validation of TLSA records is
4703 the application's responsibility. The application provides
4704 the TLSA records of its choice to OpenSSL, and these are then
4705 used to authenticate the peer.
4707 The TLSA records need not even come from DNS. They can, for
4708 example, be used to implement local end-entity certificate or
4709 trust-anchor "pinning", where the "pin" data takes the form
4710 of TLSA records, which can augment or replace verification
4711 based on the usual WebPKI public certification authorities.
4715 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4716 continues to support deprecated interfaces in default builds.
4717 However, applications are strongly advised to compile their
4718 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4719 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4720 or the 1.1.0 releases.
4722 In environments in which all applications have been ported to
4723 not use any deprecated interfaces OpenSSL's Configure script
4724 should be used with the --api=1.1.0 option to entirely remove
4725 support for the deprecated features from the library and
4726 unconditionally disable them in the installed headers.
4727 Essentially the same effect can be achieved with the "no-deprecated"
4728 argument to Configure, except that this will always restrict
4729 the build to just the latest API, rather than a fixed API
4732 As applications are ported to future revisions of the API,
4733 they should update their compile-time OPENSSL_API_COMPAT define
4734 accordingly, but in most cases should be able to continue to
4735 compile with later releases.
4737 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4738 0x10000000L and 0x00908000L, respectively. However those
4739 versions did not support the OPENSSL_API_COMPAT feature, and
4740 so applications are not typically tested for explicit support
4741 of just the undeprecated features of either release.
4745 * Add support for setting the minimum and maximum supported protocol.
4746 It can bet set via the SSL_set_min_proto_version() and
4747 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4748 MaxProtocol. It's recommended to use the new APIs to disable
4749 protocols instead of disabling individual protocols using
4750 SSL_set_options() or SSL_CONF's Protocol. This change also
4751 removes support for disabling TLS 1.2 in the OpenSSL TLS
4752 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4756 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4760 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4761 and integrates ECDSA and ECDH functionality into EC. Implementations can
4762 now redirect key generation and no longer need to convert to or from
4765 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4766 include the ec.h header file instead.
4770 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4771 ciphers who are no longer supported and drops support the ephemeral RSA key
4772 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4776 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4777 opaque. For HMAC_CTX, the following constructors and destructors
4780 HMAC_CTX *HMAC_CTX_new(void);
4781 void HMAC_CTX_free(HMAC_CTX *ctx);
4783 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4784 destroy such methods has been added. See EVP_MD_meth_new(3) and
4785 EVP_CIPHER_meth_new(3) for documentation.
4788 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4789 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4790 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4791 an already created structure.
4792 2) For consistency with the majority of our object creators and
4793 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4794 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4795 for deprecated builds.
4799 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4800 cryptographic operations to be performed asynchronously as long as an
4801 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4802 further details. Libssl has also had this capability integrated with the
4803 introduction of the new mode SSL_MODE_ASYNC and associated error
4804 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4805 pages. This work was developed in partnership with Intel Corp.
4809 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4810 always enabled now. If you want to disable the support you should
4811 exclude it using the list of supported ciphers. This also means that the
4812 "-no_ecdhe" option has been removed from s_server.
4816 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4817 SSL_{CTX_}set1_curves() which can set a list.
4821 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4822 curve you want to support using SSL_{CTX_}set1_curves().
4826 * State machine rewrite. The state machine code has been significantly
4827 refactored in order to remove much duplication of code and solve issues
4828 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4829 further details). This change does have some associated API changes.
4830 Notably the SSL_state() function has been removed and replaced by
4831 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4832 SSL_set_state() has been removed altogether. The previous handshake states
4833 defined in ssl.h and ssl3.h have also been removed.
4837 * All instances of the string "ssleay" in the public API were replaced
4838 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4839 Some error codes related to internal RSA_eay API's were renamed.
4843 * The demo files in crypto/threads were moved to demo/threads.
4847 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4850 *Matt Caswell, Rich Salz*
4852 * New ASN.1 embed macro.
4854 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4855 structure is not allocated: it is part of the parent. That is instead of
4863 This reduces memory fragmentation and make it impossible to accidentally
4864 set a mandatory field to NULL.
4866 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4867 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4868 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4873 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4877 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4878 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4879 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4880 DES and RC4 ciphersuites.
4884 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4885 This changes the decoding behaviour for some invalid messages,
4886 though the change is mostly in the more lenient direction, and
4887 legacy behaviour is preserved as much as possible.
4891 * Fix no-stdio build.
4892 *David Woodhouse <David.Woodhouse@intel.com> and also*
4893 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4895 * New testing framework
4896 The testing framework has been largely rewritten and is now using
4897 perl and the perl modules Test::Harness and an extended variant of
4898 Test::More called OpenSSL::Test to do its work. All test scripts in
4899 test/ have been rewritten into test recipes, and all direct calls to
4900 executables in test/Makefile have become individual recipes using the
4901 simplified testing OpenSSL::Test::Simple.
4903 For documentation on our testing modules, do:
4905 perldoc test/testlib/OpenSSL/Test/Simple.pm
4906 perldoc test/testlib/OpenSSL/Test.pm
4910 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4911 are used; the latter aborts on memory leaks (usually checked on exit).
4912 Some undocumented "set malloc, etc., hooks" functions were removed
4913 and others were changed. All are now documented.
4917 * In DSA_generate_parameters_ex, if the provided seed is too short,
4920 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4922 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4923 from RFC4279, RFC4785, RFC5487, RFC5489.
4925 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4926 original RSA_PSK patch.
4930 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4931 era flag was never set throughout the codebase (only read). Also removed
4932 SSL3_FLAGS_POP_BUFFER which was only used if
4933 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4937 * Changed the default name options in the "ca", "crl", "req" and "x509"
4938 to be "oneline" instead of "compat".
4942 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4943 not aware of clients that still exhibit this bug, and the workaround
4944 hasn't been working properly for a while.
4948 * The return type of BIO_number_read() and BIO_number_written() as well as
4949 the corresponding num_read and num_write members in the BIO structure has
4950 changed from unsigned long to uint64_t. On platforms where an unsigned
4951 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4956 * Given the pervasive nature of TLS extensions it is inadvisable to run
4957 OpenSSL without support for them. It also means that maintaining
4958 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4959 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4963 * Removed support for the two export grade static DH ciphersuites
4964 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4965 were newly added (along with a number of other static DH ciphersuites) to
4966 1.0.2. However the two export ones have *never* worked since they were
4967 introduced. It seems strange in any case to be adding new export
4968 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4972 * Version negotiation has been rewritten. In particular SSLv23_method(),
4973 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4974 and turned into macros which simply call the new preferred function names
4975 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4976 should use the new names instead. Also as part of this change the ssl23.h
4977 header file has been removed.
4981 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4982 code and the associated standard is no longer considered fit-for-purpose.
4986 * RT2547 was closed. When generating a private key, try to make the
4987 output file readable only by the owner. This behavior change might
4988 be noticeable when interacting with other software.
4990 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4995 * Added HTTP GET support to the ocsp command.
4999 * Changed default digest for the dgst and enc commands from MD5 to
5004 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
5008 * Added support for TLS extended master secret from
5009 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
5010 initial patch which was a great help during development.
5014 * All libssl internal structures have been removed from the public header
5015 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
5016 now redundant). Users should not attempt to access internal structures
5017 directly. Instead they should use the provided API functions.
5021 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
5022 Access to deprecated functions can be re-enabled by running config with
5023 "enable-deprecated". In addition applications wishing to use deprecated
5024 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
5025 will, by default, disable some transitive includes that previously existed
5026 in the header files (e.g. ec.h will no longer, by default, include bn.h)
5030 * Added support for OCB mode. OpenSSL has been granted a patent license
5031 compatible with the OpenSSL license for use of OCB. Details are available
5032 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
5033 for OCB can be removed by calling config with no-ocb.
5037 * SSLv2 support has been removed. It still supports receiving an SSLv2
5038 compatible client hello.
5042 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
5043 done while fixing the error code for the key-too-small case.
5045 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
5047 * CA.sh has been removed; use CA.pl instead.
5051 * Removed old DES API.
5055 * Remove various unsupported platforms:
5061 Sinix/ReliantUNIX RM400
5066 16-bit platforms such as WIN16
5070 * Clean up OPENSSL_NO_xxx #define's
5071 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
5072 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
5073 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
5074 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
5075 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
5076 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
5077 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
5078 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
5079 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
5080 - Remove MS_STATIC; it's a relic from platforms <32 bits.
5084 * Cleaned up dead code
5085 Remove all but one '#ifdef undef' which is to be looked at.
5089 * Clean up calling of xxx_free routines.
5090 Just like free(), fix most of the xxx_free routines to accept
5091 NULL. Remove the non-null checks from callers. Save much code.
5095 * Add secure heap for storage of private keys (when possible).
5096 Add BIO_s_secmem(), CBIGNUM, etc.
5097 Contributed by Akamai Technologies under our Corporate CLA.
5101 * Experimental support for a new, fast, unbiased prime candidate generator,
5102 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
5104 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
5106 * New output format NSS in the sess_id command line tool. This allows
5107 exporting the session id and the master key in NSS keylog format.
5109 *Martin Kaiser <martin@kaiser.cx>*
5111 * Harmonize version and its documentation. -f flag is used to display
5114 *mancha <mancha1@zoho.com>*
5116 * Fix eckey_priv_encode so it immediately returns an error upon a failure
5117 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
5119 *mancha <mancha1@zoho.com>*
5121 * Fix some double frees. These are not thought to be exploitable.
5123 *mancha <mancha1@zoho.com>*
5125 * A missing bounds check in the handling of the TLS heartbeat extension
5126 can be used to reveal up to 64k of memory to a connected client or
5129 Thanks for Neel Mehta of Google Security for discovering this bug and to
5130 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
5131 preparing the fix ([CVE-2014-0160])
5133 *Adam Langley, Bodo Moeller*
5135 * Fix for the attack described in the paper "Recovering OpenSSL
5136 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
5137 by Yuval Yarom and Naomi Benger. Details can be obtained from:
5138 <http://eprint.iacr.org/2014/140>
5140 Thanks to Yuval Yarom and Naomi Benger for discovering this
5141 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5143 *Yuval Yarom and Naomi Benger*
5145 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5146 this fixes a limitation in previous versions of OpenSSL.
5150 * Experimental encrypt-then-mac support.
5152 Experimental support for encrypt then mac from
5153 draft-gutmann-tls-encrypt-then-mac-02.txt
5155 To enable it set the appropriate extension number (0x42 for the test
5156 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
5158 For non-compliant peers (i.e. just about everything) this should have no
5161 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
5165 * Add EVP support for key wrapping algorithms, to avoid problems with
5166 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5167 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5168 algorithms and include tests cases.
5172 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
5177 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5178 MGF1 digest and OAEP label.
5182 * Make openssl verify return errors.
5184 *Chris Palmer <palmer@google.com> and Ben Laurie*
5186 * New function ASN1_TIME_diff to calculate the difference between two
5187 ASN1_TIME structures or one structure and the current time.
5191 * Update fips_test_suite to support multiple command line options. New
5192 test to induce all self test errors in sequence and check expected
5197 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5198 sign or verify all in one operation.
5202 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5203 test programs and fips_test_suite. Includes functionality to parse
5204 the minimal script output of fipsalgest.pl directly.
5208 * Add authorisation parameter to FIPS_module_mode_set().
5212 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5216 * Use separate DRBG fields for internal and external flags. New function
5217 FIPS_drbg_health_check() to perform on demand health checking. Add
5218 generation tests to fips_test_suite with reduced health check interval to
5219 demonstrate periodic health checking. Add "nodh" option to
5220 fips_test_suite to skip very slow DH test.
5224 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5229 * More extensive health check for DRBG checking many more failure modes.
5230 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5231 combination: call this in fips_test_suite.
5235 * Add support for canonical generation of DSA parameter 'g'. See
5238 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5239 POST to handle HMAC cases.
5243 * Add functions FIPS_module_version() and FIPS_module_version_text()
5244 to return numerical and string versions of the FIPS module number.
5248 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5249 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5250 outside the validated module in the FIPS capable OpenSSL.
5254 * Minor change to DRBG entropy callback semantics. In some cases
5255 there is no multiple of the block length between min_len and
5256 max_len. Allow the callback to return more than max_len bytes
5257 of entropy but discard any extra: it is the callback's responsibility
5258 to ensure that the extra data discarded does not impact the
5259 requested amount of entropy.
5263 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5264 information in FIPS186-3, SP800-57 and SP800-131A.
5268 * CCM support via EVP. Interface is very similar to GCM case except we
5269 must supply all data in one chunk (i.e. no update, final) and the
5270 message length must be supplied if AAD is used. Add algorithm test
5275 * Initial version of POST overhaul. Add POST callback to allow the status
5276 of POST to be monitored and/or failures induced. Modify fips_test_suite
5277 to use callback. Always run all selftests even if one fails.
5281 * XTS support including algorithm test driver in the fips_gcmtest program.
5282 Note: this does increase the maximum key length from 32 to 64 bytes but
5283 there should be no binary compatibility issues as existing applications
5284 will never use XTS mode.
5288 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5289 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5290 performs algorithm blocking for unapproved PRNG types. Also do not
5291 set PRNG type in FIPS_mode_set(): leave this to the application.
5292 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5293 the standard OpenSSL PRNG: set additional data to a date time vector.
5297 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5298 This shouldn't present any incompatibility problems because applications
5299 shouldn't be using these directly and any that are will need to rethink
5300 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5304 * Extensive self tests and health checking required by SP800-90 DRBG.
5305 Remove strength parameter from FIPS_drbg_instantiate and always
5306 instantiate at maximum supported strength.
5310 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5314 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5318 * New function DH_compute_key_padded() to compute a DH key and pad with
5319 leading zeroes if needed: this complies with SP800-56A et al.
5323 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5324 anything, incomplete, subject to change and largely untested at present.
5328 * Modify fipscanisteronly build option to only build the necessary object
5329 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5333 * Add experimental option FIPSSYMS to give all symbols in
5334 fipscanister.o and FIPS or fips prefix. This will avoid
5335 conflicts with future versions of OpenSSL. Add perl script
5336 util/fipsas.pl to preprocess assembly language source files
5337 and rename any affected symbols.
5341 * Add selftest checks and algorithm block of non-fips algorithms in
5342 FIPS mode. Remove DES2 from selftests.
5346 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5347 return internal method without any ENGINE dependencies. Add new
5348 tiny fips sign and verify functions.
5352 * New build option no-ec2m to disable characteristic 2 code.
5356 * New build option "fipscanisteronly". This only builds fipscanister.o
5357 and (currently) associated fips utilities. Uses the file Makefile.fips
5358 instead of Makefile.org as the prototype.
5362 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5363 Update fips_gcmtest to use IV generator.
5367 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5368 setting output buffer to NULL. The `*Final` function must be
5369 called although it will not retrieve any additional data. The tag
5370 can be set or retrieved with a ctrl. The IV length is by default 12
5371 bytes (96 bits) but can be set to an alternative value. If the IV
5372 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5377 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5378 underlying do_cipher function handles all cipher semantics itself
5379 including padding and finalisation. This is useful if (for example)
5380 an ENGINE cipher handles block padding itself. The behaviour of
5381 do_cipher is subtly changed if this flag is set: the return value
5382 is the number of characters written to the output buffer (zero is
5383 no longer an error code) or a negative error code. Also if the
5384 input buffer is NULL and length 0 finalisation should be performed.
5388 * If a candidate issuer certificate is already part of the constructed
5389 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5393 * Improve forward-security support: add functions
5395 void SSL_CTX_set_not_resumable_session_callback(
5396 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5397 void SSL_set_not_resumable_session_callback(
5398 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5400 for use by SSL/TLS servers; the callback function will be called whenever a
5401 new session is created, and gets to decide whether the session may be
5402 cached to make it resumable (return 0) or not (return 1). (As by the
5403 SSL/TLS protocol specifications, the session_id sent by the server will be
5404 empty to indicate that the session is not resumable; also, the server will
5405 not generate RFC 4507 (RFC 5077) session tickets.)
5407 A simple reasonable callback implementation is to return is_forward_secure.
5408 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5409 by the SSL/TLS server library, indicating whether it can provide forward
5412 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5414 * New -verify_name option in command line utilities to set verification
5419 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5420 Add CMAC pkey methods.
5424 * Experimental renegotiation in s_server -www mode. If the client
5425 browses /reneg connection is renegotiated. If /renegcert it is
5426 renegotiated requesting a certificate.
5430 * Add an "external" session cache for debugging purposes to s_server. This
5431 should help trace issues which normally are only apparent in deployed
5432 multi-process servers.
5436 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5437 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5438 BIO_set_cipher() and some obscure PEM functions were changed so they
5439 can now return an error. The RAND changes required a change to the
5440 RAND_METHOD structure.
5444 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5445 a gcc attribute to warn if the result of a function is ignored. This
5446 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5447 whose return value is often ignored.
5451 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5452 These allow SCTs (signed certificate timestamps) to be requested and
5453 validated when establishing a connection.
5455 *Rob Percival <robpercival@google.com>*
5460 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5462 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5463 used even when parsing explicit parameters, when loading a encoded key
5464 or calling `EC_GROUP_new_from_ecpkparameters()`/
5465 `EC_GROUP_new_from_ecparameters()`.
5466 This prevents bypass of security hardening and performance gains,
5467 especially for curves with specialized EC_METHODs.
5468 By default, if a key encoded with explicit parameters is loaded and later
5469 encoded, the output is still encoded with explicit parameters, even if
5470 internally a "named" EC_GROUP is used for computation.
5474 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5475 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5476 NULL. After this change, only the cofactor parameter can be NULL. It also
5477 does some minimal sanity checks on the passed order.
5482 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5483 An attack is simple, if the first CMS_recipientInfo is valid but the
5484 second CMS_recipientInfo is chosen ciphertext. If the second
5485 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5486 encryption key will be replaced by garbage, and the message cannot be
5487 decoded, but if the RSA decryption fails, the correct encryption key is
5488 used and the recipient will not notice the attack.
5489 As a work around for this potential attack the length of the decrypted
5490 key must be equal to the cipher default key length, in case the
5491 certificate is not given and all recipientInfo are tried out.
5492 The old behaviour can be re-enabled in the CMS code by setting the
5493 CMS_DEBUG_DECRYPT flag.
5498 * Document issue with installation paths in diverse Windows builds
5500 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5501 binaries and run-time config file.
5506 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5508 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5509 This changes the size when using the `genpkey` command when no size is given.
5510 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5511 generation commands to use 2048 bits by default.
5515 * Add FIPS support for Android Arm 64-bit
5517 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5518 Module in Version 2.0.10. For some reason, the corresponding target
5519 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5520 built with FIPS support on Android Arm 64-bit. This omission has been
5523 *Matthias St. Pierre*
5525 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5527 * 0-byte record padding oracle
5529 If an application encounters a fatal protocol error and then calls
5530 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5531 then OpenSSL can respond differently to the calling application if a 0 byte
5532 record is received with invalid padding compared to if a 0 byte record is
5533 received with an invalid MAC. If the application then behaves differently
5534 based on that in a way that is detectable to the remote peer, then this
5535 amounts to a padding oracle that could be used to decrypt data.
5537 In order for this to be exploitable "non-stitched" ciphersuites must be in
5538 use. Stitched ciphersuites are optimised implementations of certain
5539 commonly used ciphersuites. Also the application must call SSL_shutdown()
5540 twice even if a protocol error has occurred (applications should not do
5541 this but some do anyway).
5543 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5544 Aviram, with additional investigation by Steven Collison and Andrew
5545 Hourselt. It was reported to OpenSSL on 10th December 2018.
5550 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5554 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5556 * Microarchitecture timing vulnerability in ECC scalar multiplication
5558 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5559 shown to be vulnerable to a microarchitecture timing side channel attack.
5560 An attacker with sufficient access to mount local timing attacks during
5561 ECDSA signature generation could recover the private key.
5563 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5564 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5570 * Timing vulnerability in DSA signature generation
5572 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5573 timing side channel attack. An attacker could use variations in the signing
5574 algorithm to recover the private key.
5576 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5581 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5582 Module, accidentally introduced while backporting security fixes from the
5583 development branch and hindering the use of ECC in FIPS mode.
5587 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5589 * Client DoS due to large DH parameter
5591 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5592 malicious server can send a very large prime value to the client. This will
5593 cause the client to spend an unreasonably long period of time generating a
5594 key for this prime resulting in a hang until the client has finished. This
5595 could be exploited in a Denial Of Service attack.
5597 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5602 * Cache timing vulnerability in RSA Key Generation
5604 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5605 a cache timing side channel attack. An attacker with sufficient access to
5606 mount cache timing attacks during the RSA key generation process could
5607 recover the private key.
5609 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5610 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5615 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5616 parameter is no longer accepted, as it leads to a corrupt table. NULL
5617 pem_str is reserved for alias entries only.
5621 * Revert blinding in ECDSA sign and instead make problematic addition
5622 length-invariant. Switch even to fixed-length Montgomery multiplication.
5626 * Change generating and checking of primes so that the error rate of not
5627 being prime depends on the intended use based on the size of the input.
5628 For larger primes this will result in more rounds of Miller-Rabin.
5629 The maximal error rate for primes with more than 1080 bits is lowered
5632 *Kurt Roeckx, Annie Yousar*
5634 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5638 * Add blinding to ECDSA and DSA signatures to protect against side channel
5639 attacks discovered by Keegan Ryan (NCC Group).
5643 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5644 now allow empty (zero character) pass phrases.
5648 * Certificate time validation (X509_cmp_time) enforces stricter
5649 compliance with RFC 5280. Fractional seconds and timezone offsets
5650 are no longer allowed.
5654 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5656 * Constructed ASN.1 types with a recursive definition could exceed the stack
5658 Constructed ASN.1 types with a recursive definition (such as can be found
5659 in PKCS7) could eventually exceed the stack given malicious input with
5660 excessive recursion. This could result in a Denial Of Service attack. There
5661 are no such structures used within SSL/TLS that come from untrusted sources
5662 so this is considered safe.
5664 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5670 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5672 * Read/write after SSL object in error state
5674 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5675 mechanism. The intent was that if a fatal error occurred during a handshake
5676 then OpenSSL would move into the error state and would immediately fail if
5677 you attempted to continue the handshake. This works as designed for the
5678 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5679 SSL_connect()), however due to a bug it does not work correctly if
5680 SSL_read() or SSL_write() is called directly. In that scenario, if the
5681 handshake fails then a fatal error will be returned in the initial function
5682 call. If SSL_read()/SSL_write() is subsequently called by the application
5683 for the same SSL object then it will succeed and the data is passed without
5684 being decrypted/encrypted directly from the SSL/TLS record layer.
5686 In order to exploit this issue an application bug would have to be present
5687 that resulted in a call to SSL_read()/SSL_write() being issued after having
5688 already received a fatal error.
5690 This issue was reported to OpenSSL by David Benjamin (Google).
5695 * rsaz_1024_mul_avx2 overflow bug on x86_64
5697 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5698 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5699 Analysis suggests that attacks against RSA and DSA as a result of this
5700 defect would be very difficult to perform and are not believed likely.
5701 Attacks against DH1024 are considered just feasible, because most of the
5702 work necessary to deduce information about a private key may be performed
5703 offline. The amount of resources required for such an attack would be
5704 significant. However, for an attack on TLS to be meaningful, the server
5705 would have to share the DH1024 private key among multiple clients, which is
5706 no longer an option since CVE-2016-0701.
5708 This only affects processors that support the AVX2 but not ADX extensions
5709 like Intel Haswell (4th generation).
5711 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5712 was originally found via the OSS-Fuzz project.
5717 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5719 * bn_sqrx8x_internal carry bug on x86_64
5721 There is a carry propagating bug in the x86_64 Montgomery squaring
5722 procedure. No EC algorithms are affected. Analysis suggests that attacks
5723 against RSA and DSA as a result of this defect would be very difficult to
5724 perform and are not believed likely. Attacks against DH are considered just
5725 feasible (although very difficult) because most of the work necessary to
5726 deduce information about a private key may be performed offline. The amount
5727 of resources required for such an attack would be very significant and
5728 likely only accessible to a limited number of attackers. An attacker would
5729 additionally need online access to an unpatched system using the target
5730 private key in a scenario with persistent DH parameters and a private
5731 key that is shared between multiple clients.
5733 This only affects processors that support the BMI1, BMI2 and ADX extensions
5734 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5736 This issue was reported to OpenSSL by the OSS-Fuzz project.
5741 * Malformed X.509 IPAddressFamily could cause OOB read
5743 If an X.509 certificate has a malformed IPAddressFamily extension,
5744 OpenSSL could do a one-byte buffer overread. The most likely result
5745 would be an erroneous display of the certificate in text format.
5747 This issue was reported to OpenSSL by the OSS-Fuzz project.
5751 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5753 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5754 platform rather than 'mingw'.
5758 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5760 * Truncated packet could crash via OOB read
5762 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5763 cipher is being used, then a truncated packet can cause that host to
5764 perform an out-of-bounds read, usually resulting in a crash.
5766 This issue was reported to OpenSSL by Robert Święcki of Google.
5771 * BN_mod_exp may produce incorrect results on x86_64
5773 There is a carry propagating bug in the x86_64 Montgomery squaring
5774 procedure. No EC algorithms are affected. Analysis suggests that attacks
5775 against RSA and DSA as a result of this defect would be very difficult to
5776 perform and are not believed likely. Attacks against DH are considered just
5777 feasible (although very difficult) because most of the work necessary to
5778 deduce information about a private key may be performed offline. The amount
5779 of resources required for such an attack would be very significant and
5780 likely only accessible to a limited number of attackers. An attacker would
5781 additionally need online access to an unpatched system using the target
5782 private key in a scenario with persistent DH parameters and a private
5783 key that is shared between multiple clients. For example this can occur by
5784 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5785 similar to CVE-2015-3193 but must be treated as a separate problem.
5787 This issue was reported to OpenSSL by the OSS-Fuzz project.
5792 * Montgomery multiplication may produce incorrect results
5794 There is a carry propagating bug in the Broadwell-specific Montgomery
5795 multiplication procedure that handles input lengths divisible by, but
5796 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5797 and DH private keys are impossible. This is because the subroutine in
5798 question is not used in operations with the private key itself and an input
5799 of the attacker's direct choice. Otherwise the bug can manifest itself as
5800 transient authentication and key negotiation failures or reproducible
5801 erroneous outcome of public-key operations with specially crafted input.
5802 Among EC algorithms only Brainpool P-512 curves are affected and one
5803 presumably can attack ECDH key negotiation. Impact was not analyzed in
5804 detail, because pre-requisites for attack are considered unlikely. Namely
5805 multiple clients have to choose the curve in question and the server has to
5806 share the private key among them, neither of which is default behaviour.
5807 Even then only clients that chose the curve will be affected.
5809 This issue was publicly reported as transient failures and was not
5810 initially recognized as a security issue. Thanks to Richard Morgan for
5811 providing reproducible case.
5816 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5817 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5818 prevent issues where no progress is being made and the peer continually
5819 sends unrecognised record types, using up resources processing them.
5823 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5825 * Missing CRL sanity check
5827 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5828 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5829 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5831 This issue only affects the OpenSSL 1.0.2i
5836 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5838 * OCSP Status Request extension unbounded memory growth
5840 A malicious client can send an excessively large OCSP Status Request
5841 extension. If that client continually requests renegotiation, sending a
5842 large OCSP Status Request extension each time, then there will be unbounded
5843 memory growth on the server. This will eventually lead to a Denial Of
5844 Service attack through memory exhaustion. Servers with a default
5845 configuration are vulnerable even if they do not support OCSP. Builds using
5846 the "no-ocsp" build time option are not affected.
5848 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5853 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5856 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5862 * OOB write in MDC2_Update()
5864 An overflow can occur in MDC2_Update() either if called directly or
5865 through the EVP_DigestUpdate() function using MDC2. If an attacker
5866 is able to supply very large amounts of input data after a previous
5867 call to EVP_EncryptUpdate() with a partial block then a length check
5868 can overflow resulting in a heap corruption.
5870 The amount of data needed is comparable to SIZE_MAX which is impractical
5873 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5878 * Malformed SHA512 ticket DoS
5880 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5881 DoS attack where a malformed ticket will result in an OOB read which will
5884 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5885 a custom server callback and ticket lookup mechanism.
5887 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5892 * OOB write in BN_bn2dec()
5894 The function BN_bn2dec() does not check the return value of BN_div_word().
5895 This can cause an OOB write if an application uses this function with an
5896 overly large BIGNUM. This could be a problem if an overly large certificate
5897 or CRL is printed out from an untrusted source. TLS is not affected because
5898 record limits will reject an oversized certificate before it is parsed.
5900 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5905 * OOB read in TS_OBJ_print_bio()
5907 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5908 the total length the OID text representation would use and not the amount
5909 of data written. This will result in OOB reads when large OIDs are
5912 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5917 * Pointer arithmetic undefined behaviour
5919 Avoid some undefined pointer arithmetic
5921 A common idiom in the codebase is to check limits in the following manner:
5924 Where "p" points to some malloc'd data of SIZE bytes and
5927 "len" here could be from some externally supplied data (e.g. from a TLS
5930 The rules of C pointer arithmetic are such that "p + len" is only well
5931 defined where len <= SIZE. Therefore the above idiom is actually
5932 undefined behaviour.
5934 For example this could cause problems if some malloc implementation
5935 provides an address for "p" such that "p + len" actually overflows for
5936 values of len that are too big and therefore p + len < limit.
5938 This issue was reported to OpenSSL by Guido Vranken
5943 * Constant time flag not preserved in DSA signing
5945 Operations in the DSA signing algorithm should run in constant time in
5946 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5947 implementation means that a non-constant time codepath is followed for
5948 certain operations. This has been demonstrated through a cache-timing
5949 attack to be sufficient for an attacker to recover the private DSA key.
5951 This issue was reported by César Pereida (Aalto University), Billy Brumley
5952 (Tampere University of Technology), and Yuval Yarom (The University of
5953 Adelaide and NICTA).
5958 * DTLS buffered message DoS
5960 In a DTLS connection where handshake messages are delivered out-of-order
5961 those messages that OpenSSL is not yet ready to process will be buffered
5962 for later use. Under certain circumstances, a flaw in the logic means that
5963 those messages do not get removed from the buffer even though the handshake
5964 has been completed. An attacker could force up to approx. 15 messages to
5965 remain in the buffer when they are no longer required. These messages will
5966 be cleared when the DTLS connection is closed. The default maximum size for
5967 a message is 100k. Therefore, the attacker could force an additional 1500k
5968 to be consumed per connection. By opening many simultaneous connections an
5969 attacker could cause a DoS attack through memory exhaustion.
5971 This issue was reported to OpenSSL by Quan Luo.
5976 * DTLS replay protection DoS
5978 A flaw in the DTLS replay attack protection mechanism means that records
5979 that arrive for future epochs update the replay protection "window" before
5980 the MAC for the record has been validated. This could be exploited by an
5981 attacker by sending a record for the next epoch (which does not have to
5982 decrypt or have a valid MAC), with a very large sequence number. This means
5983 that all subsequent legitimate packets are dropped causing a denial of
5984 service for a specific DTLS connection.
5986 This issue was reported to OpenSSL by the OCAP audit team.
5991 * Certificate message OOB reads
5993 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5994 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5995 theoretical DoS risk but this has not been observed in practice on common
5998 The messages affected are client certificate, client certificate request
5999 and server certificate. As a result the attack can only be performed
6000 against a client or a server which enables client authentication.
6002 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6007 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
6009 * Prevent padding oracle in AES-NI CBC MAC check
6011 A MITM attacker can use a padding oracle attack to decrypt traffic
6012 when the connection uses an AES CBC cipher and the server support
6015 This issue was introduced as part of the fix for Lucky 13 padding
6016 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6017 constant time by making sure that always the same bytes are read and
6018 compared against either the MAC or padding bytes. But it no longer
6019 checked that there was enough data to have both the MAC and padding
6022 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6026 * Fix EVP_EncodeUpdate overflow
6028 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6029 Base64 encoding of binary data. If an attacker is able to supply very large
6030 amounts of input data then a length check can overflow resulting in a heap
6033 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6034 the `PEM_write_bio*` family of functions. These are mainly used within the
6035 OpenSSL command line applications, so any application which processes data
6036 from an untrusted source and outputs it as a PEM file should be considered
6037 vulnerable to this issue. User applications that call these APIs directly
6038 with large amounts of untrusted data may also be vulnerable.
6040 This issue was reported by Guido Vranken.
6045 * Fix EVP_EncryptUpdate overflow
6047 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6048 is able to supply very large amounts of input data after a previous call to
6049 EVP_EncryptUpdate() with a partial block then a length check can overflow
6050 resulting in a heap corruption. Following an analysis of all OpenSSL
6051 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6052 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6053 the first called function after an EVP_EncryptInit(), and therefore that
6054 specific call must be safe. The second form is where the length passed to
6055 EVP_EncryptUpdate() can be seen from the code to be some small value and
6056 therefore there is no possibility of an overflow. Since all instances are
6057 one of these two forms, it is believed that there can be no overflows in
6058 internal code due to this problem. It should be noted that
6059 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6060 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6061 of these calls have also been analysed too and it is believed there are no
6062 instances in internal usage where an overflow could occur.
6064 This issue was reported by Guido Vranken.
6069 * Prevent ASN.1 BIO excessive memory allocation
6071 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6072 a short invalid encoding can cause allocation of large amounts of memory
6073 potentially consuming excessive resources or exhausting memory.
6075 Any application parsing untrusted data through d2i BIO functions is
6076 affected. The memory based functions such as d2i_X509() are *not* affected.
6077 Since the memory based functions are used by the TLS library, TLS
6078 applications are not affected.
6080 This issue was reported by Brian Carpenter.
6087 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6088 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6089 in arbitrary stack data being returned in the buffer.
6091 This issue was reported by Guido Vranken.
6096 * Modify behavior of ALPN to invoke callback after SNI/servername
6097 callback, such that updates to the SSL_CTX affect ALPN.
6101 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6106 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6107 methods are enabled and ssl2 is disabled the methods return NULL.
6111 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
6113 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6114 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6115 provide any "EXPORT" or "LOW" strength ciphers.
6119 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6120 is by default disabled at build-time. Builds that are not configured with
6121 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6122 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6123 will need to explicitly call either of:
6125 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6127 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6129 as appropriate. Even if either of those is used, or the application
6130 explicitly uses the version-specific SSLv2_method() or its client and
6131 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6132 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6133 ciphers, and SSLv2 56-bit DES are no longer available.
6138 * Fix a double-free in DSA code
6140 A double free bug was discovered when OpenSSL parses malformed DSA private
6141 keys and could lead to a DoS attack or memory corruption for applications
6142 that receive DSA private keys from untrusted sources. This scenario is
6145 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6151 * Disable SRP fake user seed to address a server memory leak.
6153 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6155 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6156 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6157 was changed to ignore the "fake user" SRP seed, even if the seed
6160 Users should use SRP_VBASE_get1_by_user instead. Note that in
6161 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6162 also that even though configuring the SRP seed attempts to hide
6163 invalid usernames by continuing the handshake with fake
6164 credentials, this behaviour is not constant time and no strong
6165 guarantees are made that the handshake is indistinguishable from
6166 that of a valid user.
6171 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6173 In the BN_hex2bn function the number of hex digits is calculated using an
6174 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6175 large values of `i` this can result in `bn_expand` not allocating any
6176 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6177 field as NULL leading to a subsequent NULL ptr deref. For very large values
6178 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6179 In this case memory is allocated to the internal BIGNUM data field, but it
6180 is insufficiently sized leading to heap corruption. A similar issue exists
6181 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6182 is ever called by user applications with very large untrusted hex/dec data.
6183 This is anticipated to be a rare occurrence.
6185 All OpenSSL internal usage of these functions use data that is not expected
6186 to be untrusted, e.g. config file data or application command line
6187 arguments. If user developed applications generate config file data based
6188 on untrusted data then it is possible that this could also lead to security
6189 consequences. This is also anticipated to be rare.
6191 This issue was reported to OpenSSL by Guido Vranken.
6196 * Fix memory issues in `BIO_*printf` functions
6198 The internal `fmtstr` function used in processing a "%s" format string in
6199 the `BIO_*printf` functions could overflow while calculating the length of a
6200 string and cause an OOB read when printing very long strings.
6202 Additionally the internal `doapr_outch` function can attempt to write to an
6203 OOB memory location (at an offset from the NULL pointer) in the event of a
6204 memory allocation failure. In 1.0.2 and below this could be caused where
6205 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6206 could be in processing a very long "%s" format string. Memory leaks can
6209 The first issue may mask the second issue dependent on compiler behaviour.
6210 These problems could enable attacks where large amounts of untrusted data
6211 is passed to the `BIO_*printf` functions. If applications use these functions
6212 in this way then they could be vulnerable. OpenSSL itself uses these
6213 functions when printing out human-readable dumps of ASN.1 data. Therefore
6214 applications that print this data could be vulnerable if the data is from
6215 untrusted sources. OpenSSL command line applications could also be
6216 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6217 as command line arguments.
6219 Libssl is not considered directly vulnerable. Additionally certificates etc
6220 received via remote connections via libssl are also unlikely to be able to
6221 trigger these issues because of message size limits enforced within libssl.
6223 This issue was reported to OpenSSL Guido Vranken.
6228 * Side channel attack on modular exponentiation
6230 A side-channel attack was found which makes use of cache-bank conflicts on
6231 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6232 of RSA keys. The ability to exploit this issue is limited as it relies on
6233 an attacker who has control of code in a thread running on the same
6234 hyper-threaded core as the victim thread which is performing decryptions.
6236 This issue was reported to OpenSSL by Yuval Yarom, The University of
6237 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6238 Nadia Heninger, University of Pennsylvania with more information at
6239 <http://cachebleed.info>.
6244 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6245 if no keysize is specified with default_bits. This fixes an
6246 omission in an earlier change that changed all RSA/DSA key generation
6247 commands to use 2048 bits by default.
6251 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6253 * DH small subgroups
6255 Historically OpenSSL only ever generated DH parameters based on "safe"
6256 primes. More recently (in version 1.0.2) support was provided for
6257 generating X9.42 style parameter files such as those required for RFC 5114
6258 support. The primes used in such files may not be "safe". Where an
6259 application is using DH configured with parameters based on primes that are
6260 not "safe" then an attacker could use this fact to find a peer's private
6261 DH exponent. This attack requires that the attacker complete multiple
6262 handshakes in which the peer uses the same private DH exponent. For example
6263 this could be used to discover a TLS server's private DH exponent if it's
6264 reusing the private DH exponent or it's using a static DH ciphersuite.
6266 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6267 TLS. It is not on by default. If the option is not set then the server
6268 reuses the same private DH exponent for the life of the server process and
6269 would be vulnerable to this attack. It is believed that many popular
6270 applications do set this option and would therefore not be at risk.
6272 The fix for this issue adds an additional check where a "q" parameter is
6273 available (as is the case in X9.42 based parameters). This detects the
6274 only known attack, and is the only possible defense for static DH
6275 ciphersuites. This could have some performance impact.
6277 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6278 default and cannot be disabled. This could have some performance impact.
6280 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6285 * SSLv2 doesn't block disabled ciphers
6287 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6288 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6289 been disabled, provided that the SSLv2 protocol was not also disabled via
6292 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6293 and Sebastian Schinzel.
6298 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6300 * BN_mod_exp may produce incorrect results on x86_64
6302 There is a carry propagating bug in the x86_64 Montgomery squaring
6303 procedure. No EC algorithms are affected. Analysis suggests that attacks
6304 against RSA and DSA as a result of this defect would be very difficult to
6305 perform and are not believed likely. Attacks against DH are considered just
6306 feasible (although very difficult) because most of the work necessary to
6307 deduce information about a private key may be performed offline. The amount
6308 of resources required for such an attack would be very significant and
6309 likely only accessible to a limited number of attackers. An attacker would
6310 additionally need online access to an unpatched system using the target
6311 private key in a scenario with persistent DH parameters and a private
6312 key that is shared between multiple clients. For example this can occur by
6313 default in OpenSSL DHE based SSL/TLS ciphersuites.
6315 This issue was reported to OpenSSL by Hanno Böck.
6320 * Certificate verify crash with missing PSS parameter
6322 The signature verification routines will crash with a NULL pointer
6323 dereference if presented with an ASN.1 signature using the RSA PSS
6324 algorithm and absent mask generation function parameter. Since these
6325 routines are used to verify certificate signature algorithms this can be
6326 used to crash any certificate verification operation and exploited in a
6327 DoS attack. Any application which performs certificate verification is
6328 vulnerable including OpenSSL clients and servers which enable client
6331 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6336 * X509_ATTRIBUTE memory leak
6338 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6339 memory. This structure is used by the PKCS#7 and CMS routines so any
6340 application which reads PKCS#7 or CMS data from untrusted sources is
6341 affected. SSL/TLS is not affected.
6343 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6349 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6350 This changes the decoding behaviour for some invalid messages,
6351 though the change is mostly in the more lenient direction, and
6352 legacy behaviour is preserved as much as possible.
6356 * In DSA_generate_parameters_ex, if the provided seed is too short,
6359 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6361 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6363 * Alternate chains certificate forgery
6365 During certificate verification, OpenSSL will attempt to find an
6366 alternative certificate chain if the first attempt to build such a chain
6367 fails. An error in the implementation of this logic can mean that an
6368 attacker could cause certain checks on untrusted certificates to be
6369 bypassed, such as the CA flag, enabling them to use a valid leaf
6370 certificate to act as a CA and "issue" an invalid certificate.
6372 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6377 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6379 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6380 incompatibility in the handling of HMAC. The previous ABI has now been
6385 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6387 * Malformed ECParameters causes infinite loop
6389 When processing an ECParameters structure OpenSSL enters an infinite loop
6390 if the curve specified is over a specially malformed binary polynomial
6393 This can be used to perform denial of service against any
6394 system which processes public keys, certificate requests or
6395 certificates. This includes TLS clients and TLS servers with
6396 client authentication enabled.
6398 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6403 * Exploitable out-of-bounds read in X509_cmp_time
6405 X509_cmp_time does not properly check the length of the ASN1_TIME
6406 string and can read a few bytes out of bounds. In addition,
6407 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6410 An attacker can use this to craft malformed certificates and CRLs of
6411 various sizes and potentially cause a segmentation fault, resulting in
6412 a DoS on applications that verify certificates or CRLs. TLS clients
6413 that verify CRLs are affected. TLS clients and servers with client
6414 authentication enabled may be affected if they use custom verification
6417 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6418 independently by Hanno Böck.
6423 * PKCS7 crash with missing EnvelopedContent
6425 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6426 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6427 with missing content and trigger a NULL pointer dereference on parsing.
6429 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6430 structures from untrusted sources are affected. OpenSSL clients and
6431 servers are not affected.
6433 This issue was reported to OpenSSL by Michal Zalewski (Google).
6438 * CMS verify infinite loop with unknown hash function
6440 When verifying a signedData message the CMS code can enter an infinite loop
6441 if presented with an unknown hash function OID. This can be used to perform
6442 denial of service against any system which verifies signedData messages using
6444 This issue was reported to OpenSSL by Johannes Bauer.
6449 * Race condition handling NewSessionTicket
6451 If a NewSessionTicket is received by a multi-threaded client when attempting to
6452 reuse a previous ticket then a race condition can occur potentially leading to
6453 a double free of the ticket data.
6458 * Only support 256-bit or stronger elliptic curves with the
6459 'ecdh_auto' setting (server) or by default (client). Of supported
6460 curves, prefer P-256 (both).
6464 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6466 * ClientHello sigalgs DoS fix
6468 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6469 invalid signature algorithms extension a NULL pointer dereference will
6470 occur. This can be exploited in a DoS attack against the server.
6472 This issue was was reported to OpenSSL by David Ramos of Stanford
6476 *Stephen Henson and Matt Caswell*
6478 * Multiblock corrupted pointer fix
6480 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6481 feature only applies on 64 bit x86 architecture platforms that support AES
6482 NI instructions. A defect in the implementation of "multiblock" can cause
6483 OpenSSL's internal write buffer to become incorrectly set to NULL when
6484 using non-blocking IO. Typically, when the user application is using a
6485 socket BIO for writing, this will only result in a failed connection.
6486 However if some other BIO is used then it is likely that a segmentation
6487 fault will be triggered, thus enabling a potential DoS attack.
6489 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6494 * Segmentation fault in DTLSv1_listen fix
6496 The DTLSv1_listen function is intended to be stateless and processes the
6497 initial ClientHello from many peers. It is common for user code to loop
6498 over the call to DTLSv1_listen until a valid ClientHello is received with
6499 an associated cookie. A defect in the implementation of DTLSv1_listen means
6500 that state is preserved in the SSL object from one invocation to the next
6501 that can lead to a segmentation fault. Errors processing the initial
6502 ClientHello can trigger this scenario. An example of such an error could be
6503 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6506 This issue was reported to OpenSSL by Per Allansson.
6511 * Segmentation fault in ASN1_TYPE_cmp fix
6513 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6514 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6515 certificate signature algorithm consistency this can be used to crash any
6516 certificate verification operation and exploited in a DoS attack. Any
6517 application which performs certificate verification is vulnerable including
6518 OpenSSL clients and servers which enable client authentication.
6523 * Segmentation fault for invalid PSS parameters fix
6525 The signature verification routines will crash with a NULL pointer
6526 dereference if presented with an ASN.1 signature using the RSA PSS
6527 algorithm and invalid parameters. Since these routines are used to verify
6528 certificate signature algorithms this can be used to crash any
6529 certificate verification operation and exploited in a DoS attack. Any
6530 application which performs certificate verification is vulnerable including
6531 OpenSSL clients and servers which enable client authentication.
6533 This issue was was reported to OpenSSL by Brian Carpenter.
6538 * ASN.1 structure reuse memory corruption fix
6540 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6541 memory corruption via an invalid write. Such reuse is and has been
6542 strongly discouraged and is believed to be rare.
6544 Applications that parse structures containing CHOICE or ANY DEFINED BY
6545 components may be affected. Certificate parsing (d2i_X509 and related
6546 functions) are however not affected. OpenSSL clients and servers are
6552 * PKCS7 NULL pointer dereferences fix
6554 The PKCS#7 parsing code does not handle missing outer ContentInfo
6555 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6556 missing content and trigger a NULL pointer dereference on parsing.
6558 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6559 otherwise parse PKCS#7 structures from untrusted sources are
6560 affected. OpenSSL clients and servers are not affected.
6562 This issue was reported to OpenSSL by Michal Zalewski (Google).
6567 * DoS via reachable assert in SSLv2 servers fix
6569 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6570 servers that both support SSLv2 and enable export cipher suites by sending
6571 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6573 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6574 (OpenSSL development team).
6579 * Empty CKE with client auth and DHE fix
6581 If client auth is used then a server can seg fault in the event of a DHE
6582 ciphersuite being selected and a zero length ClientKeyExchange message
6583 being sent by the client. This could be exploited in a DoS attack.
6588 * Handshake with unseeded PRNG fix
6590 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6591 with an unseeded PRNG. The conditions are:
6592 - The client is on a platform where the PRNG has not been seeded
6593 automatically, and the user has not seeded manually
6594 - A protocol specific client method version has been used (i.e. not
6595 SSL_client_methodv23)
6596 - A ciphersuite is used that does not require additional random data from
6597 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6599 If the handshake succeeds then the client random that has been used will
6600 have been generated from a PRNG with insufficient entropy and therefore the
6601 output may be predictable.
6603 For example using the following command with an unseeded openssl will
6604 succeed on an unpatched platform:
6606 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6611 * Use After Free following d2i_ECPrivatekey error fix
6613 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6614 could cause a use after free condition. This, in turn, could cause a double
6615 free in several private key parsing functions (such as d2i_PrivateKey
6616 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6617 for applications that receive EC private keys from untrusted
6618 sources. This scenario is considered rare.
6620 This issue was discovered by the BoringSSL project and fixed in their
6626 * X509_to_X509_REQ NULL pointer deref fix
6628 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6629 the certificate key is invalid. This function is rarely used in practice.
6631 This issue was discovered by Brian Carpenter.
6636 * Removed the export ciphers from the DEFAULT ciphers
6640 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6642 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6643 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6644 So far those who have to target multiple platforms would compromise
6645 and argue that binary targeting say ARMv5 would still execute on
6646 ARMv8. "Universal" build resolves this compromise by providing
6647 near-optimal performance even on newer platforms.
6651 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6652 (other platforms pending).
6654 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6656 * Add support for the SignedCertificateTimestampList certificate and
6657 OCSP response extensions from RFC6962.
6661 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6662 for corner cases. (Certain input points at infinity could lead to
6663 bogus results, with non-infinity inputs mapped to infinity too.)
6667 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6668 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6669 common cases are optimized and there still is room for further
6670 improvements. Vector Permutation AES for Altivec is also added.
6674 * Add support for little-endian ppc64 Linux target.
6676 *Marcelo Cerri (IBM)*
6678 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6679 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6680 are optimized and there still is room for further improvements.
6681 Both 32- and 64-bit modes are supported.
6683 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6685 * Improved ARMv7 NEON support.
6689 * Support for SPARC Architecture 2011 crypto extensions, first
6690 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6691 SHA256/512, MD5, GHASH and modular exponentiation.
6693 *Andy Polyakov, David Miller*
6695 * Accelerated modular exponentiation for Intel processors, a.k.a.
6698 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6700 * Support for new and upcoming Intel processors, including AVX2,
6701 BMI and SHA ISA extensions. This includes additional "stitched"
6702 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6705 This work was sponsored by Intel Corp.
6709 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6710 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6711 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6715 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6716 this fixes a limitation in previous versions of OpenSSL.
6720 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6721 MGF1 digest and OAEP label.
6725 * Add EVP support for key wrapping algorithms, to avoid problems with
6726 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6727 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6728 algorithms and include tests cases.
6732 * Add functions to allocate and set the fields of an ECDSA_METHOD
6735 *Douglas E. Engert, Steve Henson*
6737 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6738 difference in days and seconds between two tm or ASN1_TIME structures.
6742 * Add -rev test option to s_server to just reverse order of characters
6743 received by client and send back to server. Also prints an abbreviated
6744 summary of the connection parameters.
6748 * New option -brief for s_client and s_server to print out a brief summary
6749 of connection parameters.
6753 * Add callbacks for arbitrary TLS extensions.
6755 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6757 * New option -crl_download in several openssl utilities to download CRLs
6758 from CRLDP extension in certificates.
6762 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6766 * New function X509_CRL_diff to generate a delta CRL from the difference
6767 of two full CRLs. Add support to "crl" utility.
6771 * New functions to set lookup_crls function and to retrieve
6772 X509_STORE from X509_STORE_CTX.
6776 * Print out deprecated issuer and subject unique ID fields in
6781 * Extend OCSP I/O functions so they can be used for simple general purpose
6782 HTTP as well as OCSP. New wrapper function which can be used to download
6783 CRLs using the OCSP API.
6787 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6791 * `SSL_CONF*` functions. These provide a common framework for application
6792 configuration using configuration files or command lines.
6796 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6797 message callback and prints the results. Needs compile time option
6798 "enable-ssl-trace". New options to s_client and s_server to enable
6803 * New ctrl and macro to retrieve supported points extensions.
6804 Print out extension in s_server and s_client.
6808 * New functions to retrieve certificate signature and signature
6813 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6818 * New Suite B modes for TLS code. These use and enforce the requirements
6819 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6820 only use Suite B curves. The Suite B modes can be set by using the
6821 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6825 * New chain verification flags for Suite B levels of security. Check
6826 algorithms are acceptable when flags are set in X509_verify_cert.
6830 * Make tls1_check_chain return a set of flags indicating checks passed
6831 by a certificate chain. Add additional tests to handle client
6832 certificates: checks for matching certificate type and issuer name
6837 * If an attempt is made to use a signature algorithm not in the peer
6838 preference list abort the handshake. If client has no suitable
6839 signature algorithms in response to a certificate request do not
6840 use the certificate.
6844 * If server EC tmp key is not in client preference list abort handshake.
6848 * Add support for certificate stores in CERT structure. This makes it
6849 possible to have different stores per SSL structure or one store in
6850 the parent SSL_CTX. Include distinct stores for certificate chain
6851 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6852 to build and store a certificate chain in CERT structure: returning
6853 an error if the chain cannot be built: this will allow applications
6854 to test if a chain is correctly configured.
6856 Note: if the CERT based stores are not set then the parent SSL_CTX
6857 store is used to retain compatibility with existing behaviour.
6861 * New function ssl_set_client_disabled to set a ciphersuite disabled
6862 mask based on the current session, check mask when sending client
6863 hello and checking the requested ciphersuite.
6867 * New ctrls to retrieve and set certificate types in a certificate
6868 request message. Print out received values in s_client. If certificate
6869 types is not set with custom values set sensible values based on
6870 supported signature algorithms.
6874 * Support for distinct client and server supported signature algorithms.
6878 * Add certificate callback. If set this is called whenever a certificate
6879 is required by client or server. An application can decide which
6880 certificate chain to present based on arbitrary criteria: for example
6881 supported signature algorithms. Add very simple example to s_server.
6882 This fixes many of the problems and restrictions of the existing client
6883 certificate callback: for example you can now clear an existing
6884 certificate and specify the whole chain.
6888 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6889 the certificate can be used for (if anything). Set valid_flags field
6890 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6891 to have similar checks in it.
6893 Add new "cert_flags" field to CERT structure and include a "strict mode".
6894 This enforces some TLS certificate requirements (such as only permitting
6895 certificate signature algorithms contained in the supported algorithms
6896 extension) which some implementations ignore: this option should be used
6897 with caution as it could cause interoperability issues.
6901 * Update and tidy signature algorithm extension processing. Work out
6902 shared signature algorithms based on preferences and peer algorithms
6903 and print them out in s_client and s_server. Abort handshake if no
6904 shared signature algorithms.
6908 * Add new functions to allow customised supported signature algorithms
6909 for SSL and SSL_CTX structures. Add options to s_client and s_server
6914 * New function SSL_certs_clear() to delete all references to certificates
6915 from an SSL structure. Before this once a certificate had been added
6916 it couldn't be removed.
6920 * Integrate hostname, email address and IP address checking with certificate
6921 verification. New verify options supporting checking in openssl utility.
6925 * Fixes and wildcard matching support to hostname and email checking
6926 functions. Add manual page.
6928 *Florian Weimer (Red Hat Product Security Team)*
6930 * New functions to check a hostname email or IP address against a
6931 certificate. Add options x509 utility to print results of checks against
6936 * Fix OCSP checking.
6938 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6940 * Initial experimental support for explicitly trusted non-root CAs.
6941 OpenSSL still tries to build a complete chain to a root but if an
6942 intermediate CA has a trust setting included that is used. The first
6943 setting is used: whether to trust (e.g., -addtrust option to the x509
6948 * Add -trusted_first option which attempts to find certificates in the
6949 trusted store even if an untrusted chain is also supplied.
6953 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6954 platform support for Linux and Android.
6958 * Support for linux-x32, ILP32 environment in x86_64 framework.
6962 * Experimental multi-implementation support for FIPS capable OpenSSL.
6963 When in FIPS mode the approved implementations are used as normal,
6964 when not in FIPS mode the internal unapproved versions are used instead.
6965 This means that the FIPS capable OpenSSL isn't forced to use the
6966 (often lower performance) FIPS implementations outside FIPS mode.
6970 * Transparently support X9.42 DH parameters when calling
6971 PEM_read_bio_DHparameters. This means existing applications can handle
6972 the new parameter format automatically.
6976 * Initial experimental support for X9.42 DH parameter format: mainly
6977 to support use of 'q' parameter for RFC5114 parameters.
6981 * Add DH parameters from RFC5114 including test data to dhtest.
6985 * Support for automatic EC temporary key parameter selection. If enabled
6986 the most preferred EC parameters are automatically used instead of
6987 hardcoded fixed parameters. Now a server just has to call:
6988 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6989 support ECDH and use the most appropriate parameters.
6993 * Enhance and tidy EC curve and point format TLS extension code. Use
6994 static structures instead of allocation if default values are used.
6995 New ctrls to set curves we wish to support and to retrieve shared curves.
6996 Print out shared curves in s_server. New options to s_server and s_client
6997 to set list of supported curves.
7001 * New ctrls to retrieve supported signature algorithms and
7002 supported curve values as an array of NIDs. Extend openssl utility
7003 to print out received values.
7007 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
7008 between NIDs and the more common NIST names such as "P-256". Enhance
7009 ecparam utility and ECC method to recognise the NIST names for curves.
7013 * Enhance SSL/TLS certificate chain handling to support different
7014 chains for each certificate instead of one chain in the parent SSL_CTX.
7018 * Support for fixed DH ciphersuite client authentication: where both
7019 server and client use DH certificates with common parameters.
7023 * Support for fixed DH ciphersuites: those requiring DH server
7028 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
7030 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
7031 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
7032 X509_CINF_get_signature were reverted post internal team review.
7037 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
7039 * OCSP Status Request extension unbounded memory growth
7041 A malicious client can send an excessively large OCSP Status Request
7042 extension. If that client continually requests renegotiation, sending a
7043 large OCSP Status Request extension each time, then there will be unbounded
7044 memory growth on the server. This will eventually lead to a Denial Of
7045 Service attack through memory exhaustion. Servers with a default
7046 configuration are vulnerable even if they do not support OCSP. Builds using
7047 the "no-ocsp" build time option are not affected.
7049 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7054 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
7057 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
7063 * OOB write in MDC2_Update()
7065 An overflow can occur in MDC2_Update() either if called directly or
7066 through the EVP_DigestUpdate() function using MDC2. If an attacker
7067 is able to supply very large amounts of input data after a previous
7068 call to EVP_EncryptUpdate() with a partial block then a length check
7069 can overflow resulting in a heap corruption.
7071 The amount of data needed is comparable to SIZE_MAX which is impractical
7074 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7079 * Malformed SHA512 ticket DoS
7081 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
7082 DoS attack where a malformed ticket will result in an OOB read which will
7085 The use of SHA512 in TLS session tickets is comparatively rare as it requires
7086 a custom server callback and ticket lookup mechanism.
7088 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7093 * OOB write in BN_bn2dec()
7095 The function BN_bn2dec() does not check the return value of BN_div_word().
7096 This can cause an OOB write if an application uses this function with an
7097 overly large BIGNUM. This could be a problem if an overly large certificate
7098 or CRL is printed out from an untrusted source. TLS is not affected because
7099 record limits will reject an oversized certificate before it is parsed.
7101 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7106 * OOB read in TS_OBJ_print_bio()
7108 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
7109 the total length the OID text representation would use and not the amount
7110 of data written. This will result in OOB reads when large OIDs are
7113 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7118 * Pointer arithmetic undefined behaviour
7120 Avoid some undefined pointer arithmetic
7122 A common idiom in the codebase is to check limits in the following manner:
7125 Where "p" points to some malloc'd data of SIZE bytes and
7128 "len" here could be from some externally supplied data (e.g. from a TLS
7131 The rules of C pointer arithmetic are such that "p + len" is only well
7132 defined where len <= SIZE. Therefore, the above idiom is actually
7133 undefined behaviour.
7135 For example this could cause problems if some malloc implementation
7136 provides an address for "p" such that "p + len" actually overflows for
7137 values of len that are too big and therefore p + len < limit.
7139 This issue was reported to OpenSSL by Guido Vranken
7144 * Constant time flag not preserved in DSA signing
7146 Operations in the DSA signing algorithm should run in constant time in
7147 order to avoid side channel attacks. A flaw in the OpenSSL DSA
7148 implementation means that a non-constant time codepath is followed for
7149 certain operations. This has been demonstrated through a cache-timing
7150 attack to be sufficient for an attacker to recover the private DSA key.
7152 This issue was reported by César Pereida (Aalto University), Billy Brumley
7153 (Tampere University of Technology), and Yuval Yarom (The University of
7154 Adelaide and NICTA).
7159 * DTLS buffered message DoS
7161 In a DTLS connection where handshake messages are delivered out-of-order
7162 those messages that OpenSSL is not yet ready to process will be buffered
7163 for later use. Under certain circumstances, a flaw in the logic means that
7164 those messages do not get removed from the buffer even though the handshake
7165 has been completed. An attacker could force up to approx. 15 messages to
7166 remain in the buffer when they are no longer required. These messages will
7167 be cleared when the DTLS connection is closed. The default maximum size for
7168 a message is 100k. Therefore, the attacker could force an additional 1500k
7169 to be consumed per connection. By opening many simultaneous connections an
7170 attacker could cause a DoS attack through memory exhaustion.
7172 This issue was reported to OpenSSL by Quan Luo.
7177 * DTLS replay protection DoS
7179 A flaw in the DTLS replay attack protection mechanism means that records
7180 that arrive for future epochs update the replay protection "window" before
7181 the MAC for the record has been validated. This could be exploited by an
7182 attacker by sending a record for the next epoch (which does not have to
7183 decrypt or have a valid MAC), with a very large sequence number. This means
7184 that all subsequent legitimate packets are dropped causing a denial of
7185 service for a specific DTLS connection.
7187 This issue was reported to OpenSSL by the OCAP audit team.
7192 * Certificate message OOB reads
7194 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7195 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7196 theoretical DoS risk but this has not been observed in practice on common
7199 The messages affected are client certificate, client certificate request
7200 and server certificate. As a result the attack can only be performed
7201 against a client or a server which enables client authentication.
7203 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7208 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7210 * Prevent padding oracle in AES-NI CBC MAC check
7212 A MITM attacker can use a padding oracle attack to decrypt traffic
7213 when the connection uses an AES CBC cipher and the server support
7216 This issue was introduced as part of the fix for Lucky 13 padding
7217 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7218 constant time by making sure that always the same bytes are read and
7219 compared against either the MAC or padding bytes. But it no longer
7220 checked that there was enough data to have both the MAC and padding
7223 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7228 * Fix EVP_EncodeUpdate overflow
7230 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7231 Base64 encoding of binary data. If an attacker is able to supply very large
7232 amounts of input data then a length check can overflow resulting in a heap
7235 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7236 the `PEM_write_bio*` family of functions. These are mainly used within the
7237 OpenSSL command line applications, so any application which processes data
7238 from an untrusted source and outputs it as a PEM file should be considered
7239 vulnerable to this issue. User applications that call these APIs directly
7240 with large amounts of untrusted data may also be vulnerable.
7242 This issue was reported by Guido Vranken.
7247 * Fix EVP_EncryptUpdate overflow
7249 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7250 is able to supply very large amounts of input data after a previous call to
7251 EVP_EncryptUpdate() with a partial block then a length check can overflow
7252 resulting in a heap corruption. Following an analysis of all OpenSSL
7253 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7254 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7255 the first called function after an EVP_EncryptInit(), and therefore that
7256 specific call must be safe. The second form is where the length passed to
7257 EVP_EncryptUpdate() can be seen from the code to be some small value and
7258 therefore there is no possibility of an overflow. Since all instances are
7259 one of these two forms, it is believed that there can be no overflows in
7260 internal code due to this problem. It should be noted that
7261 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7262 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7263 of these calls have also been analysed too and it is believed there are no
7264 instances in internal usage where an overflow could occur.
7266 This issue was reported by Guido Vranken.
7271 * Prevent ASN.1 BIO excessive memory allocation
7273 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7274 a short invalid encoding can casuse allocation of large amounts of memory
7275 potentially consuming excessive resources or exhausting memory.
7277 Any application parsing untrusted data through d2i BIO functions is
7278 affected. The memory based functions such as d2i_X509() are *not* affected.
7279 Since the memory based functions are used by the TLS library, TLS
7280 applications are not affected.
7282 This issue was reported by Brian Carpenter.
7289 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7290 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7291 in arbitrary stack data being returned in the buffer.
7293 This issue was reported by Guido Vranken.
7298 * Modify behavior of ALPN to invoke callback after SNI/servername
7299 callback, such that updates to the SSL_CTX affect ALPN.
7303 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7308 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7309 methods are enabled and ssl2 is disabled the methods return NULL.
7313 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7315 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7316 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7317 provide any "EXPORT" or "LOW" strength ciphers.
7321 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7322 is by default disabled at build-time. Builds that are not configured with
7323 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7324 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7325 will need to explicitly call either of:
7327 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7329 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7331 as appropriate. Even if either of those is used, or the application
7332 explicitly uses the version-specific SSLv2_method() or its client and
7333 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7334 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7335 ciphers, and SSLv2 56-bit DES are no longer available.
7340 * Fix a double-free in DSA code
7342 A double free bug was discovered when OpenSSL parses malformed DSA private
7343 keys and could lead to a DoS attack or memory corruption for applications
7344 that receive DSA private keys from untrusted sources. This scenario is
7347 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7353 * Disable SRP fake user seed to address a server memory leak.
7355 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7357 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7358 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7359 was changed to ignore the "fake user" SRP seed, even if the seed
7362 Users should use SRP_VBASE_get1_by_user instead. Note that in
7363 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7364 also that even though configuring the SRP seed attempts to hide
7365 invalid usernames by continuing the handshake with fake
7366 credentials, this behaviour is not constant time and no strong
7367 guarantees are made that the handshake is indistinguishable from
7368 that of a valid user.
7373 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7375 In the BN_hex2bn function the number of hex digits is calculated using an
7376 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7377 large values of `i` this can result in `bn_expand` not allocating any
7378 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7379 field as NULL leading to a subsequent NULL ptr deref. For very large values
7380 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7381 In this case memory is allocated to the internal BIGNUM data field, but it
7382 is insufficiently sized leading to heap corruption. A similar issue exists
7383 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7384 is ever called by user applications with very large untrusted hex/dec data.
7385 This is anticipated to be a rare occurrence.
7387 All OpenSSL internal usage of these functions use data that is not expected
7388 to be untrusted, e.g. config file data or application command line
7389 arguments. If user developed applications generate config file data based
7390 on untrusted data then it is possible that this could also lead to security
7391 consequences. This is also anticipated to be rare.
7393 This issue was reported to OpenSSL by Guido Vranken.
7398 * Fix memory issues in `BIO_*printf` functions
7400 The internal `fmtstr` function used in processing a "%s" format string in
7401 the `BIO_*printf` functions could overflow while calculating the length of a
7402 string and cause an OOB read when printing very long strings.
7404 Additionally the internal `doapr_outch` function can attempt to write to an
7405 OOB memory location (at an offset from the NULL pointer) in the event of a
7406 memory allocation failure. In 1.0.2 and below this could be caused where
7407 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7408 could be in processing a very long "%s" format string. Memory leaks can
7411 The first issue may mask the second issue dependent on compiler behaviour.
7412 These problems could enable attacks where large amounts of untrusted data
7413 is passed to the `BIO_*printf` functions. If applications use these functions
7414 in this way then they could be vulnerable. OpenSSL itself uses these
7415 functions when printing out human-readable dumps of ASN.1 data. Therefore
7416 applications that print this data could be vulnerable if the data is from
7417 untrusted sources. OpenSSL command line applications could also be
7418 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7419 as command line arguments.
7421 Libssl is not considered directly vulnerable. Additionally certificates etc
7422 received via remote connections via libssl are also unlikely to be able to
7423 trigger these issues because of message size limits enforced within libssl.
7425 This issue was reported to OpenSSL Guido Vranken.
7430 * Side channel attack on modular exponentiation
7432 A side-channel attack was found which makes use of cache-bank conflicts on
7433 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7434 of RSA keys. The ability to exploit this issue is limited as it relies on
7435 an attacker who has control of code in a thread running on the same
7436 hyper-threaded core as the victim thread which is performing decryptions.
7438 This issue was reported to OpenSSL by Yuval Yarom, The University of
7439 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7440 Nadia Heninger, University of Pennsylvania with more information at
7441 <http://cachebleed.info>.
7446 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7447 if no keysize is specified with default_bits. This fixes an
7448 omission in an earlier change that changed all RSA/DSA key generation
7449 commands to use 2048 bits by default.
7453 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7455 * Protection for DH small subgroup attacks
7457 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7458 switched on by default and cannot be disabled. This could have some
7463 * SSLv2 doesn't block disabled ciphers
7465 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7466 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7467 been disabled, provided that the SSLv2 protocol was not also disabled via
7470 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7471 and Sebastian Schinzel.
7476 * Reject DH handshakes with parameters shorter than 1024 bits.
7480 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7482 * Certificate verify crash with missing PSS parameter
7484 The signature verification routines will crash with a NULL pointer
7485 dereference if presented with an ASN.1 signature using the RSA PSS
7486 algorithm and absent mask generation function parameter. Since these
7487 routines are used to verify certificate signature algorithms this can be
7488 used to crash any certificate verification operation and exploited in a
7489 DoS attack. Any application which performs certificate verification is
7490 vulnerable including OpenSSL clients and servers which enable client
7493 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7498 * X509_ATTRIBUTE memory leak
7500 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7501 memory. This structure is used by the PKCS#7 and CMS routines so any
7502 application which reads PKCS#7 or CMS data from untrusted sources is
7503 affected. SSL/TLS is not affected.
7505 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7511 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7512 This changes the decoding behaviour for some invalid messages,
7513 though the change is mostly in the more lenient direction, and
7514 legacy behaviour is preserved as much as possible.
7518 * In DSA_generate_parameters_ex, if the provided seed is too short,
7519 use a random seed, as already documented.
7521 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7523 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7525 * Alternate chains certificate forgery
7527 During certificate verification, OpenSSL will attempt to find an
7528 alternative certificate chain if the first attempt to build such a chain
7529 fails. An error in the implementation of this logic can mean that an
7530 attacker could cause certain checks on untrusted certificates to be
7531 bypassed, such as the CA flag, enabling them to use a valid leaf
7532 certificate to act as a CA and "issue" an invalid certificate.
7534 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7540 * Race condition handling PSK identify hint
7542 If PSK identity hints are received by a multi-threaded client then
7543 the values are wrongly updated in the parent SSL_CTX structure. This can
7544 result in a race condition potentially leading to a double free of the
7550 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7552 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7553 incompatibility in the handling of HMAC. The previous ABI has now been
7556 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7558 * Malformed ECParameters causes infinite loop
7560 When processing an ECParameters structure OpenSSL enters an infinite loop
7561 if the curve specified is over a specially malformed binary polynomial
7564 This can be used to perform denial of service against any
7565 system which processes public keys, certificate requests or
7566 certificates. This includes TLS clients and TLS servers with
7567 client authentication enabled.
7569 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7574 * Exploitable out-of-bounds read in X509_cmp_time
7576 X509_cmp_time does not properly check the length of the ASN1_TIME
7577 string and can read a few bytes out of bounds. In addition,
7578 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7581 An attacker can use this to craft malformed certificates and CRLs of
7582 various sizes and potentially cause a segmentation fault, resulting in
7583 a DoS on applications that verify certificates or CRLs. TLS clients
7584 that verify CRLs are affected. TLS clients and servers with client
7585 authentication enabled may be affected if they use custom verification
7588 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7589 independently by Hanno Böck.
7594 * PKCS7 crash with missing EnvelopedContent
7596 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7597 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7598 with missing content and trigger a NULL pointer dereference on parsing.
7600 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7601 structures from untrusted sources are affected. OpenSSL clients and
7602 servers are not affected.
7604 This issue was reported to OpenSSL by Michal Zalewski (Google).
7609 * CMS verify infinite loop with unknown hash function
7611 When verifying a signedData message the CMS code can enter an infinite loop
7612 if presented with an unknown hash function OID. This can be used to perform
7613 denial of service against any system which verifies signedData messages using
7615 This issue was reported to OpenSSL by Johannes Bauer.
7620 * Race condition handling NewSessionTicket
7622 If a NewSessionTicket is received by a multi-threaded client when attempting to
7623 reuse a previous ticket then a race condition can occur potentially leading to
7624 a double free of the ticket data.
7629 * Reject DH handshakes with parameters shorter than 768 bits.
7631 *Kurt Roeckx and Emilia Kasper*
7633 * dhparam: generate 2048-bit parameters by default.
7635 *Kurt Roeckx and Emilia Kasper*
7637 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7639 * Segmentation fault in ASN1_TYPE_cmp fix
7641 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7642 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7643 certificate signature algorithm consistency this can be used to crash any
7644 certificate verification operation and exploited in a DoS attack. Any
7645 application which performs certificate verification is vulnerable including
7646 OpenSSL clients and servers which enable client authentication.
7651 * ASN.1 structure reuse memory corruption fix
7653 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7654 memory corruption via an invalid write. Such reuse is and has been
7655 strongly discouraged and is believed to be rare.
7657 Applications that parse structures containing CHOICE or ANY DEFINED BY
7658 components may be affected. Certificate parsing (d2i_X509 and related
7659 functions) are however not affected. OpenSSL clients and servers are
7665 * PKCS7 NULL pointer dereferences fix
7667 The PKCS#7 parsing code does not handle missing outer ContentInfo
7668 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7669 missing content and trigger a NULL pointer dereference on parsing.
7671 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7672 otherwise parse PKCS#7 structures from untrusted sources are
7673 affected. OpenSSL clients and servers are not affected.
7675 This issue was reported to OpenSSL by Michal Zalewski (Google).
7680 * DoS via reachable assert in SSLv2 servers fix
7682 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7683 servers that both support SSLv2 and enable export cipher suites by sending
7684 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7686 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7687 (OpenSSL development team).
7692 * Use After Free following d2i_ECPrivatekey error fix
7694 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7695 could cause a use after free condition. This, in turn, could cause a double
7696 free in several private key parsing functions (such as d2i_PrivateKey
7697 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7698 for applications that receive EC private keys from untrusted
7699 sources. This scenario is considered rare.
7701 This issue was discovered by the BoringSSL project and fixed in their
7707 * X509_to_X509_REQ NULL pointer deref fix
7709 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7710 the certificate key is invalid. This function is rarely used in practice.
7712 This issue was discovered by Brian Carpenter.
7717 * Removed the export ciphers from the DEFAULT ciphers
7721 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7723 * Build fixes for the Windows and OpenVMS platforms
7725 *Matt Caswell and Richard Levitte*
7727 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7729 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7730 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7731 dereference. This could lead to a Denial Of Service attack. Thanks to
7732 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7737 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7738 dtls1_buffer_record function under certain conditions. In particular this
7739 could occur if an attacker sent repeated DTLS records with the same
7740 sequence number but for the next epoch. The memory leak could be exploited
7741 by an attacker in a Denial of Service attack through memory exhaustion.
7742 Thanks to Chris Mueller for reporting this issue.
7747 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7748 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7749 method would be set to NULL which could later result in a NULL pointer
7750 dereference. Thanks to Frank Schmirler for reporting this issue.
7755 * Abort handshake if server key exchange message is omitted for ephemeral
7758 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7759 reporting this issue.
7764 * Remove non-export ephemeral RSA code on client and server. This code
7765 violated the TLS standard by allowing the use of temporary RSA keys in
7766 non-export ciphersuites and could be used by a server to effectively
7767 downgrade the RSA key length used to a value smaller than the server
7768 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7769 INRIA or reporting this issue.
7774 * Fixed issue where DH client certificates are accepted without verification.
7775 An OpenSSL server will accept a DH certificate for client authentication
7776 without the certificate verify message. This effectively allows a client to
7777 authenticate without the use of a private key. This only affects servers
7778 which trust a client certificate authority which issues certificates
7779 containing DH keys: these are extremely rare and hardly ever encountered.
7780 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7786 * Ensure that the session ID context of an SSL is updated when its
7787 SSL_CTX is updated via SSL_set_SSL_CTX.
7789 The session ID context is typically set from the parent SSL_CTX,
7790 and can vary with the CTX.
7794 * Fix various certificate fingerprint issues.
7796 By using non-DER or invalid encodings outside the signed portion of a
7797 certificate the fingerprint can be changed without breaking the signature.
7798 Although no details of the signed portion of the certificate can be changed
7799 this can cause problems with some applications: e.g. those using the
7800 certificate fingerprint for blacklists.
7802 1. Reject signatures with non zero unused bits.
7804 If the BIT STRING containing the signature has non zero unused bits reject
7805 the signature. All current signature algorithms require zero unused bits.
7807 2. Check certificate algorithm consistency.
7809 Check the AlgorithmIdentifier inside TBS matches the one in the
7810 certificate signature. NB: this will result in signature failure
7811 errors for some broken certificates.
7813 Thanks to Konrad Kraszewski from Google for reporting this issue.
7815 3. Check DSA/ECDSA signatures use DER.
7817 Re-encode DSA/ECDSA signatures and compare with the original received
7818 signature. Return an error if there is a mismatch.
7820 This will reject various cases including garbage after signature
7821 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7822 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7823 (negative or with leading zeroes).
7825 Further analysis was conducted and fixes were developed by Stephen Henson
7826 of the OpenSSL core team.
7832 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7833 results on some platforms, including x86_64. This bug occurs at random
7834 with a very low probability, and is not known to be exploitable in any
7835 way, though its exact impact is difficult to determine. Thanks to Pieter
7836 Wuille (Blockstream) who reported this issue and also suggested an initial
7837 fix. Further analysis was conducted by the OpenSSL development team and
7838 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7839 the OpenSSL core team.
7844 * Do not resume sessions on the server if the negotiated protocol
7845 version does not match the session's version. Resuming with a different
7846 version, while not strictly forbidden by the RFC, is of questionable
7847 sanity and breaks all known clients.
7849 *David Benjamin, Emilia Käsper*
7851 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7852 early CCS messages during renegotiation. (Note that because
7853 renegotiation is encrypted, this early CCS was not exploitable.)
7857 * Tighten client-side session ticket handling during renegotiation:
7858 ensure that the client only accepts a session ticket if the server sends
7859 the extension anew in the ServerHello. Previously, a TLS client would
7860 reuse the old extension state and thus accept a session ticket if one was
7861 announced in the initial ServerHello.
7863 Similarly, ensure that the client requires a session ticket if one
7864 was advertised in the ServerHello. Previously, a TLS client would
7865 ignore a missing NewSessionTicket message.
7869 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7873 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7874 sends a carefully crafted handshake message, to cause OpenSSL to fail
7875 to free up to 64k of memory causing a memory leak. This could be
7876 exploited in a Denial Of Service attack. This issue affects OpenSSL
7877 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7878 whether SRTP is used or configured. Implementations of OpenSSL that
7879 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7881 The fix was developed by the OpenSSL team.
7886 * Session Ticket Memory Leak.
7888 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7889 integrity of that ticket is first verified. In the event of a session
7890 ticket integrity check failing, OpenSSL will fail to free memory
7891 causing a memory leak. By sending a large number of invalid session
7892 tickets an attacker could exploit this issue in a Denial Of Service
7898 * Build option no-ssl3 is incomplete.
7900 When OpenSSL is configured with "no-ssl3" as a build option, servers
7901 could accept and complete an SSL 3.0 handshake, and clients could be
7902 configured to send them.
7905 *Akamai and the OpenSSL team*
7907 * Add support for TLS_FALLBACK_SCSV.
7908 Client applications doing fallback retries should call
7909 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7912 *Adam Langley, Bodo Moeller*
7914 * Add additional DigestInfo checks.
7916 Re-encode DigestInto in DER and check against the original when
7917 verifying RSA signature: this will reject any improperly encoded
7918 DigestInfo structures.
7920 Note: this is a precautionary measure and no attacks are currently known.
7924 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7926 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7927 SRP code can be overrun an internal buffer. Add sanity check that
7928 g, A, B < N to SRP code.
7930 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7931 Group for discovering this issue.
7936 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7937 TLS 1.0 instead of higher protocol versions when the ClientHello message
7938 is badly fragmented. This allows a man-in-the-middle attacker to force a
7939 downgrade to TLS 1.0 even if both the server and the client support a
7940 higher protocol version, by modifying the client's TLS records.
7942 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7943 researching this issue.
7948 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7949 to a denial of service attack. A malicious server can crash the client
7950 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7951 ciphersuite and sending carefully crafted handshake messages.
7953 Thanks to Felix Gröbert (Google) for discovering and researching this
7959 * By sending carefully crafted DTLS packets an attacker could cause openssl
7960 to leak memory. This can be exploited through a Denial of Service attack.
7961 Thanks to Adam Langley for discovering and researching this issue.
7966 * An attacker can force openssl to consume large amounts of memory whilst
7967 processing DTLS handshake messages. This can be exploited through a
7968 Denial of Service attack.
7969 Thanks to Adam Langley for discovering and researching this issue.
7974 * An attacker can force an error condition which causes openssl to crash
7975 whilst processing DTLS packets due to memory being freed twice. This
7976 can be exploited through a Denial of Service attack.
7977 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7983 * If a multithreaded client connects to a malicious server using a resumed
7984 session and the server sends an ec point format extension it could write
7985 up to 255 bytes to freed memory.
7987 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7993 * A malicious server can crash an OpenSSL client with a null pointer
7994 dereference (read) by specifying an SRP ciphersuite even though it was not
7995 properly negotiated with the client. This can be exploited through a
7996 Denial of Service attack.
7998 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7999 discovering and researching this issue.
8004 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8005 X509_name_oneline, X509_name_print_ex et al. to leak some information
8006 from the stack. Applications may be affected if they echo pretty printing
8007 output to the attacker.
8009 Thanks to Ivan Fratric (Google) for discovering this issue.
8012 *Emilia Käsper, and Steve Henson*
8014 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8015 for corner cases. (Certain input points at infinity could lead to
8016 bogus results, with non-infinity inputs mapped to infinity too.)
8020 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
8022 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8023 handshake can force the use of weak keying material in OpenSSL
8024 SSL/TLS clients and servers.
8026 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8027 researching this issue. ([CVE-2014-0224])
8029 *KIKUCHI Masashi, Steve Henson*
8031 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8032 OpenSSL DTLS client the code can be made to recurse eventually crashing
8035 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8038 *Imre Rad, Steve Henson*
8040 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8041 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8042 client or server. This is potentially exploitable to run arbitrary
8043 code on a vulnerable client or server.
8045 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8047 *Jüri Aedla, Steve Henson*
8049 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8050 are subject to a denial of service attack.
8052 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8053 this issue. ([CVE-2014-3470])
8055 *Felix Gröbert, Ivan Fratric, Steve Henson*
8057 * Harmonize version and its documentation. -f flag is used to display
8060 *mancha <mancha1@zoho.com>*
8062 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8063 in i2d_ECPrivateKey.
8065 *mancha <mancha1@zoho.com>*
8067 * Fix some double frees. These are not thought to be exploitable.
8069 *mancha <mancha1@zoho.com>*
8071 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
8073 * A missing bounds check in the handling of the TLS heartbeat extension
8074 can be used to reveal up to 64k of memory to a connected client or
8077 Thanks for Neel Mehta of Google Security for discovering this bug and to
8078 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
8079 preparing the fix ([CVE-2014-0160])
8081 *Adam Langley, Bodo Moeller*
8083 * Fix for the attack described in the paper "Recovering OpenSSL
8084 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8085 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8086 <http://eprint.iacr.org/2014/140>
8088 Thanks to Yuval Yarom and Naomi Benger for discovering this
8089 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8091 *Yuval Yarom and Naomi Benger*
8093 * TLS pad extension: draft-agl-tls-padding-03
8095 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
8096 TLS client Hello record length value would otherwise be > 255 and
8097 less that 512 pad with a dummy extension containing zeroes so it
8098 is at least 512 bytes long.
8100 *Adam Langley, Steve Henson*
8102 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
8104 * Fix for TLS record tampering bug. A carefully crafted invalid
8105 handshake could crash OpenSSL with a NULL pointer exception.
8106 Thanks to Anton Johansson for reporting this issues.
8109 * Keep original DTLS digest and encryption contexts in retransmission
8110 structures so we can use the previous session parameters if they need
8111 to be resent. ([CVE-2013-6450])
8115 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8116 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8117 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8118 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8119 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8120 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8122 *Rob Stradling, Adam Langley*
8124 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
8126 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
8127 supporting platforms or when small records were transferred.
8129 *Andy Polyakov, Steve Henson*
8131 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
8133 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8135 This addresses the flaw in CBC record processing discovered by
8136 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8137 at: <http://www.isg.rhul.ac.uk/tls/>
8139 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8140 Security Group at Royal Holloway, University of London
8141 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8142 Emilia Käsper for the initial patch.
8145 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8147 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
8148 ciphersuites which can be exploited in a denial of service attack.
8149 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
8150 and detecting this bug and to Wolfgang Ettlinger
8151 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
8156 * Return an error when checking OCSP signatures when key is NULL.
8157 This fixes a DoS attack. ([CVE-2013-0166])
8161 * Make openssl verify return errors.
8163 *Chris Palmer <palmer@google.com> and Ben Laurie*
8165 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8166 the right response is stapled. Also change SSL_get_certificate()
8167 so it returns the certificate actually sent.
8168 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8170 *Rob Stradling <rob.stradling@comodo.com>*
8172 * Fix possible deadlock when decoding public keys.
8176 * Don't use TLS 1.0 record version number in initial client hello
8181 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
8183 * Sanity check record length before skipping explicit IV in TLS
8184 1.2, 1.1 and DTLS to fix DoS attack.
8186 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8187 fuzzing as a service testing platform.
8192 * Initialise tkeylen properly when encrypting CMS messages.
8193 Thanks to Solar Designer of Openwall for reporting this issue.
8197 * In FIPS mode don't try to use composite ciphers as they are not
8202 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8204 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8205 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8206 mean any application compiled against OpenSSL 1.0.0 headers setting
8207 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8208 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8209 0x10000000L Any application which was previously compiled against
8210 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8211 will need to be recompiled as a result. Letting be results in
8212 inability to disable specifically TLS 1.1 and in client context,
8213 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8217 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8218 disable just protocol X, but all protocols above X *if* there are
8219 protocols *below* X still enabled. In more practical terms it means
8220 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8221 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8222 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8227 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8229 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8230 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8231 in CRYPTO_realloc_clean.
8233 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8234 issue and to Adam Langley <agl@chromium.org> for fixing it.
8237 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8239 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8243 * Workarounds for some broken servers that "hang" if a client hello
8244 record length exceeds 255 bytes.
8246 1. Do not use record version number > TLS 1.0 in initial client
8247 hello: some (but not all) hanging servers will now work.
8248 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8249 the number of ciphers sent in the client hello. This should be
8250 set to an even number, such as 50, for example by passing:
8251 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8252 Most broken servers should now work.
8253 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8254 TLS 1.2 client support entirely.
8258 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8262 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8264 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8265 STRING form instead of a DigestInfo.
8269 * The format used for MDC2 RSA signatures is inconsistent between EVP
8270 and the RSA_sign/RSA_verify functions. This was made more apparent when
8271 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8272 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8273 the correct format in RSA_verify so both forms transparently work.
8277 * Some servers which support TLS 1.0 can choke if we initially indicate
8278 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8279 encrypted premaster secret. As a workaround use the maximum permitted
8280 client version in client hello, this should keep such servers happy
8281 and still work with previous versions of OpenSSL.
8285 * Add support for TLS/DTLS heartbeats.
8287 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8289 * Add support for SCTP.
8291 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8293 * Improved PRNG seeding for VOS.
8295 *Paul Green <Paul.Green@stratus.com>*
8297 * Extensive assembler packs updates, most notably:
8299 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8300 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8301 - x86_64: bit-sliced AES implementation;
8302 - ARM: NEON support, contemporary platforms optimizations;
8303 - s390x: z196 support;
8304 - `*`: GHASH and GF(2^m) multiplication implementations;
8308 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8309 (removal of unnecessary code)
8311 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8313 * Add TLS key material exporter from RFC 5705.
8317 * Add DTLS-SRTP negotiation from RFC 5764.
8321 * Add Next Protocol Negotiation,
8322 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8323 disabled with a no-npn flag to config or Configure. Code donated
8326 *Adam Langley <agl@google.com> and Ben Laurie*
8328 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8329 NIST-P256, NIST-P521, with constant-time single point multiplication on
8330 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8331 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8332 Code made available under Apache License version 2.0.
8334 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8335 line to include this in your build of OpenSSL, and run "make depend" (or
8336 "make update"). This enables the following EC_METHODs:
8338 EC_GFp_nistp224_method()
8339 EC_GFp_nistp256_method()
8340 EC_GFp_nistp521_method()
8342 EC_GROUP_new_by_curve_name() will automatically use these (while
8343 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8346 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8348 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8349 all platforms. Move ssize_t definition from e_os.h to the public
8350 header file e_os2.h as it now appears in public header file cms.h
8354 * New -sigopt option to the ca, req and x509 utilities. Additional
8355 signature parameters can be passed using this option and in
8360 * Add RSA PSS signing function. This will generate and set the
8361 appropriate AlgorithmIdentifiers for PSS based on those in the
8362 corresponding EVP_MD_CTX structure. No application support yet.
8366 * Support for companion algorithm specific ASN1 signing routines.
8367 New function ASN1_item_sign_ctx() signs a pre-initialised
8368 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8369 the appropriate parameters.
8373 * Add new algorithm specific ASN1 verification initialisation function
8374 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8375 handling will be the same no matter what EVP_PKEY_METHOD is used.
8376 Add a PSS handler to support verification of PSS signatures: checked
8377 against a number of sample certificates.
8381 * Add signature printing for PSS. Add PSS OIDs.
8383 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8385 * Add algorithm specific signature printing. An individual ASN1 method
8386 can now print out signatures instead of the standard hex dump.
8388 More complex signatures (e.g. PSS) can print out more meaningful
8389 information. Include DSA version that prints out the signature
8394 * Password based recipient info support for CMS library: implementing
8399 * Split password based encryption into PBES2 and PBKDF2 functions. This
8400 neatly separates the code into cipher and PBE sections and is required
8401 for some algorithms that split PBES2 into separate pieces (such as
8402 password based CMS).
8406 * Session-handling fixes:
8407 - Fix handling of connections that are resuming with a session ID,
8408 but also support Session Tickets.
8409 - Fix a bug that suppressed issuing of a new ticket if the client
8410 presented a ticket with an expired session.
8411 - Try to set the ticket lifetime hint to something reasonable.
8412 - Make tickets shorter by excluding irrelevant information.
8413 - On the client side, don't ignore renewed tickets.
8415 *Adam Langley, Bodo Moeller (Google)*
8417 * Fix PSK session representation.
8421 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8423 This work was sponsored by Intel.
8427 * Add GCM support to TLS library. Some custom code is needed to split
8428 the IV between the fixed (from PRF) and explicit (from TLS record)
8429 portions. This adds all GCM ciphersuites supported by RFC5288 and
8430 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8431 add a special AESGCM string for GCM only.
8435 * Expand range of ctrls for AES GCM. Permit setting invocation
8436 field on decrypt and retrieval of invocation field only on encrypt.
8440 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8441 As required by RFC5289 these ciphersuites cannot be used if for
8442 versions of TLS earlier than 1.2.
8446 * For FIPS capable OpenSSL interpret a NULL default public key method
8447 as unset and return the appropriate default but do *not* set the default.
8448 This means we can return the appropriate method in applications that
8449 switch between FIPS and non-FIPS modes.
8453 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8454 ENGINE is used then we cannot handle that in the FIPS module so we
8455 keep original code iff non-FIPS operations are allowed.
8459 * Add -attime option to openssl utilities.
8461 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8463 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8467 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8468 FIPS EC methods unconditionally for now.
8472 * New build option no-ec2m to disable characteristic 2 code.
8476 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8477 all cases can be covered as some introduce binary incompatibilities.
8481 * Redirect RSA operations to FIPS module including keygen,
8482 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8486 * Add similar low-level API blocking to ciphers.
8490 * low-level digest APIs are not approved in FIPS mode: any attempt
8491 to use these will cause a fatal error. Applications that *really* want
8492 to use them can use the `private_*` version instead.
8496 * Redirect cipher operations to FIPS module for FIPS builds.
8500 * Redirect digest operations to FIPS module for FIPS builds.
8504 * Update build system to add "fips" flag which will link in fipscanister.o
8505 for static and shared library builds embedding a signature if needed.
8509 * Output TLS supported curves in preference order instead of numerical
8510 order. This is currently hardcoded for the highest order curves first.
8511 This should be configurable so applications can judge speed vs strength.
8515 * Add TLS v1.2 server support for client authentication.
8519 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8524 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8525 FIPS modules versions.
8529 * Add TLS v1.2 client side support for client authentication. Keep cache
8530 of handshake records longer as we don't know the hash algorithm to use
8531 until after the certificate request message is received.
8535 * Initial TLS v1.2 client support. Add a default signature algorithms
8536 extension including all the algorithms we support. Parse new signature
8537 format in client key exchange. Relax some ECC signing restrictions for
8538 TLS v1.2 as indicated in RFC5246.
8542 * Add server support for TLS v1.2 signature algorithms extension. Switch
8543 to new signature format when needed using client digest preference.
8544 All server ciphersuites should now work correctly in TLS v1.2. No client
8545 support yet and no support for client certificates.
8549 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8550 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8551 ciphersuites. At present only RSA key exchange ciphersuites work with
8552 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8553 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8554 and version checking.
8558 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8559 with this defined it will not be affected by any changes to ssl internal
8560 structures. Add several utility functions to allow openssl application
8561 to work with OPENSSL_NO_SSL_INTERN defined.
8565 * A long standing patch to add support for SRP from EdelWeb (Peter
8566 Sylvester and Christophe Renou) was integrated.
8567 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8568 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8571 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8575 * Permit abbreviated handshakes when renegotiating using the function
8576 SSL_renegotiate_abbreviated().
8578 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8580 * Add call to ENGINE_register_all_complete() to
8581 ENGINE_load_builtin_engines(), so some implementations get used
8582 automatically instead of needing explicit application support.
8586 * Add support for TLS key exporter as described in RFC5705.
8588 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8590 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8591 a few changes are required:
8593 Add SSL_OP_NO_TLSv1_1 flag.
8594 Add TLSv1_1 methods.
8595 Update version checking logic to handle version 1.1.
8596 Add explicit IV handling (ported from DTLS code).
8597 Add command line options to s_client/s_server.
8604 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8606 * X509_ATTRIBUTE memory leak
8608 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8609 memory. This structure is used by the PKCS#7 and CMS routines so any
8610 application which reads PKCS#7 or CMS data from untrusted sources is
8611 affected. SSL/TLS is not affected.
8613 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8619 * Race condition handling PSK identify hint
8621 If PSK identity hints are received by a multi-threaded client then
8622 the values are wrongly updated in the parent SSL_CTX structure. This can
8623 result in a race condition potentially leading to a double free of the
8629 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8631 * Malformed ECParameters causes infinite loop
8633 When processing an ECParameters structure OpenSSL enters an infinite loop
8634 if the curve specified is over a specially malformed binary polynomial
8637 This can be used to perform denial of service against any
8638 system which processes public keys, certificate requests or
8639 certificates. This includes TLS clients and TLS servers with
8640 client authentication enabled.
8642 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8647 * Exploitable out-of-bounds read in X509_cmp_time
8649 X509_cmp_time does not properly check the length of the ASN1_TIME
8650 string and can read a few bytes out of bounds. In addition,
8651 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8654 An attacker can use this to craft malformed certificates and CRLs of
8655 various sizes and potentially cause a segmentation fault, resulting in
8656 a DoS on applications that verify certificates or CRLs. TLS clients
8657 that verify CRLs are affected. TLS clients and servers with client
8658 authentication enabled may be affected if they use custom verification
8661 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8662 independently by Hanno Böck.
8667 * PKCS7 crash with missing EnvelopedContent
8669 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8670 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8671 with missing content and trigger a NULL pointer dereference on parsing.
8673 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8674 structures from untrusted sources are affected. OpenSSL clients and
8675 servers are not affected.
8677 This issue was reported to OpenSSL by Michal Zalewski (Google).
8682 * CMS verify infinite loop with unknown hash function
8684 When verifying a signedData message the CMS code can enter an infinite loop
8685 if presented with an unknown hash function OID. This can be used to perform
8686 denial of service against any system which verifies signedData messages using
8688 This issue was reported to OpenSSL by Johannes Bauer.
8693 * Race condition handling NewSessionTicket
8695 If a NewSessionTicket is received by a multi-threaded client when attempting to
8696 reuse a previous ticket then a race condition can occur potentially leading to
8697 a double free of the ticket data.
8702 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8704 * Segmentation fault in ASN1_TYPE_cmp fix
8706 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8707 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8708 certificate signature algorithm consistency this can be used to crash any
8709 certificate verification operation and exploited in a DoS attack. Any
8710 application which performs certificate verification is vulnerable including
8711 OpenSSL clients and servers which enable client authentication.
8716 * ASN.1 structure reuse memory corruption fix
8718 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8719 memory corruption via an invalid write. Such reuse is and has been
8720 strongly discouraged and is believed to be rare.
8722 Applications that parse structures containing CHOICE or ANY DEFINED BY
8723 components may be affected. Certificate parsing (d2i_X509 and related
8724 functions) are however not affected. OpenSSL clients and servers are
8730 * PKCS7 NULL pointer dereferences fix
8732 The PKCS#7 parsing code does not handle missing outer ContentInfo
8733 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8734 missing content and trigger a NULL pointer dereference on parsing.
8736 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8737 otherwise parse PKCS#7 structures from untrusted sources are
8738 affected. OpenSSL clients and servers are not affected.
8740 This issue was reported to OpenSSL by Michal Zalewski (Google).
8745 * DoS via reachable assert in SSLv2 servers fix
8747 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8748 servers that both support SSLv2 and enable export cipher suites by sending
8749 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8751 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8752 (OpenSSL development team).
8757 * Use After Free following d2i_ECPrivatekey error fix
8759 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8760 could cause a use after free condition. This, in turn, could cause a double
8761 free in several private key parsing functions (such as d2i_PrivateKey
8762 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8763 for applications that receive EC private keys from untrusted
8764 sources. This scenario is considered rare.
8766 This issue was discovered by the BoringSSL project and fixed in their
8772 * X509_to_X509_REQ NULL pointer deref fix
8774 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8775 the certificate key is invalid. This function is rarely used in practice.
8777 This issue was discovered by Brian Carpenter.
8782 * Removed the export ciphers from the DEFAULT ciphers
8786 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8788 * Build fixes for the Windows and OpenVMS platforms
8790 *Matt Caswell and Richard Levitte*
8792 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8794 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8795 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8796 dereference. This could lead to a Denial Of Service attack. Thanks to
8797 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8802 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8803 dtls1_buffer_record function under certain conditions. In particular this
8804 could occur if an attacker sent repeated DTLS records with the same
8805 sequence number but for the next epoch. The memory leak could be exploited
8806 by an attacker in a Denial of Service attack through memory exhaustion.
8807 Thanks to Chris Mueller for reporting this issue.
8812 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8813 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8814 method would be set to NULL which could later result in a NULL pointer
8815 dereference. Thanks to Frank Schmirler for reporting this issue.
8820 * Abort handshake if server key exchange message is omitted for ephemeral
8823 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8824 reporting this issue.
8829 * Remove non-export ephemeral RSA code on client and server. This code
8830 violated the TLS standard by allowing the use of temporary RSA keys in
8831 non-export ciphersuites and could be used by a server to effectively
8832 downgrade the RSA key length used to a value smaller than the server
8833 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8834 INRIA or reporting this issue.
8839 * Fixed issue where DH client certificates are accepted without verification.
8840 An OpenSSL server will accept a DH certificate for client authentication
8841 without the certificate verify message. This effectively allows a client to
8842 authenticate without the use of a private key. This only affects servers
8843 which trust a client certificate authority which issues certificates
8844 containing DH keys: these are extremely rare and hardly ever encountered.
8845 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8851 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8852 results on some platforms, including x86_64. This bug occurs at random
8853 with a very low probability, and is not known to be exploitable in any
8854 way, though its exact impact is difficult to determine. Thanks to Pieter
8855 Wuille (Blockstream) who reported this issue and also suggested an initial
8856 fix. Further analysis was conducted by the OpenSSL development team and
8857 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8858 the OpenSSL core team.
8863 * Fix various certificate fingerprint issues.
8865 By using non-DER or invalid encodings outside the signed portion of a
8866 certificate the fingerprint can be changed without breaking the signature.
8867 Although no details of the signed portion of the certificate can be changed
8868 this can cause problems with some applications: e.g. those using the
8869 certificate fingerprint for blacklists.
8871 1. Reject signatures with non zero unused bits.
8873 If the BIT STRING containing the signature has non zero unused bits reject
8874 the signature. All current signature algorithms require zero unused bits.
8876 2. Check certificate algorithm consistency.
8878 Check the AlgorithmIdentifier inside TBS matches the one in the
8879 certificate signature. NB: this will result in signature failure
8880 errors for some broken certificates.
8882 Thanks to Konrad Kraszewski from Google for reporting this issue.
8884 3. Check DSA/ECDSA signatures use DER.
8886 Re-encode DSA/ECDSA signatures and compare with the original received
8887 signature. Return an error if there is a mismatch.
8889 This will reject various cases including garbage after signature
8890 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8891 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8892 (negative or with leading zeroes).
8894 Further analysis was conducted and fixes were developed by Stephen Henson
8895 of the OpenSSL core team.
8901 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8903 * Session Ticket Memory Leak.
8905 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8906 integrity of that ticket is first verified. In the event of a session
8907 ticket integrity check failing, OpenSSL will fail to free memory
8908 causing a memory leak. By sending a large number of invalid session
8909 tickets an attacker could exploit this issue in a Denial Of Service
8915 * Build option no-ssl3 is incomplete.
8917 When OpenSSL is configured with "no-ssl3" as a build option, servers
8918 could accept and complete an SSL 3.0 handshake, and clients could be
8919 configured to send them.
8922 *Akamai and the OpenSSL team*
8924 * Add support for TLS_FALLBACK_SCSV.
8925 Client applications doing fallback retries should call
8926 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8929 *Adam Langley, Bodo Moeller*
8931 * Add additional DigestInfo checks.
8933 Re-encode DigestInto in DER and check against the original when
8934 verifying RSA signature: this will reject any improperly encoded
8935 DigestInfo structures.
8937 Note: this is a precautionary measure and no attacks are currently known.
8941 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8943 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8944 to a denial of service attack. A malicious server can crash the client
8945 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8946 ciphersuite and sending carefully crafted handshake messages.
8948 Thanks to Felix Gröbert (Google) for discovering and researching this
8954 * By sending carefully crafted DTLS packets an attacker could cause openssl
8955 to leak memory. This can be exploited through a Denial of Service attack.
8956 Thanks to Adam Langley for discovering and researching this issue.
8961 * An attacker can force openssl to consume large amounts of memory whilst
8962 processing DTLS handshake messages. This can be exploited through a
8963 Denial of Service attack.
8964 Thanks to Adam Langley for discovering and researching this issue.
8969 * An attacker can force an error condition which causes openssl to crash
8970 whilst processing DTLS packets due to memory being freed twice. This
8971 can be exploited through a Denial of Service attack.
8972 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8978 * If a multithreaded client connects to a malicious server using a resumed
8979 session and the server sends an ec point format extension it could write
8980 up to 255 bytes to freed memory.
8982 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8988 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8989 X509_name_oneline, X509_name_print_ex et al. to leak some information
8990 from the stack. Applications may be affected if they echo pretty printing
8991 output to the attacker.
8993 Thanks to Ivan Fratric (Google) for discovering this issue.
8996 *Emilia Käsper, and Steve Henson*
8998 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8999 for corner cases. (Certain input points at infinity could lead to
9000 bogus results, with non-infinity inputs mapped to infinity too.)
9004 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
9006 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
9007 handshake can force the use of weak keying material in OpenSSL
9008 SSL/TLS clients and servers.
9010 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
9011 researching this issue. ([CVE-2014-0224])
9013 *KIKUCHI Masashi, Steve Henson*
9015 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
9016 OpenSSL DTLS client the code can be made to recurse eventually crashing
9019 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
9022 *Imre Rad, Steve Henson*
9024 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
9025 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
9026 client or server. This is potentially exploitable to run arbitrary
9027 code on a vulnerable client or server.
9029 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
9031 *Jüri Aedla, Steve Henson*
9033 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
9034 are subject to a denial of service attack.
9036 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
9037 this issue. ([CVE-2014-3470])
9039 *Felix Gröbert, Ivan Fratric, Steve Henson*
9041 * Harmonize version and its documentation. -f flag is used to display
9044 *mancha <mancha1@zoho.com>*
9046 * Fix eckey_priv_encode so it immediately returns an error upon a failure
9047 in i2d_ECPrivateKey.
9049 *mancha <mancha1@zoho.com>*
9051 * Fix some double frees. These are not thought to be exploitable.
9053 *mancha <mancha1@zoho.com>*
9055 * Fix for the attack described in the paper "Recovering OpenSSL
9056 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
9057 by Yuval Yarom and Naomi Benger. Details can be obtained from:
9058 <http://eprint.iacr.org/2014/140>
9060 Thanks to Yuval Yarom and Naomi Benger for discovering this
9061 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
9063 *Yuval Yarom and Naomi Benger*
9065 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
9067 * Keep original DTLS digest and encryption contexts in retransmission
9068 structures so we can use the previous session parameters if they need
9069 to be resent. ([CVE-2013-6450])
9073 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
9074 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
9075 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
9076 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
9077 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
9078 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
9080 *Rob Stradling, Adam Langley*
9082 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
9084 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
9086 This addresses the flaw in CBC record processing discovered by
9087 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
9088 at: <http://www.isg.rhul.ac.uk/tls/>
9090 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9091 Security Group at Royal Holloway, University of London
9092 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
9093 Emilia Käsper for the initial patch.
9096 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
9098 * Return an error when checking OCSP signatures when key is NULL.
9099 This fixes a DoS attack. ([CVE-2013-0166])
9103 * Call OCSP Stapling callback after ciphersuite has been chosen, so
9104 the right response is stapled. Also change SSL_get_certificate()
9105 so it returns the certificate actually sent.
9106 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
9107 (This is a backport)
9109 *Rob Stradling <rob.stradling@comodo.com>*
9111 * Fix possible deadlock when decoding public keys.
9115 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
9117 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
9120 * Sanity check record length before skipping explicit IV in DTLS
9123 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
9124 fuzzing as a service testing platform.
9129 * Initialise tkeylen properly when encrypting CMS messages.
9130 Thanks to Solar Designer of Openwall for reporting this issue.
9134 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
9136 * Check for potentially exploitable overflows in asn1_d2i_read_bio
9137 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
9138 in CRYPTO_realloc_clean.
9140 Thanks to Tavis Ormandy, Google Security Team, for discovering this
9141 issue and to Adam Langley <agl@chromium.org> for fixing it.
9144 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
9146 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
9148 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
9149 in CMS and PKCS7 code. When RSA decryption fails use a random key for
9150 content decryption and always return the same error. Note: this attack
9151 needs on average 2^20 messages so it only affects automated senders. The
9152 old behaviour can be re-enabled in the CMS code by setting the
9153 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
9154 an MMA defence is not necessary.
9155 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
9156 this issue. ([CVE-2012-0884])
9160 * Fix CVE-2011-4619: make sure we really are receiving a
9161 client hello before rejecting multiple SGC restarts. Thanks to
9162 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
9166 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
9168 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
9169 Thanks to Antonio Martin, Enterprise Secure Access Research and
9170 Development, Cisco Systems, Inc. for discovering this bug and
9171 preparing a fix. ([CVE-2012-0050])
9175 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
9177 * Nadhem Alfardan and Kenny Paterson have discovered an extension
9178 of the Vaudenay padding oracle attack on CBC mode encryption
9179 which enables an efficient plaintext recovery attack against
9180 the OpenSSL implementation of DTLS. Their attack exploits timing
9181 differences arising during decryption processing. A research
9182 paper describing this attack can be found at:
9183 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
9184 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9185 Security Group at Royal Holloway, University of London
9186 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9187 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9188 for preparing the fix. ([CVE-2011-4108])
9190 *Robin Seggelmann, Michael Tuexen*
9192 * Clear bytes used for block padding of SSL 3.0 records.
9195 *Adam Langley (Google)*
9197 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9198 Kadianakis <desnacked@gmail.com> for discovering this issue and
9199 Adam Langley for preparing the fix. ([CVE-2011-4619])
9201 *Adam Langley (Google)*
9203 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9205 *Andrey Kulikov <amdeich@gmail.com>*
9207 * Prevent malformed RFC3779 data triggering an assertion failure.
9208 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9209 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9211 *Rob Austein <sra@hactrn.net>*
9213 * Improved PRNG seeding for VOS.
9215 *Paul Green <Paul.Green@stratus.com>*
9217 * Fix ssl_ciph.c set-up race.
9219 *Adam Langley (Google)*
9221 * Fix spurious failures in ecdsatest.c.
9223 *Emilia Käsper (Google)*
9225 * Fix the BIO_f_buffer() implementation (which was mixing different
9226 interpretations of the `..._len` fields).
9228 *Adam Langley (Google)*
9230 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9231 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9232 threads won't reuse the same blinding coefficients.
9234 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9235 lock to call BN_BLINDING_invert_ex, and avoids one use of
9236 BN_BLINDING_update for each BN_BLINDING structure (previously,
9237 the last update always remained unused).
9239 *Emilia Käsper (Google)*
9241 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9243 *Bob Buckholz (Google)*
9245 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9247 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9248 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9250 *Kaspar Brand <ossl@velox.ch>*
9252 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9253 for multi-threaded use of ECDH. ([CVE-2011-3210])
9255 *Adam Langley (Google)*
9257 * Fix x509_name_ex_d2i memory leak on bad inputs.
9261 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9262 signature public key algorithm by using OID xref utilities instead.
9263 Before this you could only use some ECC ciphersuites with SHA1 only.
9267 * Add protection against ECDSA timing attacks as mentioned in the paper
9268 by Billy Bob Brumley and Nicola Tuveri, see:
9269 <http://eprint.iacr.org/2011/232.pdf>
9271 *Billy Bob Brumley and Nicola Tuveri*
9273 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9275 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9277 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9279 * Fix bug in string printing code: if *any* escaping is enabled we must
9280 escape the escape character (backslash) or the resulting string is
9285 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9287 * Disable code workaround for ancient and obsolete Netscape browsers
9288 and servers: an attacker can use it in a ciphersuite downgrade attack.
9289 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9293 * Fixed J-PAKE implementation error, originally discovered by
9294 Sebastien Martini, further info and confirmation from Stefan
9295 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9299 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9301 * Fix extension code to avoid race conditions which can result in a buffer
9302 overrun vulnerability: resumed sessions must not be modified as they can
9303 be shared by multiple threads. CVE-2010-3864
9307 * Fix WIN32 build system to correctly link an ENGINE directory into
9312 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9314 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9317 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9319 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9321 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9322 context. The operation can be customised via the ctrl mechanism in
9323 case ENGINEs want to include additional functionality.
9327 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9331 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9332 output hashes compatible with older versions of OpenSSL.
9334 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9336 * Fix compression algorithm handling: if resuming a session use the
9337 compression algorithm of the resumed session instead of determining
9338 it from client hello again. Don't allow server to change algorithm.
9342 * Add load_crls() function to commands tidying load_certs() too. Add option
9343 to verify utility to allow additional CRLs to be included.
9347 * Update OCSP request code to permit adding custom headers to the request:
9348 some responders need this.
9352 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9355 *Julia Lawall <julia@diku.dk>*
9357 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9358 needlessly dereferenced structures, used obsolete functions and
9359 didn't handle all updated verify codes correctly.
9363 * Disable MD2 in the default configuration.
9367 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9368 indicate the initial BIO being pushed or popped. This makes it possible
9369 to determine whether the BIO is the one explicitly called or as a result
9370 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9371 it handles reference counts correctly and doesn't zero out the I/O bio
9372 when it is not being explicitly popped. WARNING: applications which
9373 included workarounds for the old buggy behaviour will need to be modified
9374 or they could free up already freed BIOs.
9378 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9379 renaming to all platforms (within the 0.9.8 branch, this was
9380 done conditionally on Netware platforms to avoid a name clash).
9382 *Guenter <lists@gknw.net>*
9384 * Add ECDHE and PSK support to DTLS.
9386 *Michael Tuexen <tuexen@fh-muenster.de>*
9388 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9393 * Add "missing" function EVP_MD_flags() (without this the only way to
9394 retrieve a digest flags is by accessing the structure directly. Update
9395 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9396 or cipher is registered as in the "from" argument. Print out all
9397 registered digests in the dgst usage message instead of manually
9398 attempting to work them out.
9402 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9403 this allows the use of compression and extensions. Change default cipher
9404 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9405 by default unless an application cipher string requests it.
9409 * Alter match criteria in PKCS12_parse(). It used to try to use local
9410 key ids to find matching certificates and keys but some PKCS#12 files
9411 don't follow the (somewhat unwritten) rules and this strategy fails.
9412 Now just gather all certificates together and the first private key
9413 then look for the first certificate that matches the key.
9417 * Support use of registered digest and cipher names for dgst and cipher
9418 commands instead of having to add each one as a special case. So now
9425 openssl dgst -sha256 foo
9427 and this works for ENGINE based algorithms too.
9431 * Update Gost ENGINE to support parameter files.
9433 *Victor B. Wagner <vitus@cryptocom.ru>*
9435 * Support GeneralizedTime in ca utility.
9437 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9439 * Enhance the hash format used for certificate directory links. The new
9440 form uses the canonical encoding (meaning equivalent names will work
9441 even if they aren't identical) and uses SHA1 instead of MD5. This form
9442 is incompatible with the older format and as a result c_rehash should
9443 be used to rebuild symbolic links.
9447 * Make PKCS#8 the default write format for private keys, replacing the
9448 traditional format. This form is standardised, more secure and doesn't
9449 include an implicit MD5 dependency.
9453 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9454 committed to OpenSSL should pass this lot as a minimum.
9458 * Add session ticket override functionality for use by EAP-FAST.
9460 *Jouni Malinen <j@w1.fi>*
9462 * Modify HMAC functions to return a value. Since these can be implemented
9463 in an ENGINE errors can occur.
9467 * Type-checked OBJ_bsearch_ex.
9471 * Type-checked OBJ_bsearch. Also some constification necessitated
9472 by type-checking. Still to come: TXT_DB, bsearch(?),
9473 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9478 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9479 seconds to a tm structure directly, instead of going through OS
9480 specific date routines. This avoids any issues with OS routines such
9481 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9482 and X509_time_adj_ex() to cover the extended range. The existing
9483 X509_time_adj() is still usable and will no longer have any date issues.
9487 * Delta CRL support. New use deltas option which will attempt to locate
9488 and search any appropriate delta CRLs available.
9490 This work was sponsored by Google.
9494 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9495 code and add additional score elements. Validate alternate CRL paths
9496 as part of the CRL checking and indicate a new error "CRL path validation
9497 error" in this case. Applications wanting additional details can use
9498 the verify callback and check the new "parent" field. If this is not
9499 NULL CRL path validation is taking place. Existing applications won't
9500 see this because it requires extended CRL support which is off by
9503 This work was sponsored by Google.
9507 * Support for freshest CRL extension.
9509 This work was sponsored by Google.
9513 * Initial indirect CRL support. Currently only supported in the CRLs
9514 passed directly and not via lookup. Process certificate issuer
9515 CRL entry extension and lookup CRL entries by bother issuer name
9516 and serial number. Check and process CRL issuer entry in IDP extension.
9518 This work was sponsored by Google.
9522 * Add support for distinct certificate and CRL paths. The CRL issuer
9523 certificate is validated separately in this case. Only enabled if
9524 an extended CRL support flag is set: this flag will enable additional
9525 CRL functionality in future.
9527 This work was sponsored by Google.
9531 * Add support for policy mappings extension.
9533 This work was sponsored by Google.
9537 * Fixes to pathlength constraint, self issued certificate handling,
9538 policy processing to align with RFC3280 and PKITS tests.
9540 This work was sponsored by Google.
9544 * Support for name constraints certificate extension. DN, email, DNS
9545 and URI types are currently supported.
9547 This work was sponsored by Google.
9551 * To cater for systems that provide a pointer-based thread ID rather
9552 than numeric, deprecate the current numeric thread ID mechanism and
9553 replace it with a structure and associated callback type. This
9554 mechanism allows a numeric "hash" to be extracted from a thread ID in
9555 either case, and on platforms where pointers are larger than 'long',
9556 mixing is done to help ensure the numeric 'hash' is usable even if it
9557 can't be guaranteed unique. The default mechanism is to use "&errno"
9558 as a pointer-based thread ID to distinguish between threads.
9560 Applications that want to provide their own thread IDs should now use
9561 CRYPTO_THREADID_set_callback() to register a callback that will call
9562 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9564 Note that ERR_remove_state() is now deprecated, because it is tied
9565 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9566 to free the current thread's error state should be replaced by
9567 ERR_remove_thread_state(NULL).
9569 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9570 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9571 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9572 application was previously providing a numeric thread callback that
9573 was inappropriate for distinguishing threads, then uniqueness might
9574 have been obtained with &errno that happened immediately in the
9575 intermediate development versions of OpenSSL; this is no longer the
9576 case, the numeric thread callback will now override the automatic use
9579 *Geoff Thorpe, with help from Bodo Moeller*
9581 * Initial support for different CRL issuing certificates. This covers a
9582 simple case where the self issued certificates in the chain exist and
9583 the real CRL issuer is higher in the existing chain.
9585 This work was sponsored by Google.
9589 * Removed effectively defunct crypto/store from the build.
9593 * Revamp of STACK to provide stronger type-checking. Still to come:
9594 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9595 ASN1_STRING, CONF_VALUE.
9599 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9600 RAM on SSL connections. This option can save about 34k per idle SSL.
9604 * Revamp of LHASH to provide stronger type-checking. Still to come:
9605 STACK, TXT_DB, bsearch, qsort.
9609 * Initial support for Cryptographic Message Syntax (aka CMS) based
9610 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9611 support for data, signedData, compressedData, digestedData and
9612 encryptedData, envelopedData types included. Scripts to check against
9613 RFC4134 examples draft and interop and consistency checks of many
9614 content types and variants.
9618 * Add options to enc utility to support use of zlib compression BIO.
9622 * Extend mk1mf to support importing of options and assembly language
9623 files from Configure script, currently only included in VC-WIN32.
9624 The assembly language rules can now optionally generate the source
9625 files from the associated perl scripts.
9629 * Implement remaining functionality needed to support GOST ciphersuites.
9630 Interop testing has been performed using CryptoPro implementations.
9632 *Victor B. Wagner <vitus@cryptocom.ru>*
9634 * s390x assembler pack.
9638 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9643 * Implement Opaque PRF Input TLS extension as specified in
9644 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9645 official specification yet and no extension type assignment by
9646 IANA exists, this extension (for now) will have to be explicitly
9647 enabled when building OpenSSL by providing the extension number
9648 to use. For example, specify an option
9650 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9652 to the "config" or "Configure" script to enable the extension,
9653 assuming extension number 0x9527 (which is a completely arbitrary
9654 and unofficial assignment based on the MD5 hash of the Internet
9655 Draft). Note that by doing so, you potentially lose
9656 interoperability with other TLS implementations since these might
9657 be using the same extension number for other purposes.
9659 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9660 opaque PRF input value to use in the handshake. This will create
9661 an internal copy of the length-'len' string at 'src', and will
9662 return non-zero for success.
9664 To get more control and flexibility, provide a callback function
9667 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9668 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9672 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9675 Callback function 'cb' will be called in handshakes, and is
9676 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9677 Argument 'arg' is for application purposes (the value as given to
9678 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9679 be provided to the callback function). The callback function
9680 has to return non-zero to report success: usually 1 to use opaque
9681 PRF input just if possible, or 2 to enforce use of the opaque PRF
9682 input. In the latter case, the library will abort the handshake
9683 if opaque PRF input is not successfully negotiated.
9685 Arguments 'peerinput' and 'len' given to the callback function
9686 will always be NULL and 0 in the case of a client. A server will
9687 see the client's opaque PRF input through these variables if
9688 available (NULL and 0 otherwise). Note that if the server
9689 provides an opaque PRF input, the length must be the same as the
9690 length of the client's opaque PRF input.
9692 Note that the callback function will only be called when creating
9693 a new session (session resumption can resume whatever was
9694 previously negotiated), and will not be called in SSL 2.0
9695 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9696 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9697 for applications that need to enforce opaque PRF input.
9701 * Update ssl code to support digests other than SHA1+MD5 for handshake
9704 *Victor B. Wagner <vitus@cryptocom.ru>*
9706 * Add RFC4507 support to OpenSSL. This includes the corrections in
9707 RFC4507bis. The encrypted ticket format is an encrypted encoded
9708 SSL_SESSION structure, that way new session features are automatically
9711 If a client application caches session in an SSL_SESSION structure
9712 support is transparent because tickets are now stored in the encoded
9715 The SSL_CTX structure automatically generates keys for ticket
9716 protection in servers so again support should be possible
9717 with no application modification.
9719 If a client or server wishes to disable RFC4507 support then the option
9720 SSL_OP_NO_TICKET can be set.
9722 Add a TLS extension debugging callback to allow the contents of any client
9723 or server extensions to be examined.
9725 This work was sponsored by Google.
9729 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9730 OpenSSL should now compile cleanly on gcc 4.2
9732 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9734 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9735 support including streaming MAC support: this is required for GOST
9736 ciphersuite support.
9738 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9740 * Add option -stream to use PKCS#7 streaming in smime utility. New
9741 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9742 to output in BER and PEM format.
9746 * Experimental support for use of HMAC via EVP_PKEY interface. This
9747 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9748 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9749 ENGINE support for HMAC keys which are unextractable. New -mac and
9750 -macopt options to dgst utility.
9754 * New option -sigopt to dgst utility. Update dgst to use
9755 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9756 alternative signing parameters such as X9.31 or PSS in the dgst
9761 * Change ssl_cipher_apply_rule(), the internal function that does
9762 the work each time a ciphersuite string requests enabling
9763 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9764 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9765 the order of disabled ciphersuites such that those ciphersuites
9766 that most recently went from enabled to disabled not only stay
9767 in order with respect to each other, but also have higher priority
9768 than other disabled ciphersuites the next time ciphersuites are
9771 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9772 the same ciphersuites as with "HIGH" alone, but in a specific
9773 order where the PSK ciphersuites come first (since they are the
9774 most recently disabled ciphersuites when "HIGH" is parsed).
9776 Also, change ssl_create_cipher_list() (using this new
9777 functionality) such that between otherwise identical
9778 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9783 * Change ssl_create_cipher_list() so that it automatically
9784 arranges the ciphersuites in reasonable order before starting
9785 to process the rule string. Thus, the definition for "DEFAULT"
9786 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9787 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9788 This makes it much easier to arrive at a reasonable default order
9789 in applications for which anonymous ciphers are OK (meaning
9790 that you can't actually use DEFAULT).
9792 *Bodo Moeller; suggested by Victor Duchovni*
9794 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9795 processing) into multiple integers instead of setting
9796 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9797 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9798 (These masks as well as the individual bit definitions are hidden
9799 away into the non-exported interface ssl/ssl_locl.h, so this
9800 change to the definition of the SSL_CIPHER structure shouldn't
9801 affect applications.) This give us more bits for each of these
9802 categories, so there is no longer a need to coagulate AES128 and
9803 AES256 into a single algorithm bit, and to coagulate Camellia128
9804 and Camellia256 into a single algorithm bit, which has led to all
9807 Thus, among other things, the kludge introduced in 0.9.7m and
9808 0.9.8e for masking out AES256 independently of AES128 or masking
9809 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9811 With the change, we also introduce new ciphersuite aliases that
9812 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9817 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9818 Use the leftmost N bytes of the signature input if the input is
9819 larger than the prime q (with N being the size in bytes of q).
9823 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9824 it yet and it is largely untested.
9828 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9832 * Initial incomplete changes to avoid need for function casts in OpenSSL
9833 some compilers (gcc 4.2 and later) reject their use. Safestack is
9834 reimplemented. Update ASN1 to avoid use of legacy functions.
9838 * Win32/64 targets are linked with Winsock2.
9842 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9843 to external functions. This can be used to increase CRL handling
9844 efficiency especially when CRLs are very large by (for example) storing
9845 the CRL revoked certificates in a database.
9849 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9850 new CRLs added to a directory can be used. New command line option
9851 -verify_return_error to s_client and s_server. This causes real errors
9852 to be returned by the verify callback instead of carrying on no matter
9853 what. This reflects the way a "real world" verify callback would behave.
9857 * GOST engine, supporting several GOST algorithms and public key formats.
9858 Kindly donated by Cryptocom.
9862 * Partial support for Issuing Distribution Point CRL extension. CRLs
9863 partitioned by DP are handled but no indirect CRL or reason partitioning
9864 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9865 selected via a scoring technique which handles IDP and AKID in CRLs.
9869 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9870 will ultimately be used for all verify operations: this will remove the
9871 X509_STORE dependency on certificate verification and allow alternative
9872 lookup methods. X509_STORE based implementations of these two callbacks.
9876 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9877 Modify get_crl() to find a valid (unexpired) CRL if possible.
9881 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9882 this would be called X509_CRL_cmp() but that name is already used by
9883 a function that just compares CRL issuer names. Cache several CRL
9884 extensions in X509_CRL structure and cache CRLDP in X509.
9888 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9889 this maps equivalent X509_NAME structures into a consistent structure.
9890 Name comparison can then be performed rapidly using memcmp().
9894 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9899 * Allow digests to supply their own micalg string for S/MIME type using
9900 the ctrl EVP_MD_CTRL_MICALG.
9904 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9905 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9906 ctrl. It can then customise the structure before and/or after signing
9911 * New function OBJ_add_sigid() to allow application defined signature OIDs
9912 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9913 to free up any added signature OIDs.
9917 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9918 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9919 digest and cipher tables. New options added to openssl utility:
9920 list-message-digest-algorithms and list-cipher-algorithms.
9924 * Change the array representation of binary polynomials: the list
9925 of degrees of non-zero coefficients is now terminated with -1.
9926 Previously it was terminated with 0, which was also part of the
9927 value; thus, the array representation was not applicable to
9928 polynomials where t^0 has coefficient zero. This change makes
9929 the array representation useful in a more general context.
9933 * Various modifications and fixes to SSL/TLS cipher string
9934 handling. For ECC, the code now distinguishes between fixed ECDH
9935 with RSA certificates on the one hand and with ECDSA certificates
9936 on the other hand, since these are separate ciphersuites. The
9937 unused code for Fortezza ciphersuites has been removed.
9939 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9940 (not "ECDHE"). For consistency with the code for DH
9941 certificates, use of ECDH certificates is now considered ECDH
9942 authentication, not RSA or ECDSA authentication (the latter is
9943 merely the CA's signing algorithm and not actively used in the
9946 The temporary ciphersuite alias "ECCdraft" is no longer
9947 available, and ECC ciphersuites are no longer excluded from "ALL"
9948 and "DEFAULT". The following aliases now exist for RFC 4492
9949 ciphersuites, most of these by analogy with the DH case:
9951 kECDHr - ECDH cert, signed with RSA
9952 kECDHe - ECDH cert, signed with ECDSA
9953 kECDH - ECDH cert (signed with either RSA or ECDSA)
9954 kEECDH - ephemeral ECDH
9955 ECDH - ECDH cert or ephemeral ECDH
9961 AECDH - anonymous ECDH
9962 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9966 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9967 Use correct micalg parameters depending on digest(s) in signed message.
9971 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9972 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9976 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9977 an engine to register a method. Add ENGINE lookups for methods and
9978 functional reference processing.
9982 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9983 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9988 * New -resign option to smime utility. This adds one or more signers
9989 to an existing PKCS#7 signedData structure. Also -md option to use an
9990 alternative message digest algorithm for signing.
9994 * Tidy up PKCS#7 routines and add new functions to make it easier to
9995 create PKCS7 structures containing multiple signers. Update smime
9996 application to support multiple signers.
10000 * New -macalg option to pkcs12 utility to allow setting of an alternative
10005 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
10006 Reorganize PBE internals to lookup from a static table using NIDs,
10007 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
10008 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
10009 PRF which will be automatically used with PBES2.
10013 * Replace the algorithm specific calls to generate keys in "req" with the
10018 * Update PKCS#7 enveloped data routines to use new API. This is now
10019 supported by any public key method supporting the encrypt operation. A
10020 ctrl is added to allow the public key algorithm to examine or modify
10021 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
10026 * Add a ctrl to asn1 method to allow a public key algorithm to express
10027 a default digest type to use. In most cases this will be SHA1 but some
10028 algorithms (such as GOST) need to specify an alternative digest. The
10029 return value indicates how strong the preference is 1 means optional and
10030 2 is mandatory (that is it is the only supported type). Modify
10031 ASN1_item_sign() to accept a NULL digest argument to indicate it should
10032 use the default md. Update openssl utilities to use the default digest
10033 type for signing if it is not explicitly indicated.
10037 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
10038 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
10039 signing method from the key type. This effectively removes the link
10040 between digests and public key types.
10044 * Add an OID cross reference table and utility functions. Its purpose is to
10045 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
10046 rsaEncryption. This will allow some of the algorithm specific hackery
10047 needed to use the correct OID to be removed.
10051 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
10052 structures for PKCS7_sign(). They are now set up by the relevant public
10057 * Add provisional EC pkey method with support for ECDSA and ECDH.
10061 * Add support for key derivation (agreement) in the API, DH method and
10066 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
10067 public and private key formats. As a side effect these add additional
10068 command line functionality not previously available: DSA signatures can be
10069 generated and verified using pkeyutl and DH key support and generation in
10076 *Oliver Tappe <zooey@hirschkaefer.de>*
10078 * New make target "install_html_docs" installs HTML renditions of the
10081 *Oliver Tappe <zooey@hirschkaefer.de>*
10083 * New utility "genpkey" this is analogous to "genrsa" etc except it can
10084 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
10085 support key and parameter generation and add initial key generation
10086 functionality for RSA.
10090 * Add functions for main EVP_PKEY_method operations. The undocumented
10091 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
10092 `EVP_PKEY_{encrypt,decrypt}_old`.
10096 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
10097 key API, doesn't do much yet.
10101 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
10102 public key algorithms. New option to openssl utility:
10103 "list-public-key-algorithms" to print out info.
10107 * Implement the Supported Elliptic Curves Extension for
10108 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10112 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
10113 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
10117 * New utilities pkey and pkeyparam. These are similar to algorithm specific
10118 utilities such as rsa, dsa, dsaparam etc except they process any key
10123 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
10124 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
10125 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
10130 * Initial support for pluggable public key ASN1.
10131 De-spaghettify the public key ASN1 handling. Move public and private
10132 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
10133 algorithm specific handling to a single module within the relevant
10134 algorithm directory. Add functions to allow (near) opaque processing
10135 of public and private key structures.
10139 * Implement the Supported Point Formats Extension for
10140 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10144 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
10145 for the psk identity [hint] and the psk callback functions to the
10146 SSL_SESSION, SSL and SSL_CTX structure.
10149 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
10153 SSL_CTX_use_psk_identity_hint
10154 SSL_get_psk_identity_hint
10155 SSL_get_psk_identity
10156 SSL_use_psk_identity_hint
10158 *Mika Kousa and Pasi Eronen of Nokia Corporation*
10160 * Add RFC 3161 compliant time stamp request creation, response generation
10161 and response verification functionality.
10163 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
10165 * Add initial support for TLS extensions, specifically for the server_name
10166 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10167 have new members for a hostname. The SSL data structure has an
10168 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10169 stored in that context to allow for session resumption, even after the
10170 SSL has been switched to a new SSL_CTX in reaction to a client's
10171 server_name extension.
10173 New functions (subject to change):
10175 SSL_get_servername()
10176 SSL_get_servername_type()
10179 New CTRL codes and macros (subject to change):
10181 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10182 - SSL_CTX_set_tlsext_servername_callback()
10183 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10184 - SSL_CTX_set_tlsext_servername_arg()
10185 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10187 openssl s_client has a new '-servername ...' option.
10189 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10190 '-key2 ...', '-servername_fatal' (subject to change). This allows
10191 testing the HostName extension for a specific single hostname ('-cert'
10192 and '-key' remain fallbacks for handshakes without HostName
10193 negotiation). If the unrecognized_name alert has to be sent, this by
10194 default is a warning; it becomes fatal with the '-servername_fatal'
10197 *Peter Sylvester, Remy Allais, Christophe Renou*
10199 * Whirlpool hash implementation is added.
10203 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10204 bn(64,32). Because of instruction set limitations it doesn't have
10205 any negative impact on performance. This was done mostly in order
10206 to make it possible to share assembler modules, such as bn_mul_mont
10207 implementations, between 32- and 64-bit builds without hassle.
10211 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10212 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10217 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10218 dedicated Montgomery multiplication procedure, is introduced.
10219 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10220 "64-bit" performance on certain 32-bit targets.
10224 * New option SSL_OP_NO_COMP to disable use of compression selectively
10225 in SSL structures. New SSL ctrl to set maximum send fragment size.
10226 Save memory by setting the I/O buffer sizes dynamically instead of
10227 using the maximum available value.
10231 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10232 in addition to the text details.
10236 * Very, very preliminary EXPERIMENTAL support for printing of general
10237 ASN1 structures. This currently produces rather ugly output and doesn't
10238 handle several customised structures at all.
10242 * Integrated support for PVK file format and some related formats such
10243 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10244 these in the 'rsa' and 'dsa' utilities.
10248 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10252 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10253 place for the (very old) "NETSCAPE" format certificates which are now
10254 handled using new ASN1 code equivalents.
10258 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10259 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10260 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10264 * Modify CRL distribution points extension code to print out previously
10265 unsupported fields. Enhance extension setting code to allow setting of
10270 * Add print and set support for Issuing Distribution Point CRL extension.
10274 * Change 'Configure' script to enable Camellia by default.
10281 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10283 * When rejecting SSL/TLS records due to an incorrect version number, never
10284 update s->server with a new major version number. As of
10285 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10286 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10287 the previous behavior could result in a read attempt at NULL when
10288 receiving specific incorrect SSL/TLS records once record payload
10289 protection is active. ([CVE-2010-0740])
10291 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10293 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10294 could be crashed if the relevant tables were not present (e.g. chrooted).
10296 *Tomas Hoger <thoger@redhat.com>*
10298 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10300 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10302 *Martin Olsson, Neel Mehta*
10304 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10305 accommodate for stack sorting, always a write lock!).
10309 * On some versions of WIN32 Heap32Next is very slow. This can cause
10310 excessive delays in the RAND_poll(): over a minute. As a workaround
10311 include a time check in the inner Heap32Next loop too.
10315 * The code that handled flushing of data in SSL/TLS originally used the
10316 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10317 the problem outlined in PR#1949. The fix suggested there however can
10318 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10319 of Apache). So instead simplify the code to flush unconditionally.
10320 This should be fine since flushing with no data to flush is a no op.
10324 * Handle TLS versions 2.0 and later properly and correctly use the
10325 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10326 off ancient servers have a habit of sticking around for a while...
10330 * Modify compression code so it frees up structures without using the
10331 ex_data callbacks. This works around a problem where some applications
10332 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10333 restarting) then use compression (e.g. SSL with compression) later.
10334 This results in significant per-connection memory leaks and
10335 has caused some security issues including CVE-2008-1678 and
10340 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10341 change when encrypting or decrypting.
10345 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10346 connect and renegotiate with servers which do not support RI.
10347 Until RI is more widely deployed this option is enabled by default.
10351 * Add "missing" ssl ctrls to clear options and mode.
10355 * If client attempts to renegotiate and doesn't support RI respond with
10356 a no_renegotiation alert as required by RFC5746. Some renegotiating
10357 TLS clients will continue a connection gracefully when they receive
10358 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10359 waiting for a server hello which it will never receive. Now we treat a
10360 received no_renegotiation alert as a fatal error. This is because
10361 applications requesting a renegotiation might well expect it to succeed
10362 and would have no code in place to handle the server denying it so the
10363 only safe thing to do is to terminate the connection.
10367 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10368 peer supports secure renegotiation and 0 otherwise. Print out peer
10369 renegotiation support in s_client/s_server.
10373 * Replace the highly broken and deprecated SPKAC certification method with
10374 the updated NID creation version. This should correctly handle UTF8.
10378 * Implement RFC5746. Re-enable renegotiation but require the extension
10379 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10380 turns out to be a bad idea. It has been replaced by
10381 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10382 SSL_CTX_set_options(). This is really not recommended unless you
10383 know what you are doing.
10385 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10387 * Fixes to stateless session resumption handling. Use initial_ctx when
10388 issuing and attempting to decrypt tickets in case it has changed during
10389 servername handling. Use a non-zero length session ID when attempting
10390 stateless session resumption: this makes it possible to determine if
10391 a resumption has occurred immediately after receiving server hello
10392 (several places in OpenSSL subtly assume this) instead of later in
10397 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10398 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10399 fixes for a few places where the return code is not checked
10402 *Julia Lawall <julia@diku.dk>*
10404 * Add --strict-warnings option to Configure script to include devteam
10405 warnings in other configurations.
10409 * Add support for --libdir option and LIBDIR variable in makefiles. This
10410 makes it possible to install openssl libraries in locations which
10411 have names other than "lib", for example "/usr/lib64" which some
10414 *Steve Henson, based on patch from Jeremy Utley*
10416 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10417 X690 8.9.12 and can produce some misleading textual output of OIDs.
10419 *Steve Henson, reported by Dan Kaminsky*
10421 * Delete MD2 from algorithm tables. This follows the recommendation in
10422 several standards that it is not used in new applications due to
10423 several cryptographic weaknesses. For binary compatibility reasons
10424 the MD2 API is still compiled in by default.
10428 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10433 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10434 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10437 *Guenter <lists@gknw.net>*
10439 * Fix the server certificate chain building code to use X509_verify_cert(),
10440 it used to have an ad-hoc builder which was unable to cope with anything
10441 other than a simple chain.
10443 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10445 * Don't check self signed certificate signatures in X509_verify_cert()
10446 by default (a flag can override this): it just wastes time without
10447 adding any security. As a useful side effect self signed root CAs
10448 with non-FIPS digests are now usable in FIPS mode.
10452 * In dtls1_process_out_of_seq_message() the check if the current message
10453 is already buffered was missing. For every new message was memory
10454 allocated, allowing an attacker to perform an denial of service attack
10455 with sending out of seq handshake messages until there is no memory
10456 left. Additionally every future message was buffered, even if the
10457 sequence number made no sense and would be part of another handshake.
10458 So only messages with sequence numbers less than 10 in advance will be
10459 buffered. ([CVE-2009-1378])
10461 *Robin Seggelmann, discovered by Daniel Mentz*
10463 * Records are buffered if they arrive with a future epoch to be
10464 processed after finishing the corresponding handshake. There is
10465 currently no limitation to this buffer allowing an attacker to perform
10466 a DOS attack with sending records with future epochs until there is no
10467 memory left. This patch adds the pqueue_size() function to determine
10468 the size of a buffer and limits the record buffer to 100 entries.
10471 *Robin Seggelmann, discovered by Daniel Mentz*
10473 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10474 parent structure is freed. ([CVE-2009-1379])
10478 * Handle non-blocking I/O properly in SSL_shutdown() call.
10480 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10482 * Add `2.5.4.*` OIDs
10484 *Ilya O. <vrghost@gmail.com>*
10486 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10488 * Disable renegotiation completely - this fixes a severe security
10489 problem ([CVE-2009-3555]) at the cost of breaking all
10490 renegotiation. Renegotiation can be re-enabled by setting
10491 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10492 run-time. This is really not recommended unless you know what
10497 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10499 * Don't set val to NULL when freeing up structures, it is freed up by
10500 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10501 zeroing past the valid field. ([CVE-2009-0789])
10503 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10505 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10506 checked correctly. This would allow some invalid signed attributes to
10507 appear to verify correctly. ([CVE-2009-0591])
10509 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10511 * Reject UniversalString and BMPString types with invalid lengths. This
10512 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10513 a legal length. ([CVE-2009-0590])
10517 * Set S/MIME signing as the default purpose rather than setting it
10518 unconditionally. This allows applications to override it at the store
10523 * Permit restricted recursion of ASN1 strings. This is needed in practice
10524 to handle some structures.
10528 * Improve efficiency of mem_gets: don't search whole buffer each time
10531 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10533 * New -hex option for openssl rand.
10537 * Print out UTF8String and NumericString when parsing ASN1.
10541 * Support NumericString type for name components.
10545 * Allow CC in the environment to override the automatically chosen
10546 compiler. Note that nothing is done to ensure flags work with the
10551 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10553 * Properly check EVP_VerifyFinal() and similar return values
10556 *Ben Laurie, Bodo Moeller, Google Security Team*
10558 * Enable TLS extensions by default.
10562 * Allow the CHIL engine to be loaded, whether the application is
10563 multithreaded or not. (This does not release the developer from the
10564 obligation to set up the dynamic locking callbacks.)
10566 *Sander Temme <sander@temme.net>*
10568 * Use correct exit code if there is an error in dgst command.
10570 *Steve Henson; problem pointed out by Roland Dirlewanger*
10572 * Tweak Configure so that you need to say "experimental-jpake" to enable
10573 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10577 * Add experimental JPAKE support, including demo authentication in
10578 s_client and s_server.
10582 * Set the comparison function in v3_addr_canonize().
10584 *Rob Austein <sra@hactrn.net>*
10586 * Add support for XMPP STARTTLS in s_client.
10588 *Philip Paeps <philip@freebsd.org>*
10590 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10591 to ensure that even with this option, only ciphersuites in the
10592 server's preference list will be accepted. (Note that the option
10593 applies only when resuming a session, so the earlier behavior was
10594 just about the algorithm choice for symmetric cryptography.)
10598 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10600 * Fix NULL pointer dereference if a DTLS server received
10601 ChangeCipherSpec as first record ([CVE-2009-1386]).
10605 * Fix a state transition in s3_srvr.c and d1_srvr.c
10606 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10608 *Nagendra Modadugu*
10610 * The fix in 0.9.8c that supposedly got rid of unsafe
10611 double-checked locking was incomplete for RSA blinding,
10612 addressing just one layer of what turns out to have been
10613 doubly unsafe triple-checked locking.
10615 So now fix this for real by retiring the MONT_HELPER macro
10616 in crypto/rsa/rsa_eay.c.
10618 *Bodo Moeller; problem pointed out by Marius Schilder*
10620 * Various precautionary measures:
10622 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10624 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10625 (NB: This would require knowledge of the secret session ticket key
10626 to exploit, in which case you'd be SOL either way.)
10628 - Change bn_nist.c so that it will properly handle input BIGNUMs
10629 outside the expected range.
10631 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10634 *Neel Mehta, Bodo Moeller*
10636 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10637 the load fails. Useful for distros.
10639 *Ben Laurie and the FreeBSD team*
10641 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10645 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10649 * Expand ENGINE to support engine supplied SSL client certificate functions.
10651 This work was sponsored by Logica.
10655 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10656 keystores. Support for SSL/TLS client authentication too.
10657 Not compiled unless enable-capieng specified to Configure.
10659 This work was sponsored by Logica.
10663 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10664 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10665 attribute creation routines such as certificate requests and PKCS#12
10670 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10672 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10673 handshake which could lead to a client crash as found using the
10674 Codenomicon TLS test suite ([CVE-2008-1672])
10676 *Steve Henson, Mark Cox*
10678 * Fix double free in TLS server name extensions which could lead to
10679 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10683 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10685 Clear the error queue to ensure that error entries left from
10686 older function calls do not interfere with the correct operation.
10688 *Lutz Jaenicke, Erik de Castro Lopo*
10690 * Remove root CA certificates of commercial CAs:
10692 The OpenSSL project does not recommend any specific CA and does not
10693 have any policy with respect to including or excluding any CA.
10694 Therefore, it does not make any sense to ship an arbitrary selection
10695 of root CA certificates with the OpenSSL software.
10699 * RSA OAEP patches to fix two separate invalid memory reads.
10700 The first one involves inputs when 'lzero' is greater than
10701 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10702 before the beginning of from). The second one involves inputs where
10703 the 'db' section contains nothing but zeroes (there is a one-byte
10704 invalid read after the end of 'db').
10706 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10708 * Partial backport from 0.9.9-dev:
10710 Introduce bn_mul_mont (dedicated Montgomery multiplication
10711 procedure) as a candidate for BIGNUM assembler implementation.
10712 While 0.9.9-dev uses assembler for various architectures, only
10713 x86_64 is available by default here in the 0.9.8 branch, and
10714 32-bit x86 is available through a compile-time setting.
10716 To try the 32-bit x86 assembler implementation, use Configure
10717 option "enable-montasm" (which exists only for this backport).
10719 As "enable-montasm" for 32-bit x86 disclaims code stability
10720 anyway, in this constellation we activate additional code
10721 backported from 0.9.9-dev for further performance improvements,
10722 namely BN_from_montgomery_word. (To enable this otherwise,
10723 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10725 *Andy Polyakov (backport partially by Bodo Moeller)*
10727 * Add TLS session ticket callback. This allows an application to set
10728 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10729 values. This is useful for key rollover for example where several key
10730 sets may exist with different names.
10734 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10735 This was broken until now in 0.9.8 releases, such that the only way
10736 a registered ENGINE could be used (assuming it initialises
10737 successfully on the host) was to explicitly set it as the default
10738 for the relevant algorithms. This is in contradiction with 0.9.7
10739 behaviour and the documentation. With this fix, when an ENGINE is
10740 registered into a given algorithm's table of implementations, the
10741 'uptodate' flag is reset so that auto-discovery will be used next
10742 time a new context for that algorithm attempts to select an
10745 *Ian Lister (tweaked by Geoff Thorpe)*
10747 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10748 implementation in the following ways:
10750 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10753 Lack of BER streaming support means one pass streaming processing is
10754 only supported if data is detached: setting the streaming flag is
10755 ignored for embedded content.
10757 CMS support is disabled by default and must be explicitly enabled
10758 with the enable-cms configuration option.
10762 * Update the GMP engine glue to do direct copies between BIGNUM and
10763 mpz_t when openssl and GMP use the same limb size. Otherwise the
10764 existing "conversion via a text string export" trick is still used.
10766 *Paul Sheer <paulsheer@gmail.com>*
10768 * Zlib compression BIO. This is a filter BIO which compressed and
10769 uncompresses any data passed through it.
10773 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10774 RFC3394 compatible AES key wrapping.
10778 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10779 sets string data without copying. X509_ALGOR_set0() and
10780 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10781 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10782 from an X509_ATTRIBUTE structure optionally checking it occurs only
10783 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10788 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10789 to get the expected BN_FLG_CONSTTIME behavior.
10791 *Bodo Moeller (Google)*
10795 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10796 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10797 - added some more tests to do_tests.pl
10798 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10799 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10800 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10801 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10802 - various changes to netware.pl to enable gcc-cross builds on Win32
10804 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10805 - various changes to fix missing prototype warnings
10806 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10807 - added AES, WHIRLPOOL and CPUID assembler code to build files
10808 - added missing AES assembler make rules to mk1mf.pl
10809 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10811 *Guenter Knauf <eflash@gmx.net>*
10813 * Implement certificate status request TLS extension defined in RFC3546.
10814 A client can set the appropriate parameters and receive the encoded
10815 OCSP response via a callback. A server can query the supplied parameters
10816 and set the encoded OCSP response in the callback. Add simplified examples
10817 to s_client and s_server.
10821 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10823 * Fix various bugs:
10824 + Binary incompatibility of ssl_ctx_st structure
10825 + DTLS interoperation with non-compliant servers
10826 + Don't call get_session_cb() without proposed session
10827 + Fix ia64 assembler code
10829 *Andy Polyakov, Steve Henson*
10831 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10833 * DTLS Handshake overhaul. There were longstanding issues with
10834 OpenSSL DTLS implementation, which were making it impossible for
10835 RFC 4347 compliant client to communicate with OpenSSL server.
10836 Unfortunately just fixing these incompatibilities would "cut off"
10837 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10838 server keeps tolerating non RFC compliant syntax. The opposite is
10839 not true, 0.9.8f client can not communicate with earlier server.
10840 This update even addresses CVE-2007-4995.
10844 * Changes to avoid need for function casts in OpenSSL: some compilers
10845 (gcc 4.2 and later) reject their use.
10846 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10849 * Add RFC4507 support to OpenSSL. This includes the corrections in
10850 RFC4507bis. The encrypted ticket format is an encrypted encoded
10851 SSL_SESSION structure, that way new session features are automatically
10854 If a client application caches session in an SSL_SESSION structure
10855 support is transparent because tickets are now stored in the encoded
10858 The SSL_CTX structure automatically generates keys for ticket
10859 protection in servers so again support should be possible
10860 with no application modification.
10862 If a client or server wishes to disable RFC4507 support then the option
10863 SSL_OP_NO_TICKET can be set.
10865 Add a TLS extension debugging callback to allow the contents of any client
10866 or server extensions to be examined.
10868 This work was sponsored by Google.
10872 * Add initial support for TLS extensions, specifically for the server_name
10873 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10874 have new members for a hostname. The SSL data structure has an
10875 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10876 stored in that context to allow for session resumption, even after the
10877 SSL has been switched to a new SSL_CTX in reaction to a client's
10878 server_name extension.
10880 New functions (subject to change):
10882 SSL_get_servername()
10883 SSL_get_servername_type()
10886 New CTRL codes and macros (subject to change):
10888 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10889 - SSL_CTX_set_tlsext_servername_callback()
10890 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10891 - SSL_CTX_set_tlsext_servername_arg()
10892 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10894 openssl s_client has a new '-servername ...' option.
10896 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10897 '-key2 ...', '-servername_fatal' (subject to change). This allows
10898 testing the HostName extension for a specific single hostname ('-cert'
10899 and '-key' remain fallbacks for handshakes without HostName
10900 negotiation). If the unrecognized_name alert has to be sent, this by
10901 default is a warning; it becomes fatal with the '-servername_fatal'
10904 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10906 * Add AES and SSE2 assembly language support to VC++ build.
10910 * Mitigate attack on final subtraction in Montgomery reduction.
10914 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10915 (which previously caused an internal error).
10919 * Squeeze another 10% out of IGE mode when in != out.
10923 * AES IGE mode speedup.
10925 *Dean Gaudet (Google)*
10927 * Add the Korean symmetric 128-bit cipher SEED (see
10928 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10929 add SEED ciphersuites from RFC 4162:
10931 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10932 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10933 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10934 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10936 To minimize changes between patchlevels in the OpenSSL 0.9.8
10937 series, SEED remains excluded from compilation unless OpenSSL
10938 is configured with 'enable-seed'.
10940 *KISA, Bodo Moeller*
10942 * Mitigate branch prediction attacks, which can be practical if a
10943 single processor is shared, allowing a spy process to extract
10944 information. For detailed background information, see
10945 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10946 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10947 and Necessary Software Countermeasures"). The core of the change
10948 are new versions BN_div_no_branch() and
10949 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10950 respectively, which are slower, but avoid the security-relevant
10951 conditional branches. These are automatically called by BN_div()
10952 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10953 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10954 remove a conditional branch.
10956 BN_FLG_CONSTTIME is the new name for the previous
10957 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10958 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10959 in the exponent causes BN_mod_exp_mont() to use the alternative
10960 implementation in BN_mod_exp_mont_consttime().) The old name
10961 remains as a deprecated alias.
10963 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10964 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10965 constant-time implementations for more than just exponentiation.
10966 Here too the old name is kept as a deprecated alias.
10968 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10969 the BN_BLINDING structure gets an independent copy of the
10970 modulus. This means that the previous `BIGNUM *m` argument to
10971 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10972 essentially becomes `const BIGNUM *m`, although we can't actually
10973 change this in the header file before 0.9.9. It allows
10974 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10975 enable BN_FLG_CONSTTIME.
10977 *Matthew D Wood (Intel Corp)*
10979 * In the SSL/TLS server implementation, be strict about session ID
10980 context matching (which matters if an application uses a single
10981 external cache for different purposes). Previously,
10982 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10983 set. This did ensure strict client verification, but meant that,
10984 with applications using a single external cache for quite
10985 different requirements, clients could circumvent ciphersuite
10986 restrictions for a given session ID context by starting a session
10987 in a different context.
10991 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10992 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10993 authentication-only ciphersuites.
10997 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10998 not complete and could lead to a possible single byte overflow
10999 ([CVE-2007-5135]) [Ben Laurie]
11001 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
11003 * Since AES128 and AES256 (and similarly Camellia128 and
11004 Camellia256) share a single mask bit in the logic of
11005 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11006 kludge to work properly if AES128 is available and AES256 isn't
11007 (or if Camellia128 is available and Camellia256 isn't).
11011 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
11012 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
11013 When a point or a seed is encoded in a BIT STRING, we need to
11014 prevent the removal of trailing zero bits to get the proper DER
11015 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
11016 of a NamedBitList, for which trailing 0 bits need to be removed.)
11020 * Have SSL/TLS server implementation tolerate "mismatched" record
11021 protocol version while receiving ClientHello even if the
11022 ClientHello is fragmented. (The server can't insist on the
11023 particular protocol version it has chosen before the ServerHello
11024 message has informed the client about his choice.)
11028 * Add RFC 3779 support.
11030 *Rob Austein for ARIN, Ben Laurie*
11032 * Load error codes if they are not already present instead of using a
11033 static variable. This allows them to be cleanly unloaded and reloaded.
11034 Improve header file function name parsing.
11038 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
11039 or CAPABILITY handshake as required by RFCs.
11041 *Goetz Babin-Ebell*
11043 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
11045 * Introduce limits to prevent malicious keys being able to
11046 cause a denial of service. ([CVE-2006-2940])
11048 *Steve Henson, Bodo Moeller*
11050 * Fix ASN.1 parsing of certain invalid structures that can result
11051 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11053 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11054 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11056 * Fix SSL client code which could crash if connecting to a
11057 malicious SSLv2 server. ([CVE-2006-4343])
11059 *Tavis Ormandy and Will Drewry, Google Security Team*
11061 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
11062 match only those. Before that, "AES256-SHA" would be interpreted
11063 as a pattern and match "AES128-SHA" too (since AES128-SHA got
11064 the same strength classification in 0.9.7h) as we currently only
11065 have a single AES bit in the ciphersuite description bitmap.
11066 That change, however, also applied to ciphersuite strings such as
11067 "RC4-MD5" that intentionally matched multiple ciphersuites --
11068 namely, SSL 2.0 ciphersuites in addition to the more common ones
11069 from SSL 3.0/TLS 1.0.
11071 So we change the selection algorithm again: Naming an explicit
11072 ciphersuite selects this one ciphersuite, and any other similar
11073 ciphersuite (same bitmap) from *other* protocol versions.
11074 Thus, "RC4-MD5" again will properly select both the SSL 2.0
11075 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
11077 Since SSL 2.0 does not have any ciphersuites for which the
11078 128/256 bit distinction would be relevant, this works for now.
11079 The proper fix will be to use different bits for AES128 and
11080 AES256, which would have avoided the problems from the beginning;
11081 however, bits are scarce, so we can only do this in a new release
11082 (not just a patchlevel) when we can change the SSL_CIPHER
11083 definition to split the single 'unsigned long mask' bitmap into
11084 multiple values to extend the available space.
11088 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
11090 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11091 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11093 * Add AES IGE and biIGE modes.
11097 * Change the Unix randomness entropy gathering to use poll() when
11098 possible instead of select(), since the latter has some
11099 undesirable limitations.
11101 *Darryl Miles via Richard Levitte and Bodo Moeller*
11103 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
11104 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
11105 cannot be implicitly activated as part of, e.g., the "AES" alias.
11106 However, please upgrade to OpenSSL 0.9.9[-dev] for
11107 non-experimental use of the ECC ciphersuites to get TLS extension
11108 support, which is required for curve and point format negotiation
11109 to avoid potential handshake problems.
11113 * Disable rogue ciphersuites:
11115 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11116 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11117 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11119 The latter two were purportedly from
11120 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11123 Also deactivate the remaining ciphersuites from
11124 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11125 unofficial, and the ID has long expired.
11129 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11130 dual-core machines) and other potential thread-safety issues.
11134 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
11135 versions), which is now available for royalty-free use
11136 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
11137 Also, add Camellia TLS ciphersuites from RFC 4132.
11139 To minimize changes between patchlevels in the OpenSSL 0.9.8
11140 series, Camellia remains excluded from compilation unless OpenSSL
11141 is configured with 'enable-camellia'.
11145 * Disable the padding bug check when compression is in use. The padding
11146 bug check assumes the first packet is of even length, this is not
11147 necessarily true if compression is enabled and can result in false
11148 positives causing handshake failure. The actual bug test is ancient
11149 code so it is hoped that implementations will either have fixed it by
11150 now or any which still have the bug do not support compression.
11154 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
11156 * When applying a cipher rule check to see if string match is an explicit
11157 cipher suite and only match that one cipher suite if it is.
11161 * Link in manifests for VC++ if needed.
11163 *Austin Ziegler <halostatue@gmail.com>*
11165 * Update support for ECC-based TLS ciphersuites according to
11166 draft-ietf-tls-ecc-12.txt with proposed changes (but without
11167 TLS extensions, which are supported starting with the 0.9.9
11168 branch, not in the OpenSSL 0.9.8 branch).
11172 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
11173 opaque EVP_CIPHER_CTX handling.
11177 * Fixes and enhancements to zlib compression code. We now only use
11178 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
11179 to conform with the standards mentioned here:
11180 <http://www.zlib.net/DLL_FAQ.txt>
11181 Static zlib linking now works on Windows and the new --with-zlib-include
11182 --with-zlib-lib options to Configure can be used to supply the location
11183 of the headers and library. Gracefully handle case where zlib library
11188 * Several fixes and enhancements to the OID generation code. The old code
11189 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11190 handle numbers larger than ULONG_MAX, truncated printing and had a
11191 non standard OBJ_obj2txt() behaviour.
11195 * Add support for building of engines under engine/ as shared libraries
11196 under VC++ build system.
11200 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11201 Hopefully, we will not see any false combination of paths any more.
11205 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11207 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11208 (part of SSL_OP_ALL). This option used to disable the
11209 countermeasure against man-in-the-middle protocol-version
11210 rollback in the SSL 2.0 server implementation, which is a bad
11211 idea. ([CVE-2005-2969])
11213 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11214 for Information Security, National Institute of Advanced Industrial
11215 Science and Technology [AIST], Japan)*
11217 * Add two function to clear and return the verify parameter flags.
11221 * Keep cipherlists sorted in the source instead of sorting them at
11222 runtime, thus removing the need for a lock.
11226 * Avoid some small subgroup attacks in Diffie-Hellman.
11228 *Nick Mathewson and Ben Laurie*
11230 * Add functions for well-known primes.
11234 * Extended Windows CE support.
11236 *Satoshi Nakamura and Andy Polyakov*
11238 * Initialize SSL_METHOD structures at compile time instead of during
11239 runtime, thus removing the need for a lock.
11243 * Make PKCS7_decrypt() work even if no certificate is supplied by
11244 attempting to decrypt each encrypted key in turn. Add support to
11249 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11251 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11254 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11258 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11259 key into the same file any more.
11263 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11267 * Add -utf8 command line and config file option to 'ca'.
11269 *Stefan <stf@udoma.org*
11271 * Removed the macro des_crypt(), as it seems to conflict with some
11272 libraries. Use DES_crypt().
11276 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11277 involves renaming the source and generated shared-libs for
11278 both. The engines will accept the corrected or legacy ids
11279 ('ncipher' and '4758_cca' respectively) when binding. NB,
11280 this only applies when building 'shared'.
11282 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11284 * Add attribute functions to EVP_PKEY structure. Modify
11285 PKCS12_create() to recognize a CSP name attribute and
11286 use it. Make -CSP option work again in pkcs12 utility.
11290 * Add new functionality to the bn blinding code:
11291 - automatic re-creation of the BN_BLINDING parameters after
11292 a fixed number of uses (currently 32)
11293 - add new function for parameter creation
11294 - introduce flags to control the update behaviour of the
11295 BN_BLINDING parameters
11296 - hide BN_BLINDING structure
11297 Add a second BN_BLINDING slot to the RSA structure to improve
11298 performance when a single RSA object is shared among several
11303 * Add support for DTLS.
11305 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11307 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11308 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11312 * Remove buggy and incomplete DH cert support from
11313 ssl/ssl_rsa.c and ssl/s3_both.c
11317 * Use SHA-1 instead of MD5 as the default digest algorithm for
11318 the `apps/openssl` commands.
11322 * Compile clean with "-Wall -Wmissing-prototypes
11323 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11324 DEBUG_SAFESTACK must also be set.
11328 * Change ./Configure so that certain algorithms can be disabled by default.
11329 The new counterpiece to "no-xxx" is "enable-xxx".
11331 The patented RC5 and MDC2 algorithms will now be disabled unless
11332 "enable-rc5" and "enable-mdc2", respectively, are specified.
11334 (IDEA remains enabled despite being patented. This is because IDEA
11335 is frequently required for interoperability, and there is no license
11336 fee for non-commercial use. As before, "no-idea" can be used to
11337 avoid this algorithm.)
11341 * Add processing of proxy certificates (see RFC 3820). This work was
11342 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11343 EGEE (Enabling Grids for E-science in Europe).
11347 * RC4 performance overhaul on modern architectures/implementations, such
11348 as Intel P4, IA-64 and AMD64.
11352 * New utility extract-section.pl. This can be used specify an alternative
11353 section number in a pod file instead of having to treat each file as
11354 a separate case in Makefile. This can be done by adding two lines to the
11357 =for comment openssl_section:XXX
11359 The blank line is mandatory.
11363 * New arguments -certform, -keyform and -pass for s_client and s_server
11364 to allow alternative format key and certificate files and passphrase
11369 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11370 update associated structures and add various utility functions.
11372 Add new policy related verify parameters, include policy checking in
11373 standard verify code. Enhance 'smime' application with extra parameters
11374 to support policy checking and print out.
11378 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11379 Nehemiah processors. These extensions support AES encryption in hardware
11380 as well as RNG (though RNG support is currently disabled).
11382 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11384 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11388 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11390 *Andy Polyakov and a number of other people*
11392 * Improved PowerPC platform support. Most notably BIGNUM assembler
11393 implementation contributed by IBM.
11395 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11397 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11398 exponent rather than 'unsigned long'. There is a corresponding change to
11399 the new 'rsa_keygen' element of the RSA_METHOD structure.
11401 *Jelte Jansen, Geoff Thorpe*
11403 * Functionality for creating the initial serial number file is now
11404 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11406 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11407 number file to 1, which is bound to cause problems. To avoid
11408 the problems while respecting compatibility between different 0.9.7
11409 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11410 CA.pl for serial number initialization. With the new release 0.9.8,
11411 we can fix the problem directly in the 'ca' utility.)
11415 * Reduced header interdependencies by declaring more opaque objects in
11416 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11417 give fewer recursive includes, which could break lazy source code - so
11418 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11419 developers should define this symbol when building and using openssl to
11420 ensure they track the recommended behaviour, interfaces, [etc], but
11421 backwards-compatible behaviour prevails when this isn't defined.
11425 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11429 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11430 This will generate a random key of the appropriate length based on the
11431 cipher context. The EVP_CIPHER can provide its own random key generation
11432 routine to support keys of a specific form. This is used in the des and
11433 3des routines to generate a key of the correct parity. Update S/MIME
11434 code to use new functions and hence generate correct parity DES keys.
11435 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11436 valid (weak or incorrect parity).
11440 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11441 as looking them up. This is useful when the verified structure may contain
11442 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11443 present unless the new PKCS7_NO_CRL flag is asserted.
11447 * Extend ASN1 oid configuration module. It now additionally accepts the
11450 shortName = some long name, 1.2.3.4
11454 * Reimplemented the BN_CTX implementation. There is now no more static
11455 limitation on the number of variables it can handle nor the depth of the
11456 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11457 information can now expand as required, and rather than having a single
11458 static array of bignums, BN_CTX now uses a linked-list of such arrays
11459 allowing it to expand on demand whilst maintaining the usefulness of
11460 BN_CTX's "bundling".
11464 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11465 to allow all RSA operations to function using a single BN_CTX.
11469 * Preliminary support for certificate policy evaluation and checking. This
11470 is initially intended to pass the tests outlined in "Conformance Testing
11471 of Relying Party Client Certificate Path Processing Logic" v1.07.
11475 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11476 remained unused and not that useful. A variety of other little bignum
11477 tweaks and fixes have also been made continuing on from the audit (see
11482 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11483 associated ASN1, EVP and SSL functions and old ASN1 macros.
11487 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11488 and this should never fail. So the return value from the use of
11489 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11490 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11494 * BN_CTX_get() should return zero-valued bignums, providing the same
11495 initialised value as BN_new().
11497 *Geoff Thorpe, suggested by Ulf Möller*
11499 * Support for inhibitAnyPolicy certificate extension.
11503 * An audit of the BIGNUM code is underway, for which debugging code is
11504 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11505 is considered valid when processing BIGNUMs, and causes execution to
11506 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11507 further steps are taken to deliberately pollute unused data in BIGNUM
11508 structures to try and expose faulty code further on. For now, openssl will
11509 (in its default mode of operation) continue to tolerate the inconsistent
11510 forms that it has tolerated in the past, but authors and packagers should
11511 consider trying openssl and their own applications when compiled with
11512 these debugging symbols defined. It will help highlight potential bugs in
11513 their own code, and will improve the test coverage for OpenSSL itself. At
11514 some point, these tighter rules will become openssl's default to improve
11515 maintainability, though the assert()s and other overheads will remain only
11516 in debugging configurations. See bn.h for more details.
11518 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11520 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11521 that can only be obtained through BN_CTX_new() (which implicitly
11522 initialises it). The presence of this function only made it possible
11523 to overwrite an existing structure (and cause memory leaks).
11527 * Because of the callback-based approach for implementing LHASH as a
11528 template type, lh_insert() adds opaque objects to hash-tables and
11529 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11530 to clean up those corresponding objects before destroying the hash table
11531 (and losing the object pointers). So some over-zealous constifications in
11532 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11533 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11534 prototyped to have "const" restrictions on the object pointers they are
11535 given (and so aren't required to cast them away any more).
11539 * The tmdiff.h API was so ugly and minimal that our own timing utility
11540 (speed) prefers to use its own implementation. The two implementations
11541 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11542 its object type properly exposed (MS_TM) instead of casting to/from
11543 `char *`. This may still change yet if someone realises MS_TM and
11545 aren't necessarily the greatest nomenclatures - but this is what was used
11546 internally to the implementation so I've used that for now.
11550 * Ensure that deprecated functions do not get compiled when
11551 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11552 the self-tests were still using deprecated key-generation functions so
11553 these have been updated also.
11557 * Reorganise PKCS#7 code to separate the digest location functionality
11558 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11559 New function PKCS7_set_digest() to set the digest type for PKCS#7
11560 digestedData type. Add additional code to correctly generate the
11561 digestedData type and add support for this type in PKCS7 initialization
11566 * New function PKCS7_set0_type_other() this initializes a PKCS7
11567 structure of type "other".
11571 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11572 sure the loop does correctly stop and breaking ("division by zero")
11573 modulus operations are not performed. The (pre-generated) prime
11574 table crypto/bn/bn_prime.h was already correct, but it could not be
11575 re-generated on some platforms because of the "division by zero"
11576 situation in the script.
11578 *Ralf S. Engelschall*
11580 * Update support for ECC-based TLS ciphersuites according to
11581 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11582 SHA-1 now is only used for "small" curves (where the
11583 representation of a field element takes up to 24 bytes); for
11584 larger curves, the field element resulting from ECDH is directly
11585 used as premaster secret.
11587 *Douglas Stebila (Sun Microsystems Laboratories)*
11589 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11590 curve secp160r1 to the tests.
11592 *Douglas Stebila (Sun Microsystems Laboratories)*
11594 * Add the possibility to load symbols globally with DSO.
11596 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11598 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11599 control of the error stack.
11603 * Add support for STORE in ENGINE.
11607 * Add the STORE type. The intention is to provide a common interface
11608 to certificate and key stores, be they simple file-based stores, or
11609 HSM-type store, or LDAP stores, or...
11610 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11614 * Add a generic structure called OPENSSL_ITEM. This can be used to
11615 pass a list of arguments to any function as well as provide a way
11616 for a function to pass data back to the caller.
11620 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11621 works like BUF_strdup() but can be used to duplicate a portion of
11622 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11627 * Add the function sk_find_ex() which works like sk_find(), but will
11628 return an index to an element even if an exact match couldn't be
11629 found. The index is guaranteed to point at the element where the
11630 searched-for key would be inserted to preserve sorting order.
11634 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11635 takes an extra flags argument for optional functionality. Currently,
11636 the following flags are defined:
11638 OBJ_BSEARCH_VALUE_ON_NOMATCH
11639 This one gets OBJ_bsearch_ex() to return a pointer to the first
11640 element where the comparing function returns a negative or zero
11643 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11644 This one gets OBJ_bsearch_ex() to return a pointer to the first
11645 element where the comparing function returns zero. This is useful
11646 if there are more than one element where the comparing function
11651 * Make it possible to create self-signed certificates with 'openssl ca'
11652 in such a way that the self-signed certificate becomes part of the
11653 CA database and uses the same mechanisms for serial number generation
11654 as all other certificate signing. The new flag '-selfsign' enables
11655 this functionality. Adapt CA.sh and CA.pl.in.
11659 * Add functionality to check the public key of a certificate request
11660 against a given private. This is useful to check that a certificate
11661 request can be signed by that key (self-signing).
11665 * Make it possible to have multiple active certificates with the same
11666 subject in the CA index file. This is done only if the keyword
11667 'unique_subject' is set to 'no' in the main CA section (default
11668 if 'CA_default') of the configuration file. The value is saved
11669 with the database itself in a separate index attribute file,
11670 named like the index file with '.attr' appended to the name.
11674 * Generate multi-valued AVAs using '+' notation in config files for
11679 * Support for nameConstraints certificate extension.
11683 * Support for policyConstraints certificate extension.
11687 * Support for policyMappings certificate extension.
11691 * Make sure the default DSA_METHOD implementation only uses its
11692 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11693 and change its own handlers to be NULL so as to remove unnecessary
11694 indirection. This lets alternative implementations fallback to the
11695 default implementation more easily.
11699 * Support for directoryName in GeneralName related extensions
11704 * Make it possible to link applications using Makefile.shared.
11705 Make that possible even when linking against static libraries!
11709 * Support for single pass processing for S/MIME signing. This now
11710 means that S/MIME signing can be done from a pipe, in addition
11711 cleartext signing (multipart/signed type) is effectively streaming
11712 and the signed data does not need to be all held in memory.
11714 This is done with a new flag PKCS7_STREAM. When this flag is set
11715 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11716 is done after the data is output (and digests calculated) in
11717 SMIME_write_PKCS7().
11721 * Add full support for -rpath/-R, both in shared libraries and
11722 applications, at least on the platforms where it's known how
11727 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11728 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11729 will now compute a table of multiples of the generator that
11730 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11731 faster (notably in the case of a single point multiplication,
11732 scalar * generator).
11734 *Nils Larsch, Bodo Moeller*
11736 * IPv6 support for certificate extensions. The various extensions
11737 which use the IP:a.b.c.d can now take IPv6 addresses using the
11738 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11743 * Added an ENGINE that implements RSA by performing private key
11744 exponentiations with the GMP library. The conversions to and from
11745 GMP's mpz_t format aren't optimised nor are any montgomery forms
11746 cached, and on x86 it appears OpenSSL's own performance has caught up.
11747 However there are likely to be other architectures where GMP could
11748 provide a boost. This ENGINE is not built in by default, but it can be
11749 specified at Configure time and should be accompanied by the necessary
11750 linker additions, eg;
11751 ./config -DOPENSSL_USE_GMP -lgmp
11755 * "openssl engine" will not display ENGINE/DSO load failure errors when
11756 testing availability of engines with "-t" - the old behaviour is
11757 produced by increasing the feature's verbosity with "-tt".
11761 * ECDSA routines: under certain error conditions uninitialized BN objects
11762 could be freed. Solution: make sure initialization is performed early
11763 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11768 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11769 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11770 software implementations. For DSA and DH, parameter generation can
11771 also be overridden by providing the appropriate method callbacks.
11775 * Change the "progress" mechanism used in key-generation and
11776 primality testing to functions that take a new BN_GENCB pointer in
11777 place of callback/argument pairs. The new API functions have `_ex`
11778 postfixes and the older functions are reimplemented as wrappers for
11779 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11780 declarations of the old functions to help (graceful) attempts to
11781 migrate to the new functions. Also, the new key-generation API
11782 functions operate on a caller-supplied key-structure and return
11783 success/failure rather than returning a key or NULL - this is to
11784 help make "keygen" another member function of RSA_METHOD etc.
11786 Example for using the new callback interface:
11788 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11789 void *my_arg = ...;
11792 BN_GENCB_set(&my_cb, my_callback, my_arg);
11794 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11795 /* For the meaning of a, b in calls to my_callback(), see the
11796 * documentation of the function that calls the callback.
11797 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11798 * my_callback should return 1 if it wants BN_is_prime_ex()
11799 * to continue, or 0 to stop.
11804 * Change the ZLIB compression method to be stateful, and make it
11805 available to TLS with the number defined in
11806 draft-ietf-tls-compression-04.txt.
11810 * Add the ASN.1 structures and functions for CertificatePair, which
11811 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11813 CertificatePair ::= SEQUENCE {
11814 forward [0] Certificate OPTIONAL,
11815 reverse [1] Certificate OPTIONAL,
11816 -- at least one of the pair shall be present -- }
11818 Also implement the PEM functions to read and write certificate
11819 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11821 This needed to be defined, mostly for the sake of the LDAP
11822 attribute crossCertificatePair, but may prove useful elsewhere as
11827 * Make it possible to inhibit symlinking of shared libraries in
11828 Makefile.shared, for Cygwin's sake.
11832 * Extend the BIGNUM API by creating a function
11833 void BN_set_negative(BIGNUM *a, int neg);
11834 and a macro that behave like
11835 int BN_is_negative(const BIGNUM *a);
11837 to avoid the need to access 'a->neg' directly in applications.
11841 * Implement fast modular reduction for pseudo-Mersenne primes
11842 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11843 EC_GROUP_new_curve_GFp() will now automatically use this
11846 *Nils Larsch <nla@trustcenter.de>*
11848 * Add new lock type (CRYPTO_LOCK_BN).
11852 * Change the ENGINE framework to automatically load engines
11853 dynamically from specific directories unless they could be
11854 found to already be built in or loaded. Move all the
11855 current engines except for the cryptodev one to a new
11856 directory engines/.
11857 The engines in engines/ are built as shared libraries if
11858 the "shared" options was given to ./Configure or ./config.
11859 Otherwise, they are inserted in libcrypto.a.
11860 /usr/local/ssl/engines is the default directory for dynamic
11861 engines, but that can be overridden at configure time through
11862 the usual use of --prefix and/or --openssldir, and at run
11863 time with the environment variable OPENSSL_ENGINES.
11865 *Geoff Thorpe and Richard Levitte*
11867 * Add Makefile.shared, a helper makefile to build shared
11868 libraries. Adapt Makefile.org.
11872 * Add version info to Win32 DLLs.
11874 *Peter 'Luna' Runestig" <peter@runestig.com>*
11876 * Add new 'medium level' PKCS#12 API. Certificates and keys
11877 can be added using this API to created arbitrary PKCS#12
11878 files while avoiding the low-level API.
11880 New options to PKCS12_create(), key or cert can be NULL and
11881 will then be omitted from the output file. The encryption
11882 algorithm NIDs can be set to -1 for no encryption, the mac
11883 iteration count can be set to 0 to omit the mac.
11885 Enhance pkcs12 utility by making the -nokeys and -nocerts
11886 options work when creating a PKCS#12 file. New option -nomac
11887 to omit the mac, NONE can be set for an encryption algorithm.
11888 New code is modified to use the enhanced PKCS12_create()
11889 instead of the low-level API.
11893 * Extend ASN1 encoder to support indefinite length constructed
11894 encoding. This can output sequences tags and octet strings in
11895 this form. Modify pk7_asn1.c to support indefinite length
11896 encoding. This is experimental and needs additional code to
11897 be useful, such as an ASN1 bio and some enhanced streaming
11900 Extend template encode functionality so that tagging is passed
11901 down to the template encoder.
11905 * Let 'openssl req' fail if an argument to '-newkey' is not
11906 recognized instead of using RSA as a default.
11910 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11911 As these are not official, they are not included in "ALL";
11912 the "ECCdraft" ciphersuite group alias can be used to select them.
11914 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11916 * Add ECDH engine support.
11918 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11920 * Add ECDH in new directory crypto/ecdh/.
11922 *Douglas Stebila (Sun Microsystems Laboratories)*
11924 * Let BN_rand_range() abort with an error after 100 iterations
11925 without success (which indicates a broken PRNG).
11929 * Change BN_mod_sqrt() so that it verifies that the input value
11930 is really the square of the return value. (Previously,
11931 BN_mod_sqrt would show GIGO behaviour.)
11935 * Add named elliptic curves over binary fields from X9.62, SECG,
11936 and WAP/WTLS; add OIDs that were still missing.
11938 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11940 * Extend the EC library for elliptic curves over binary fields
11941 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11944 EC_GF2m_simple_method
11948 EC_GROUP_new_curve_GF2m
11949 EC_GROUP_set_curve_GF2m
11950 EC_GROUP_get_curve_GF2m
11951 EC_POINT_set_affine_coordinates_GF2m
11952 EC_POINT_get_affine_coordinates_GF2m
11953 EC_POINT_set_compressed_coordinates_GF2m
11955 Point compression for binary fields is disabled by default for
11956 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11959 As binary polynomials are represented as BIGNUMs, various members
11960 of the EC_GROUP and EC_POINT data structures can be shared
11961 between the implementations for prime fields and binary fields;
11962 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11963 are essentially identical to their `..._GFp` counterparts.
11964 (For simplicity, the `..._GFp` prefix has been dropped from
11965 various internal method names.)
11967 An internal 'field_div' method (similar to 'field_mul' and
11968 'field_sqr') has been added; this is used only for binary fields.
11970 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11972 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11973 through methods ('mul', 'precompute_mult').
11975 The generic implementations (now internally called 'ec_wNAF_mul'
11976 and 'ec_wNAF_precomputed_mult') remain the default if these
11977 methods are undefined.
11979 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11981 * New function EC_GROUP_get_degree, which is defined through
11982 EC_METHOD. For curves over prime fields, this returns the bit
11983 length of the modulus.
11985 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11987 * New functions EC_GROUP_dup, EC_POINT_dup.
11988 (These simply call ..._new and ..._copy).
11990 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11992 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11993 Polynomials are represented as BIGNUMs (where the sign bit is not
11994 used) in the following functions [macros]:
11997 BN_GF2m_sub [= BN_GF2m_add]
11998 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11999 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
12000 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
12002 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
12003 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
12004 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
12005 BN_GF2m_cmp [= BN_ucmp]
12007 (Note that only the 'mod' functions are actually for fields GF(2^m).
12008 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
12010 For some functions, an the irreducible polynomial defining a
12011 field can be given as an 'unsigned int[]' with strictly
12012 decreasing elements giving the indices of those bits that are set;
12013 i.e., p[] represents the polynomial
12014 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
12016 p[0] > p[1] > ... > p[k] = 0.
12017 This applies to the following functions:
12020 BN_GF2m_mod_mul_arr
12021 BN_GF2m_mod_sqr_arr
12022 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
12023 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
12024 BN_GF2m_mod_exp_arr
12025 BN_GF2m_mod_sqrt_arr
12026 BN_GF2m_mod_solve_quad_arr
12030 Conversion can be performed by the following functions:
12035 bntest.c has additional tests for binary polynomial arithmetic.
12037 Two implementations for BN_GF2m_mod_div() are available.
12038 The default algorithm simply uses BN_GF2m_mod_inv() and
12039 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
12040 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
12041 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
12043 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12045 * Add new error code 'ERR_R_DISABLED' that can be used when some
12046 functionality is disabled at compile-time.
12048 *Douglas Stebila <douglas.stebila@sun.com>*
12050 * Change default behaviour of 'openssl asn1parse' so that more
12051 information is visible when viewing, e.g., a certificate:
12053 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
12054 mode the content of non-printable OCTET STRINGs is output in a
12055 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
12056 avoid the appearance of a printable string.
12058 *Nils Larsch <nla@trustcenter.de>*
12060 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
12062 EC_GROUP_set_asn1_flag()
12063 EC_GROUP_get_asn1_flag()
12064 EC_GROUP_set_point_conversion_form()
12065 EC_GROUP_get_point_conversion_form()
12066 These control ASN1 encoding details:
12067 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
12068 has been set to OPENSSL_EC_NAMED_CURVE.
12069 - Points are encoded in uncompressed form by default; options for
12070 asn1_for are as for point2oct, namely
12071 POINT_CONVERSION_COMPRESSED
12072 POINT_CONVERSION_UNCOMPRESSED
12073 POINT_CONVERSION_HYBRID
12075 Also add 'seed' and 'seed_len' members to EC_GROUP with access
12077 EC_GROUP_set_seed()
12078 EC_GROUP_get0_seed()
12079 EC_GROUP_get_seed_len()
12080 This is used only for ASN1 purposes (so far).
12082 *Nils Larsch <nla@trustcenter.de>*
12084 * Add 'field_type' member to EC_METHOD, which holds the NID
12085 of the appropriate field type OID. The new function
12086 EC_METHOD_get_field_type() returns this value.
12088 *Nils Larsch <nla@trustcenter.de>*
12091 EC_POINT_point2bn()
12092 EC_POINT_bn2point()
12093 EC_POINT_point2hex()
12094 EC_POINT_hex2point()
12095 providing useful interfaces to EC_POINT_point2oct() and
12096 EC_POINT_oct2point().
12098 *Nils Larsch <nla@trustcenter.de>*
12100 * Change internals of the EC library so that the functions
12101 EC_GROUP_set_generator()
12102 EC_GROUP_get_generator()
12103 EC_GROUP_get_order()
12104 EC_GROUP_get_cofactor()
12105 are implemented directly in crypto/ec/ec_lib.c and not dispatched
12106 to methods, which would lead to unnecessary code duplication when
12107 adding different types of curves.
12109 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
12111 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
12112 arithmetic, and such that modified wNAFs are generated
12113 (which avoid length expansion in many cases).
12117 * Add a function EC_GROUP_check_discriminant() (defined via
12118 EC_METHOD) that verifies that the curve discriminant is non-zero.
12120 Add a function EC_GROUP_check() that makes some sanity tests
12121 on a EC_GROUP, its generator and order. This includes
12122 EC_GROUP_check_discriminant().
12124 *Nils Larsch <nla@trustcenter.de>*
12126 * Add ECDSA in new directory crypto/ecdsa/.
12128 Add applications 'openssl ecparam' and 'openssl ecdsa'
12129 (these are based on 'openssl dsaparam' and 'openssl dsa').
12131 ECDSA support is also included in various other files across the
12132 library. Most notably,
12133 - 'openssl req' now has a '-newkey ecdsa:file' option;
12134 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
12135 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
12136 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
12137 them suitable for ECDSA where domain parameters must be
12138 extracted before the specific public key;
12139 - ECDSA engine support has been added.
12141 *Nils Larsch <nla@trustcenter.de>*
12143 * Include some named elliptic curves, and add OIDs from X9.62,
12144 SECG, and WAP/WTLS. Each curve can be obtained from the new
12146 EC_GROUP_new_by_curve_name(),
12147 and the list of available named curves can be obtained with
12148 EC_get_builtin_curves().
12149 Also add a 'curve_name' member to EC_GROUP objects, which can be
12151 EC_GROUP_set_curve_name()
12152 EC_GROUP_get_curve_name()
12154 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
12156 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12157 was actually never needed) and in BN_mul(). The removal in BN_mul()
12158 required a small change in bn_mul_part_recursive() and the addition
12159 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12160 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12161 bn_sub_words() and bn_add_words() except they take arrays with
12166 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
12168 * Cleanse PEM buffers before freeing them since they may contain
12171 *Benjamin Bennett <ben@psc.edu>*
12173 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
12174 a ciphersuite string such as "DEFAULT:RSA" cannot enable
12175 authentication-only ciphersuites.
12179 * Since AES128 and AES256 share a single mask bit in the logic of
12180 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
12181 kludge to work properly if AES128 is available and AES256 isn't.
12185 * Expand security boundary to match 1.1.1 module.
12189 * Remove redundant features: hash file source, editing of test vectors
12190 modify fipsld to use external fips_premain.c signature.
12194 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12195 run algorithm test programs.
12199 * Make algorithm test programs more tolerant of whitespace.
12203 * Have SSL/TLS server implementation tolerate "mismatched" record
12204 protocol version while receiving ClientHello even if the
12205 ClientHello is fragmented. (The server can't insist on the
12206 particular protocol version it has chosen before the ServerHello
12207 message has informed the client about his choice.)
12211 * Load error codes if they are not already present instead of using a
12212 static variable. This allows them to be cleanly unloaded and reloaded.
12216 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12218 * Introduce limits to prevent malicious keys being able to
12219 cause a denial of service. ([CVE-2006-2940])
12221 *Steve Henson, Bodo Moeller*
12223 * Fix ASN.1 parsing of certain invalid structures that can result
12224 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12226 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12227 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12229 * Fix SSL client code which could crash if connecting to a
12230 malicious SSLv2 server. ([CVE-2006-4343])
12232 *Tavis Ormandy and Will Drewry, Google Security Team*
12234 * Change ciphersuite string processing so that an explicit
12235 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12236 will no longer include "AES128-SHA"), and any other similar
12237 ciphersuite (same bitmap) from *other* protocol versions (so that
12238 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12239 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12240 changes from 0.9.8b and 0.9.8d.
12244 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12246 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12247 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12249 * Change the Unix randomness entropy gathering to use poll() when
12250 possible instead of select(), since the latter has some
12251 undesirable limitations.
12253 *Darryl Miles via Richard Levitte and Bodo Moeller*
12255 * Disable rogue ciphersuites:
12257 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12258 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12259 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12261 The latter two were purportedly from
12262 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12265 Also deactivate the remaining ciphersuites from
12266 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12267 unofficial, and the ID has long expired.
12271 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12272 dual-core machines) and other potential thread-safety issues.
12276 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12278 * Adapt fipsld and the build system to link against the validated FIPS
12279 module in FIPS mode.
12283 * Fixes for VC++ 2005 build under Windows.
12287 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12288 from a Windows bash shell such as MSYS. It is autodetected from the
12289 "config" script when run from a VC++ environment. Modify standard VC++
12290 build to use fipscanister.o from the GNU make build.
12294 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12296 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12297 The value now differs depending on if you build for FIPS or not.
12298 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12299 safely run with a non-FIPSed libcrypto, as it may crash because of
12300 the difference induced by this change.
12304 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12306 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12307 (part of SSL_OP_ALL). This option used to disable the
12308 countermeasure against man-in-the-middle protocol-version
12309 rollback in the SSL 2.0 server implementation, which is a bad
12310 idea. ([CVE-2005-2969])
12312 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12313 for Information Security, National Institute of Advanced Industrial
12314 Science and Technology [AIST, Japan)]*
12316 * Minimal support for X9.31 signatures and PSS padding modes. This is
12317 mainly for FIPS compliance and not fully integrated at this stage.
12321 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12322 the exponentiation using a fixed-length exponent. (Otherwise,
12323 the information leaked through timing could expose the secret key
12324 after many signatures; cf. Bleichenbacher's attack on DSA with
12329 * Make a new fixed-window mod_exp implementation the default for
12330 RSA, DSA, and DH private-key operations so that the sequence of
12331 squares and multiplies and the memory access pattern are
12332 independent of the particular secret key. This will mitigate
12333 cache-timing and potential related attacks.
12335 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12336 and this is automatically used by BN_mod_exp_mont() if the new flag
12337 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12338 will use this BN flag for private exponents unless the flag
12339 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12340 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12342 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12344 * Change the client implementation for SSLv23_method() and
12345 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12346 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12347 (Previously, the SSL 2.0 backwards compatible Client Hello
12348 message format would be used even with SSL_OP_NO_SSLv2.)
12352 * Add support for smime-type MIME parameter in S/MIME messages which some
12357 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12358 a threadsafe manner. Modify rsa code to use new function and add calls
12359 to dsa and dh code (which had race conditions before).
12363 * Include the fixed error library code in the C error file definitions
12364 instead of fixing them up at runtime. This keeps the error code
12365 structures constant.
12369 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12371 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12374 * Fixes for newer kerberos headers. NB: the casts are needed because
12375 the 'length' field is signed on one version and unsigned on another
12376 with no (?) obvious way to tell the difference, without these VC++
12377 complains. Also the "definition" of FAR (blank) is no longer included
12378 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12379 some needed definitions.
12383 * Undo Cygwin change.
12387 * Added support for proxy certificates according to RFC 3820.
12388 Because they may be a security thread to unaware applications,
12389 they must be explicitly allowed in run-time. See
12390 docs/HOWTO/proxy_certificates.txt for further information.
12394 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12396 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12397 server and client random values. Previously
12398 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12399 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12401 This change has negligible security impact because:
12403 1. Server and client random values still have 24 bytes of pseudo random
12406 2. Server and client random values are sent in the clear in the initial
12409 3. The master secret is derived using the premaster secret (48 bytes in
12410 size for static RSA ciphersuites) as well as client server and random
12413 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12416 *Stephen Henson, reported by UK NISCC*
12418 * Use Windows randomness collection on Cygwin.
12422 * Fix hang in EGD/PRNGD query when communication socket is closed
12423 prematurely by EGD/PRNGD.
12425 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12427 * Prompt for pass phrases when appropriate for PKCS12 input format.
12431 * Back-port of selected performance improvements from development
12432 branch, as well as improved support for PowerPC platforms.
12436 * Add lots of checks for memory allocation failure, error codes to indicate
12437 failure and freeing up memory if a failure occurs.
12439 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12441 * Add new -passin argument to dgst.
12445 * Perform some character comparisons of different types in X509_NAME_cmp:
12446 this is needed for some certificates that re-encode DNs into UTF8Strings
12447 (in violation of RFC3280) and can't or won't issue name rollover
12452 * Make an explicit check during certificate validation to see that
12453 the CA setting in each certificate on the chain is correct. As a
12454 side effect always do the following basic checks on extensions,
12455 not just when there's an associated purpose to the check:
12457 - if there is an unhandled critical extension (unless the user
12458 has chosen to ignore this fault)
12459 - if the path length has been exceeded (if one is set at all)
12460 - that certain extensions fit the associated purpose (if one has
12465 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12467 * Avoid a race condition when CRLs are checked in a multi threaded
12468 environment. This would happen due to the reordering of the revoked
12469 entries during signature checking and serial number lookup. Now the
12470 encoding is cached and the serial number sort performed under a lock.
12471 Add new STACK function sk_is_sorted().
12475 * Add Delta CRL to the extension code.
12479 * Various fixes to s3_pkt.c so alerts are sent properly.
12481 *David Holmes <d.holmes@f5.com>*
12483 * Reduce the chances of duplicate issuer name and serial numbers (in
12484 violation of RFC3280) using the OpenSSL certificate creation utilities.
12485 This is done by creating a random 64 bit value for the initial serial
12486 number when a serial number file is created or when a self signed
12487 certificate is created using 'openssl req -x509'. The initial serial
12488 number file is created using 'openssl x509 -next_serial' in CA.pl
12489 rather than being initialized to 1.
12493 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12495 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12496 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12498 *Joe Orton, Steve Henson*
12500 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12503 *Joe Orton, Steve Henson*
12505 * Make it possible to have multiple active certificates with the same
12506 subject in the CA index file. This is done only if the keyword
12507 'unique_subject' is set to 'no' in the main CA section (default
12508 if 'CA_default') of the configuration file. The value is saved
12509 with the database itself in a separate index attribute file,
12510 named like the index file with '.attr' appended to the name.
12514 * X509 verify fixes. Disable broken certificate workarounds when
12515 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12516 keyUsage extension present. Don't accept CRLs with unhandled critical
12517 extensions: since verify currently doesn't process CRL extensions this
12518 rejects a CRL with *any* critical extensions. Add new verify error codes
12523 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12524 A clarification of RFC2560 will require the use of OCTET STRINGs and
12525 some implementations cannot handle the current raw format. Since OpenSSL
12526 copies and compares OCSP nonces as opaque blobs without any attempt at
12527 parsing them this should not create any compatibility issues.
12531 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12532 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12533 this HMAC (and other) operations are several times slower than OpenSSL
12538 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12540 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12542 * Use the correct content when signing type "other".
12546 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12548 * Fix various bugs revealed by running the NISCC test suite:
12550 Stop out of bounds reads in the ASN1 code when presented with
12551 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12553 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12555 If verify callback ignores invalid public key errors don't try to check
12556 certificate signature with the NULL public key.
12560 * New -ignore_err option in ocsp application to stop the server
12561 exiting on the first error in a request.
12565 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12566 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12571 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12572 extra data after the compression methods not only for TLS 1.0
12573 but also for SSL 3.0 (as required by the specification).
12575 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12577 * Change X509_certificate_type() to mark the key as exported/exportable
12578 when it's 512 *bits* long, not 512 bytes.
12582 * Change AES_cbc_encrypt() so it outputs exact multiple of
12583 blocks during encryption.
12587 * Various fixes to base64 BIO and non blocking I/O. On write
12588 flushes were not handled properly if the BIO retried. On read
12589 data was not being buffered properly and had various logic bugs.
12590 This also affects blocking I/O when the data being decoded is a
12595 * Various S/MIME bugfixes and compatibility changes:
12596 output correct application/pkcs7 MIME type if
12597 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12598 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12599 of files as .eml work). Correctly handle very long lines in MIME
12604 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12606 * Countermeasure against the Klima-Pokorny-Rosa extension of
12607 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12608 a protocol version number mismatch like a decryption error
12609 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12613 * Turn on RSA blinding by default in the default implementation
12614 to avoid a timing attack. Applications that don't want it can call
12615 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12616 They would be ill-advised to do so in most cases.
12618 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12620 * Change RSA blinding code so that it works when the PRNG is not
12621 seeded (in this case, the secret RSA exponent is abused as
12622 an unpredictable seed -- if it is not unpredictable, there
12623 is no point in blinding anyway). Make RSA blinding thread-safe
12624 by remembering the creator's thread ID in rsa->blinding and
12625 having all other threads use local one-time blinding factors
12626 (this requires more computation than sharing rsa->blinding, but
12627 avoids excessive locking; and if an RSA object is not shared
12628 between threads, blinding will still be very fast).
12632 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12633 ENGINE as defaults for all supported algorithms irrespective of
12634 the 'flags' parameter. 'flags' is now honoured, so applications
12635 should make sure they are passing it correctly.
12639 * Target "mingw" now allows native Windows code to be generated in
12640 the Cygwin environment as well as with the MinGW compiler.
12644 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12646 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12647 via timing by performing a MAC computation even if incorrect
12648 block cipher padding has been found. This is a countermeasure
12649 against active attacks where the attacker has to distinguish
12650 between bad padding and a MAC verification error. ([CVE-2003-0078])
12652 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12653 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12654 Martin Vuagnoux (EPFL, Ilion)*
12656 * Make the no-err option work as intended. The intention with no-err
12657 is not to have the whole error stack handling routines removed from
12658 libcrypto, it's only intended to remove all the function name and
12659 reason texts, thereby removing some of the footprint that may not
12660 be interesting if those errors aren't displayed anyway.
12662 NOTE: it's still possible for any application or module to have its
12663 own set of error texts inserted. The routines are there, just not
12664 used by default when no-err is given.
12668 * Add support for FreeBSD on IA64.
12670 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12672 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12673 Kerberos function mit_des_cbc_cksum(). Before this change,
12674 the value returned by DES_cbc_cksum() was like the one from
12675 mit_des_cbc_cksum(), except the bytes were swapped.
12677 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12679 * Allow an application to disable the automatic SSL chain building.
12680 Before this a rather primitive chain build was always performed in
12681 ssl3_output_cert_chain(): an application had no way to send the
12682 correct chain if the automatic operation produced an incorrect result.
12684 Now the chain builder is disabled if either:
12686 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12688 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12690 The reasoning behind this is that an application would not want the
12691 auto chain building to take place if extra chain certificates are
12692 present and it might also want a means of sending no additional
12693 certificates (for example the chain has two certificates and the
12698 * Add the possibility to build without the ENGINE framework.
12700 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12702 * Under Win32 gmtime() can return NULL: check return value in
12703 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12707 * DSA routines: under certain error conditions uninitialized BN objects
12708 could be freed. Solution: make sure initialization is performed early
12709 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12710 Nils Larsch <nla@trustcenter.de> via PR#459)
12714 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12715 checked on reconnect on the client side, therefore session resumption
12716 could still fail with a "ssl session id is different" error. This
12717 behaviour is masked when SSL_OP_ALL is used due to
12718 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12719 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12720 followup to PR #377.
12724 * IA-32 assembler support enhancements: unified ELF targets, support
12725 for SCO/Caldera platforms, fix for Cygwin shared build.
12729 * Add support for FreeBSD on sparc64. As a consequence, support for
12730 FreeBSD on non-x86 processors is separate from x86 processors on
12731 the config script, much like the NetBSD support.
12733 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12735 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12737 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12740 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12741 code (06) was taken as the first octet of the session ID and the last
12742 octet was ignored consequently. As a result SSLv2 client side session
12743 caching could not have worked due to the session ID mismatch between
12745 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12750 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12751 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12756 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12757 seems that in spite of existing for more than a year, many application
12758 author have done nothing to provide the necessary callbacks, which
12759 means that this particular engine will not work properly anywhere.
12760 This is a very unfortunate situation which forces us, in the name
12761 of usability, to give the hw_ncipher.c a static lock, which is part
12763 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12764 appear in 0.9.8 or later. We EXPECT application authors to have
12765 dealt properly with this when 0.9.8 is released (unless we actually
12766 make such changes in the libcrypto locking code that changes will
12767 have to be made anyway).
12771 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12772 octets have been read, EOF or an error occurs. Without this change
12773 some truncated ASN1 structures will not produce an error.
12777 * Disable Heimdal support, since it hasn't been fully implemented.
12778 Still give the possibility to force the use of Heimdal, but with
12779 warnings and a request that patches get sent to openssl-dev.
12783 * Add the VC-CE target, introduce the WINCE sysname, and add
12784 INSTALL.WCE and appropriate conditionals to make it build.
12786 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12788 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12789 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12790 edit numbers of the version.
12792 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12794 * Introduce safe string copy and catenation functions
12795 (BUF_strlcpy() and BUF_strlcat()).
12797 *Ben Laurie (CHATS) and Richard Levitte*
12799 * Avoid using fixed-size buffers for one-line DNs.
12801 *Ben Laurie (CHATS)*
12803 * Add BUF_MEM_grow_clean() to avoid information leakage when
12804 resizing buffers containing secrets, and use where appropriate.
12806 *Ben Laurie (CHATS)*
12808 * Avoid using fixed size buffers for configuration file location.
12810 *Ben Laurie (CHATS)*
12812 * Avoid filename truncation for various CA files.
12814 *Ben Laurie (CHATS)*
12816 * Use sizeof in preference to magic numbers.
12818 *Ben Laurie (CHATS)*
12820 * Avoid filename truncation in cert requests.
12822 *Ben Laurie (CHATS)*
12824 * Add assertions to check for (supposedly impossible) buffer
12827 *Ben Laurie (CHATS)*
12829 * Don't cache truncated DNS entries in the local cache (this could
12830 potentially lead to a spoofing attack).
12832 *Ben Laurie (CHATS)*
12834 * Fix various buffers to be large enough for hex/decimal
12835 representations in a platform independent manner.
12837 *Ben Laurie (CHATS)*
12839 * Add CRYPTO_realloc_clean() to avoid information leakage when
12840 resizing buffers containing secrets, and use where appropriate.
12842 *Ben Laurie (CHATS)*
12844 * Add BIO_indent() to avoid much slightly worrying code to do
12847 *Ben Laurie (CHATS)*
12849 * Convert sprintf()/BIO_puts() to BIO_printf().
12851 *Ben Laurie (CHATS)*
12853 * buffer_gets() could terminate with the buffer only half
12856 *Ben Laurie (CHATS)*
12858 * Add assertions to prevent user-supplied crypto functions from
12859 overflowing internal buffers by having large block sizes, etc.
12861 *Ben Laurie (CHATS)*
12863 * New OPENSSL_assert() macro (similar to assert(), but enabled
12866 *Ben Laurie (CHATS)*
12868 * Eliminate unused copy of key in RC4.
12870 *Ben Laurie (CHATS)*
12872 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12874 *Ben Laurie (CHATS)*
12876 * Fix off-by-one error in EGD path.
12878 *Ben Laurie (CHATS)*
12880 * If RANDFILE path is too long, ignore instead of truncating.
12882 *Ben Laurie (CHATS)*
12884 * Eliminate unused and incorrectly sized X.509 structure
12887 *Ben Laurie (CHATS)*
12889 * Eliminate unused and dangerous function knumber().
12891 *Ben Laurie (CHATS)*
12893 * Eliminate unused and dangerous structure, KSSL_ERR.
12895 *Ben Laurie (CHATS)*
12897 * Protect against overlong session ID context length in an encoded
12898 session object. Since these are local, this does not appear to be
12901 *Ben Laurie (CHATS)*
12903 * Change from security patch (see 0.9.6e below) that did not affect
12904 the 0.9.6 release series:
12906 Remote buffer overflow in SSL3 protocol - an attacker could
12907 supply an oversized master key in Kerberos-enabled versions.
12910 *Ben Laurie (CHATS)*
12912 * Change the SSL kerb5 codes to match RFC 2712.
12916 * Make -nameopt work fully for req and add -reqopt switch.
12918 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12920 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12922 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12924 * Make sure tests can be performed even if the corresponding algorithms
12925 have been removed entirely. This was also the last step to make
12926 OpenSSL compilable with DJGPP under all reasonable conditions.
12928 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12930 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12931 to allow version independent disabling of normally unselected ciphers,
12932 which may be activated as a side-effect of selecting a single cipher.
12934 (E.g., cipher list string "RSA" enables ciphersuites that are left
12935 out of "ALL" because they do not provide symmetric encryption.
12936 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12938 *Lutz Jaenicke, Bodo Moeller*
12940 * Add appropriate support for separate platform-dependent build
12941 directories. The recommended way to make a platform-dependent
12942 build directory is the following (tested on Linux), maybe with
12945 # Place yourself outside of the OpenSSL source tree. In
12946 # this example, the environment variable OPENSSL_SOURCE
12947 # is assumed to contain the absolute OpenSSL source directory.
12948 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12949 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12950 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12951 mkdir -p `dirname $F`
12952 ln -s $OPENSSL_SOURCE/$F $F
12955 To be absolutely sure not to disturb the source tree, a "make clean"
12956 is a good thing. If it isn't successful, don't worry about it,
12957 it probably means the source directory is very clean.
12961 * Make sure any ENGINE control commands make local copies of string
12962 pointers passed to them whenever necessary. Otherwise it is possible
12963 the caller may have overwritten (or deallocated) the original string
12964 data when a later ENGINE operation tries to use the stored values.
12966 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12968 * Improve diagnostics in file reading and command-line digests.
12970 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12972 * Add AES modes CFB and OFB to the object database. Correct an
12973 error in AES-CFB decryption.
12977 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12978 allows existing EVP_CIPHER_CTX structures to be reused after
12979 calling `EVP_*Final()`. This behaviour is used by encryption
12980 BIOs and some applications. This has the side effect that
12981 applications must explicitly clean up cipher contexts with
12982 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12986 * Check the values of dna and dnb in bn_mul_recursive before calling
12987 bn_mul_comba (a non zero value means the a or b arrays do not contain
12988 n2 elements) and fallback to bn_mul_normal if either is not zero.
12992 * Fix escaping of non-ASCII characters when using the -subj option
12993 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12997 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12998 form for "surname", serialNumber has no short form.
12999 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
13000 therefore remove "mail" short name for "internet 7".
13001 The OID for unique identifiers in X509 certificates is
13002 x500UniqueIdentifier, not uniqueIdentifier.
13003 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
13007 * Add an "init" command to the ENGINE config module and auto initialize
13008 ENGINEs. Without any "init" command the ENGINE will be initialized
13009 after all ctrl commands have been executed on it. If init=1 the
13010 ENGINE is initialized at that point (ctrls before that point are run
13011 on the uninitialized ENGINE and after on the initialized one). If
13012 init=0 then the ENGINE will not be initialized at all.
13016 * Fix the 'app_verify_callback' interface so that the user-defined
13017 argument is actually passed to the callback: In the
13018 SSL_CTX_set_cert_verify_callback() prototype, the callback
13019 declaration has been changed from
13022 int (*cb)(X509_STORE_CTX *,void *);
13023 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
13024 i=s->ctx->app_verify_callback(&ctx)
13025 has been changed into
13026 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
13028 To update applications using SSL_CTX_set_cert_verify_callback(),
13029 a dummy argument can be added to their callback functions.
13031 *D. K. Smetters <smetters@parc.xerox.com>*
13033 * Added the '4758cca' ENGINE to support IBM 4758 cards.
13035 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
13037 * Add and OPENSSL_LOAD_CONF define which will cause
13038 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
13039 This allows older applications to transparently support certain
13040 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
13041 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
13042 load the config file and OPENSSL_add_all_algorithms_conf() which will
13043 always load it have also been added.
13047 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
13048 Adjust NIDs and EVP layer.
13050 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13052 * Config modules support in openssl utility.
13054 Most commands now load modules from the config file,
13055 though in a few (such as version) this isn't done
13056 because it couldn't be used for anything.
13058 In the case of ca and req the config file used is
13059 the same as the utility itself: that is the -config
13060 command line option can be used to specify an
13065 * Move default behaviour from OPENSSL_config(). If appname is NULL
13066 use "openssl_conf" if filename is NULL use default openssl config file.
13070 * Add an argument to OPENSSL_config() to allow the use of an alternative
13071 config section name. Add a new flag to tolerate a missing config file
13072 and move code to CONF_modules_load_file().
13076 * Support for crypto accelerator cards from Accelerated Encryption
13077 Processing, www.aep.ie. (Use engine 'aep')
13078 The support was copied from 0.9.6c [engine] and adapted/corrected
13079 to work with the new engine framework.
13081 *AEP Inc. and Richard Levitte*
13083 * Support for SureWare crypto accelerator cards from Baltimore
13084 Technologies. (Use engine 'sureware')
13085 The support was copied from 0.9.6c [engine] and adapted
13086 to work with the new engine framework.
13090 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
13091 make the newer ENGINE framework commands for the CHIL engine work.
13093 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
13095 * Make it possible to produce shared libraries on ReliantUNIX.
13097 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
13099 * Add the configuration target debug-linux-ppro.
13100 Make 'openssl rsa' use the general key loading routines
13101 implemented in `apps.c`, and make those routines able to
13102 handle the key format FORMAT_NETSCAPE and the variant
13105 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13107 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13109 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13111 * Add -keyform to rsautl, and document -engine.
13113 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
13115 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
13116 BIO_R_NO_SUCH_FILE error code rather than the generic
13117 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
13121 * Add new functions
13122 ERR_peek_last_error
13123 ERR_peek_last_error_line
13124 ERR_peek_last_error_line_data.
13125 These are similar to
13127 ERR_peek_error_line
13128 ERR_peek_error_line_data,
13129 but report on the latest error recorded rather than the first one
13130 still in the error queue.
13132 *Ben Laurie, Bodo Moeller*
13134 * default_algorithms option in ENGINE config module. This allows things
13136 default_algorithms = ALL
13137 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
13141 * Preliminary ENGINE config module.
13145 * New experimental application configuration code.
13149 * Change the AES code to follow the same name structure as all other
13150 symmetric ciphers, and behave the same way. Move everything to
13151 the directory crypto/aes, thereby obsoleting crypto/rijndael.
13153 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13155 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
13157 *Ben Laurie and Theo de Raadt*
13159 * Add option to output public keys in req command.
13161 *Massimiliano Pala madwolf@openca.org*
13163 * Use wNAFs in EC_POINTs_mul() for improved efficiency
13164 (up to about 10% better than before for P-192 and P-224).
13168 * New functions/macros
13170 SSL_CTX_set_msg_callback(ctx, cb)
13171 SSL_CTX_set_msg_callback_arg(ctx, arg)
13172 SSL_set_msg_callback(ssl, cb)
13173 SSL_set_msg_callback_arg(ssl, arg)
13175 to request calling a callback function
13177 void cb(int write_p, int version, int content_type,
13178 const void *buf, size_t len, SSL *ssl, void *arg)
13180 whenever a protocol message has been completely received
13181 (write_p == 0) or sent (write_p == 1). Here 'version' is the
13182 protocol version according to which the SSL library interprets
13183 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
13184 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
13185 the content type as defined in the SSL 3.0/TLS 1.0 protocol
13186 specification (change_cipher_spec(20), alert(21), handshake(22)).
13187 'buf' and 'len' point to the actual message, 'ssl' to the
13188 SSL object, and 'arg' is the application-defined value set by
13189 SSL[_CTX]_set_msg_callback_arg().
13191 'openssl s_client' and 'openssl s_server' have new '-msg' options
13192 to enable a callback that displays all protocol messages.
13196 * Change the shared library support so shared libraries are built as
13197 soon as the corresponding static library is finished, and thereby get
13198 openssl and the test programs linked against the shared library.
13199 This still only happens when the keyword "shard" has been given to
13200 the configuration scripts.
13202 NOTE: shared library support is still an experimental thing, and
13203 backward binary compatibility is still not guaranteed.
13205 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13207 * Add support for Subject Information Access extension.
13209 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13211 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13212 additional bytes when new memory had to be allocated, not just
13213 when reusing an existing buffer.
13217 * New command line and configuration option 'utf8' for the req command.
13218 This allows field values to be specified as UTF8 strings.
13222 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13223 runs for the former and machine-readable output for the latter.
13227 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13228 of the e-mail address in the DN (i.e., it will go into a certificate
13229 extension only). The new configuration file option 'email_in_dn = no'
13230 has the same effect.
13232 *Massimiliano Pala madwolf@openca.org*
13234 * Change all functions with names starting with `des_` to be starting
13235 with `DES_` instead. Add wrappers that are compatible with libdes,
13236 but are named `_ossl_old_des_*`. Finally, add macros that map the
13237 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13238 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13239 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13242 Since we provide two compatibility mappings, the user needs to
13243 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13244 compatibility is desired. The default (i.e., when that macro
13245 isn't defined) is OpenSSL 0.9.6c compatibility.
13247 There are also macros that enable and disable the support of old
13248 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13249 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13250 are defined, the default will apply: to support the old des routines.
13252 In either case, one must include openssl/des.h to get the correct
13253 definitions. Do not try to just include openssl/des_old.h, that
13256 NOTE: This is a major break of an old API into a new one. Software
13257 authors are encouraged to switch to the `DES_` style functions. Some
13258 time in the future, des_old.h and the libdes compatibility functions
13259 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13260 default), and then completely removed.
13264 * Test for certificates which contain unsupported critical extensions.
13265 If such a certificate is found during a verify operation it is
13266 rejected by default: this behaviour can be overridden by either
13267 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13268 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13269 X509_supported_extension() has also been added which returns 1 if a
13270 particular extension is supported.
13274 * Modify the behaviour of EVP cipher functions in similar way to digests
13275 to retain compatibility with existing code.
13279 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13280 compatibility with existing code. In particular the 'ctx' parameter does
13281 not have to be to be initialized before the call to EVP_DigestInit() and
13282 it is tidied up after a call to EVP_DigestFinal(). New function
13283 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13284 EVP_MD_CTX_copy() changed to not require the destination to be
13285 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13286 requires the destination to be valid.
13288 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13289 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13293 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13294 so that complete 'Handshake' protocol structures are kept in memory
13295 instead of overwriting 'msg_type' and 'length' with 'body' data.
13299 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13301 *Massimo Santin via Richard Levitte*
13303 * Major restructuring to the underlying ENGINE code. This includes
13304 reduction of linker bloat, separation of pure "ENGINE" manipulation
13305 (initialisation, etc) from functionality dealing with implementations
13306 of specific crypto interfaces. This change also introduces integrated
13307 support for symmetric ciphers and digest implementations - so ENGINEs
13308 can now accelerate these by providing EVP_CIPHER and EVP_MD
13309 implementations of their own. This is detailed in
13310 [crypto/engine/README.md](crypto/engine/README.md)
13311 as it couldn't be adequately described here. However, there are a few
13312 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13313 were changed in the original introduction of ENGINE code have now
13314 reverted back - the hooking from this code to ENGINE is now a good
13315 deal more passive and at run-time, operations deal directly with
13316 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13317 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13318 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13319 they were not being used by the framework as there is no concept of a
13320 BIGNUM_METHOD and they could not be generalised to the new
13321 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13322 ENGINE_cpy() has been removed as it cannot be consistently defined in
13327 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13331 * Change mkdef.pl to sort symbols that get the same entry number,
13332 and make sure the automatically generated functions `ERR_load_*`
13333 become part of libeay.num as well.
13337 * New function SSL_renegotiate_pending(). This returns true once
13338 renegotiation has been requested (either SSL_renegotiate() call
13339 or HelloRequest/ClientHello received from the peer) and becomes
13340 false once a handshake has been completed.
13341 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13342 sends a HelloRequest, but does not ensure that a handshake takes
13343 place. SSL_renegotiate_pending() is useful for checking if the
13344 client has followed the request.)
13348 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13349 By default, clients may request session resumption even during
13350 renegotiation (if session ID contexts permit); with this option,
13351 session resumption is possible only in the first handshake.
13353 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13354 more bits available for options that should not be part of
13355 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13359 * Add some demos for certificate and certificate request creation.
13363 * Make maximum certificate chain size accepted from the peer application
13364 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13365 "Douglas E. Engert" <deengert@anl.gov>.
13369 * Add support for shared libraries for Unixware-7
13370 (Boyd Lynn Gerber <gerberb@zenez.com>).
13374 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13375 be done prior to destruction. Use this to unload error strings from
13376 ENGINEs that load their own error strings. NB: This adds two new API
13377 functions to "get" and "set" this destroy handler in an ENGINE.
13381 * Alter all existing ENGINE implementations (except "openssl" and
13382 "openbsd") to dynamically instantiate their own error strings. This
13383 makes them more flexible to be built both as statically-linked ENGINEs
13384 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13385 Also, add stub code to each that makes building them as self-contained
13386 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13390 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13391 implementations into applications that are completely implemented in
13392 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13393 commands that can be used to configure what shared-library to load and
13394 to control aspects of the way it is handled. Also, made an update to
13395 the [README-Engine.md](README-Engine.md) file
13396 that brings its information up-to-date and
13397 provides some information and instructions on the "dynamic" ENGINE
13398 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13402 * Make it possible to unload ranges of ERR strings with a new
13403 "ERR_unload_strings" function.
13407 * Add a copy() function to EVP_MD.
13411 * Make EVP_MD routines take a context pointer instead of just the
13412 md_data void pointer.
13416 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13417 that the digest can only process a single chunk of data
13418 (typically because it is provided by a piece of
13419 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13420 is only going to provide a single chunk of data, and hence the
13421 framework needn't accumulate the data for oneshot drivers.
13425 * As with "ERR", make it possible to replace the underlying "ex_data"
13426 functions. This change also alters the storage and management of global
13427 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13428 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13429 index counters. The API functions that use this state have been changed
13430 to take a "class_index" rather than pointers to the class's local STACK
13431 and counter, and there is now an API function to dynamically create new
13432 classes. This centralisation allows us to (a) plug a lot of the
13433 thread-safety problems that existed, and (b) makes it possible to clean
13434 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13435 such data would previously have always leaked in application code and
13436 workarounds were in place to make the memory debugging turn a blind eye
13437 to it. Application code that doesn't use this new function will still
13438 leak as before, but their memory debugging output will announce it now
13439 rather than letting it slide.
13441 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13442 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13443 has a return value to indicate success or failure.
13447 * Make it possible to replace the underlying "ERR" functions such that the
13448 global state (2 LHASH tables and 2 locks) is only used by the "default"
13449 implementation. This change also adds two functions to "get" and "set"
13450 the implementation prior to it being automatically set the first time
13451 any other ERR function takes place. Ie. an application can call "get",
13452 pass the return value to a module it has just loaded, and that module
13453 can call its own "set" function using that value. This means the
13454 module's "ERR" operations will use (and modify) the error state in the
13455 application and not in its own statically linked copy of OpenSSL code.
13459 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13460 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13461 the operation, and provides a more encapsulated way for external code
13462 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13463 to use these functions rather than manually incrementing the counts.
13465 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13469 * Add EVP test program.
13473 * Add symmetric cipher support to ENGINE. Expect the API to change!
13477 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13478 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13479 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13480 These allow a CRL to be built without having to access X509_CRL fields
13481 directly. Modify 'ca' application to use new functions.
13485 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13486 bug workarounds. Rollback attack detection is a security feature.
13487 The problem will only arise on OpenSSL servers when TLSv1 is not
13488 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13489 Software authors not wanting to support TLSv1 will have special reasons
13490 for their choice and can explicitly enable this option.
13492 *Bodo Moeller, Lutz Jaenicke*
13494 * Rationalise EVP so it can be extended: don't include a union of
13495 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13496 (similar to those existing for EVP_CIPHER_CTX).
13501 EVP_MD_CTX_init(&md); /* new function call */
13502 EVP_DigestInit(&md, EVP_sha1());
13503 EVP_DigestUpdate(&md, in, len);
13504 EVP_DigestFinal(&md, out, NULL);
13505 EVP_MD_CTX_cleanup(&md); /* new function call */
13509 * Make DES key schedule conform to the usual scheme, as well as
13510 correcting its structure. This means that calls to DES functions
13511 now have to pass a pointer to a des_key_schedule instead of a
13512 plain des_key_schedule (which was actually always a pointer
13515 des_key_schedule ks;
13517 des_set_key_checked(..., &ks);
13518 des_ncbc_encrypt(..., &ks, ...);
13520 (Note that a later change renames 'des_...' into 'DES_...'.)
13524 * Initial reduction of linker bloat: the use of some functions, such as
13525 PEM causes large amounts of unused functions to be linked in due to
13526 poor organisation. For example pem_all.c contains every PEM function
13527 which has a knock on effect of linking in large amounts of (unused)
13528 ASN1 code. Grouping together similar functions and splitting unrelated
13529 functions prevents this.
13533 * Cleanup of EVP macros.
13537 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13538 correct `_ecb suffix`.
13542 * Add initial OCSP responder support to ocsp application. The
13543 revocation information is handled using the text based index
13544 use by the ca application. The responder can either handle
13545 requests generated internally, supplied in files (for example
13546 via a CGI script) or using an internal minimal server.
13550 * Add configuration choices to get zlib compression for TLS.
13554 * Changes to Kerberos SSL for RFC 2712 compliance:
13555 1. Implemented real KerberosWrapper, instead of just using
13556 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13557 2. Implemented optional authenticator field of KerberosWrapper.
13559 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13560 and authenticator structs; see crypto/krb5/.
13562 Generalized Kerberos calls to support multiple Kerberos libraries.
13563 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13564 via Richard Levitte*
13566 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13567 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13568 values for each of the key sizes rather than having just
13569 parameters (and 'speed' generating keys each time).
13573 * Speed up EVP routines.
13576 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13577 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13578 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13579 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13581 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13582 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13583 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13586 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13588 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13592 * Added the OS2-EMX target.
13594 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13596 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13597 New functions to support `NCONF` routines in extension code.
13598 New function `CONF_set_nconf()`
13599 to allow functions which take an `NCONF` to also handle the old `LHASH`
13600 structure: this means that the old `CONF` compatible routines can be
13601 retained (in particular w.rt. extensions) without having to duplicate the
13602 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13606 * Enhance the general user interface with mechanisms for inner control
13607 and with possibilities to have yes/no kind of prompts.
13611 * Change all calls to low-level digest routines in the library and
13612 applications to use EVP. Add missing calls to HMAC_cleanup() and
13613 don't assume HMAC_CTX can be copied using memcpy().
13615 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13617 * Add the possibility to control engines through control names but with
13618 arbitrary arguments instead of just a string.
13619 Change the key loaders to take a UI_METHOD instead of a callback
13620 function pointer. NOTE: this breaks binary compatibility with earlier
13621 versions of OpenSSL [engine].
13622 Adapt the nCipher code for these new conditions and add a card insertion
13627 * Enhance the general user interface with mechanisms to better support
13628 dialog box interfaces, application-defined prompts, the possibility
13629 to use defaults (for example default passwords from somewhere else)
13630 and interrupts/cancellations.
13634 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13635 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13639 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13640 tidy up some unnecessarily weird code in 'sk_new()').
13642 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13644 * Change the key loading routines for ENGINEs to use the same kind
13645 callback (pem_password_cb) as all other routines that need this
13650 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13651 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13652 than this minimum value is recommended.
13656 * New random seeder for OpenVMS, using the system process statistics
13657 that are easily reachable.
13661 * Windows apparently can't transparently handle global
13662 variables defined in DLLs. Initialisations such as:
13664 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13666 won't compile. This is used by the any applications that need to
13667 declare their own ASN1 modules. This was fixed by adding the option
13668 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13669 needed for static libraries under Win32.
13673 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13674 setting of purpose and trust fields. New X509_STORE trust and
13675 purpose functions and tidy up setting in other SSL functions.
13679 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13680 structure. These are inherited by X509_STORE_CTX when it is
13681 initialised. This allows various defaults to be set in the
13682 X509_STORE structure (such as flags for CRL checking and custom
13683 purpose or trust settings) for functions which only use X509_STORE_CTX
13684 internally such as S/MIME.
13686 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13687 trust settings if they are not set in X509_STORE. This allows X509_STORE
13688 purposes and trust (in S/MIME for example) to override any set by default.
13690 Add command line options for CRL checking to smime, s_client and s_server
13695 * Initial CRL based revocation checking. If the CRL checking flag(s)
13696 are set then the CRL is looked up in the X509_STORE structure and
13697 its validity and signature checked, then if the certificate is found
13698 in the CRL the verify fails with a revoked error.
13700 Various new CRL related callbacks added to X509_STORE_CTX structure.
13702 Command line options added to 'verify' application to support this.
13704 This needs some additional work, such as being able to handle multiple
13705 CRLs with different times, extension based lookup (rather than just
13706 by subject name) and ultimately more complete V2 CRL extension
13711 * Add a general user interface API (crypto/ui/). This is designed
13712 to replace things like des_read_password and friends (backward
13713 compatibility functions using this new API are provided).
13714 The purpose is to remove prompting functions from the DES code
13715 section as well as provide for prompting through dialog boxes in
13716 a window system and the like.
13720 * Add "ex_data" support to ENGINE so implementations can add state at a
13721 per-structure level rather than having to store it globally.
13725 * Make it possible for ENGINE structures to be copied when retrieved by
13726 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13727 This causes the "original" ENGINE structure to act like a template,
13728 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13729 operational state can be localised to each ENGINE structure, despite the
13730 fact they all share the same "methods". New ENGINE structures returned in
13731 this case have no functional references and the return value is the single
13732 structural reference. This matches the single structural reference returned
13733 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13738 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13739 needs to match any other type at all we need to manually clear the
13744 * Changes to the "openssl engine" utility to include;
13745 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13746 about an ENGINE's available control commands.
13747 - executing control commands from command line arguments using the
13748 '-pre' and '-post' switches. '-post' is only used if '-t' is
13749 specified and the ENGINE is successfully initialised. The syntax for
13750 the individual commands are colon-separated, for example;
13751 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13755 * New dynamic control command support for ENGINEs. ENGINEs can now
13756 declare their own commands (numbers), names (strings), descriptions,
13757 and input types for run-time discovery by calling applications. A
13758 subset of these commands are implicitly classed as "executable"
13759 depending on their input type, and only these can be invoked through
13760 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13761 can be based on user input, config files, etc). The distinction is
13762 that "executable" commands cannot return anything other than a boolean
13763 result and can only support numeric or string input, whereas some
13764 discoverable commands may only be for direct use through
13765 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13766 pointers, or other custom uses. The "executable" commands are to
13767 support parameterisations of ENGINE behaviour that can be
13768 unambiguously defined by ENGINEs and used consistently across any
13769 OpenSSL-based application. Commands have been added to all the
13770 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13771 control over shared-library paths without source code alterations.
13775 * Changed all ENGINE implementations to dynamically allocate their
13776 ENGINEs rather than declaring them statically. Apart from this being
13777 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13778 this also allows the implementations to compile without using the
13779 internal engine_int.h header.
13783 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13784 'const' value. Any code that should be able to modify a RAND_METHOD
13785 should already have non-const pointers to it (ie. they should only
13786 modify their own ones).
13790 * Made a variety of little tweaks to the ENGINE code.
13791 - "atalla" and "ubsec" string definitions were moved from header files
13792 to C code. "nuron" string definitions were placed in variables
13793 rather than hard-coded - allowing parameterisation of these values
13794 later on via ctrl() commands.
13795 - Removed unused "#if 0"'d code.
13796 - Fixed engine list iteration code so it uses ENGINE_free() to release
13797 structural references.
13798 - Constified the RAND_METHOD element of ENGINE structures.
13799 - Constified various get/set functions as appropriate and added
13800 missing functions (including a catch-all ENGINE_cpy that duplicates
13801 all ENGINE values onto a new ENGINE except reference counts/state).
13802 - Removed NULL parameter checks in get/set functions. Setting a method
13803 or function to NULL is a way of cancelling out a previously set
13804 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13805 and doesn't justify the extra error symbols and code.
13806 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13807 flags from engine_int.h to engine.h.
13808 - Changed prototypes for ENGINE handler functions (init(), finish(),
13809 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13813 * Implement binary inversion algorithm for BN_mod_inverse in addition
13814 to the algorithm using long division. The binary algorithm can be
13815 used only if the modulus is odd. On 32-bit systems, it is faster
13816 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13817 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13818 up to 450 bits. In 64-bit environments, the binary algorithm
13819 appears to be advantageous for much longer moduli; here we use it
13820 for moduli up to 2048 bits.
13824 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13825 could not support the combine flag in choice fields.
13829 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13830 extensions from a certificate request to the certificate.
13834 * Allow multiple 'certopt' and 'nameopt' options to be separated
13835 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13836 file: this allows the display of the certificate about to be
13837 signed to be customised, to allow certain fields to be included
13838 or excluded and extension details. The old system didn't display
13839 multicharacter strings properly, omitted fields not in the policy
13840 and couldn't display additional details such as extensions.
13844 * Function EC_POINTs_mul for multiple scalar multiplication
13845 of an arbitrary number of elliptic curve points
13846 \sum scalars[i]*points[i],
13847 optionally including the generator defined for the EC_GROUP:
13848 scalar*generator + \sum scalars[i]*points[i].
13850 EC_POINT_mul is a simple wrapper function for the typical case
13851 that the point list has just one item (besides the optional
13856 * First EC_METHODs for curves over GF(p):
13858 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13859 operations and provides various method functions that can also
13860 operate with faster implementations of modular arithmetic.
13862 EC_GFp_mont_method() reuses most functions that are part of
13863 EC_GFp_simple_method, but uses Montgomery arithmetic.
13865 *Bodo Moeller; point addition and point doubling
13866 implementation directly derived from source code provided by
13867 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13869 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13870 crypto/ec/ec_lib.c):
13872 Curves are EC_GROUP objects (with an optional group generator)
13873 based on EC_METHODs that are built into the library.
13875 Points are EC_POINT objects based on EC_GROUP objects.
13877 Most of the framework would be able to handle curves over arbitrary
13878 finite fields, but as there are no obvious types for fields other
13879 than GF(p), some functions are limited to that for now.
13883 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13884 that the file contains a complete HTTP response.
13888 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13889 change the def and num file printf format specifier from "%-40sXXX"
13890 to "%-39s XXX". The latter will always guarantee a space after the
13891 field while the former will cause them to run together if the field
13892 is 40 of more characters long.
13896 * Constify the cipher and digest 'method' functions and structures
13897 and modify related functions to take constant EVP_MD and EVP_CIPHER
13902 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13903 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13907 * Modify `EVP_Digest*()` routines so they now return values. Although the
13908 internal software routines can never fail additional hardware versions
13913 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13915 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13916 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13919 ERR_R_NESTED_ASN1_ERROR
13921 ERR_R_MISSING_ASN1_EOS
13922 were 4 .. 9, conflicting with
13923 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13925 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13926 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13928 Add new error code 'ERR_R_INTERNAL_ERROR'.
13932 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13937 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13938 sets the subject name for a new request or supersedes the
13939 subject name in a given request. Formats that can be parsed are
13940 'CN=Some Name, OU=myOU, C=IT'
13942 'CN=Some Name/OU=myOU/C=IT'.
13944 Add options '-batch' and '-verbose' to 'openssl req'.
13946 *Massimiliano Pala <madwolf@hackmasters.net>*
13948 * Introduce the possibility to access global variables through
13949 functions on platform were that's the best way to handle exporting
13950 global variables in shared libraries. To enable this functionality,
13951 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13952 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13953 is normally done by Configure or something similar).
13955 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13956 in the source file (foo.c) like this:
13958 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13959 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13961 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13962 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13964 OPENSSL_DECLARE_GLOBAL(int,foo);
13965 #define foo OPENSSL_GLOBAL_REF(foo)
13966 OPENSSL_DECLARE_GLOBAL(double,bar);
13967 #define bar OPENSSL_GLOBAL_REF(bar)
13969 The #defines are very important, and therefore so is including the
13970 header file everywhere where the defined globals are used.
13972 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13973 of ASN.1 items, but that structure is a bit different.
13975 The largest change is in util/mkdef.pl which has been enhanced with
13976 better and easier to understand logic to choose which symbols should
13977 go into the Windows .def files as well as a number of fixes and code
13978 cleanup (among others, algorithm keywords are now sorted
13979 lexicographically to avoid constant rewrites).
13983 * In BN_div() keep a copy of the sign of 'num' before writing the
13984 result to 'rm' because if rm==num the value will be overwritten
13985 and produce the wrong result if 'num' is negative: this caused
13986 problems with BN_mod() and BN_nnmod().
13990 * Function OCSP_request_verify(). This checks the signature on an
13991 OCSP request and verifies the signer certificate. The signer
13992 certificate is just checked for a generic purpose and OCSP request
13997 * Add OCSP_check_validity() function to check the validity of OCSP
13998 responses. OCSP responses are prepared in real time and may only
13999 be a few seconds old. Simply checking that the current time lies
14000 between thisUpdate and nextUpdate max reject otherwise valid responses
14001 caused by either OCSP responder or client clock inaccuracy. Instead
14002 we allow thisUpdate and nextUpdate to fall within a certain period of
14003 the current time. The age of the response can also optionally be
14004 checked. Two new options -validity_period and -status_age added to
14009 * If signature or public key algorithm is unrecognized print out its
14010 OID rather that just UNKNOWN.
14014 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
14015 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
14016 ID to be generated from the issuer certificate alone which can then be
14017 passed to OCSP_id_issuer_cmp().
14021 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
14022 ASN1 modules to export functions returning ASN1_ITEM pointers
14023 instead of the ASN1_ITEM structures themselves. This adds several
14024 new macros which allow the underlying ASN1 function/structure to
14025 be accessed transparently. As a result code should not use ASN1_ITEM
14026 references directly (such as &X509_it) but instead use the relevant
14027 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
14028 use of the new ASN1 code on platforms where exporting structures
14029 is problematical (for example in shared libraries) but exporting
14030 functions returning pointers to structures is not.
14034 * Add support for overriding the generation of SSL/TLS session IDs.
14035 These callbacks can be registered either in an SSL_CTX or per SSL.
14036 The purpose of this is to allow applications to control, if they wish,
14037 the arbitrary values chosen for use as session IDs, particularly as it
14038 can be useful for session caching in multiple-server environments. A
14039 command-line switch for testing this (and any client code that wishes
14040 to use such a feature) has been added to "s_server".
14042 *Geoff Thorpe, Lutz Jaenicke*
14044 * Modify mkdef.pl to recognise and parse preprocessor conditionals
14045 of the form `#if defined(...) || defined(...) || ...` and
14046 `#if !defined(...) && !defined(...) && ...`. This also avoids
14047 the growing number of special cases it was previously handling.
14051 * Make all configuration macros available for application by making
14052 sure they are available in opensslconf.h, by giving them names starting
14053 with `OPENSSL_` to avoid conflicts with other packages and by making
14054 sure e_os2.h will cover all platform-specific cases together with
14056 Additionally, it is now possible to define configuration/platform-
14057 specific names (called "system identities"). In the C code, these
14058 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
14059 macro with the name beginning with `OPENSSL_SYS_`, which is determined
14060 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
14065 * New option -set_serial to 'req' and 'x509' this allows the serial
14066 number to use to be specified on the command line. Previously self
14067 signed certificates were hard coded with serial number 0 and the
14068 CA options of 'x509' had to use a serial number in a file which was
14073 * New options to 'ca' utility to support V2 CRL entry extensions.
14074 Currently CRL reason, invalidity date and hold instruction are
14075 supported. Add new CRL extensions to V3 code and some new objects.
14079 * New function EVP_CIPHER_CTX_set_padding() this is used to
14080 disable standard block padding (aka PKCS#5 padding) in the EVP
14081 API, which was previously mandatory. This means that the data is
14082 not padded in any way and so the total length much be a multiple
14083 of the block size, otherwise an error occurs.
14087 * Initial (incomplete) OCSP SSL support.
14091 * New function OCSP_parse_url(). This splits up a URL into its host,
14092 port and path components: primarily to parse OCSP URLs. New -url
14093 option to ocsp utility.
14097 * New nonce behavior. The return value of OCSP_check_nonce() now
14098 reflects the various checks performed. Applications can decide
14099 whether to tolerate certain situations such as an absent nonce
14100 in a response when one was present in a request: the ocsp application
14101 just prints out a warning. New function OCSP_add1_basic_nonce()
14102 this is to allow responders to include a nonce in a response even if
14103 the request is nonce-less.
14107 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
14108 skipped when using openssl x509 multiple times on a single input file,
14109 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
14113 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
14114 set string type: to handle setting ASN1_TIME structures. Fix ca
14115 utility to correctly initialize revocation date of CRLs.
14119 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
14120 the clients preferred ciphersuites and rather use its own preferences.
14121 Should help to work around M$ SGC (Server Gated Cryptography) bug in
14122 Internet Explorer by ensuring unchanged hash method during stepup.
14123 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
14127 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
14128 to aes and add a new 'exist' option to print out symbols that don't
14133 * Additional options to ocsp utility to allow flags to be set and
14134 additional certificates supplied.
14138 * Add the option -VAfile to 'openssl ocsp', so the user can give the
14139 OCSP client a number of certificate to only verify the response
14144 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
14145 handle the new API. Currently only ECB, CBC modes supported. Add new
14148 Add TLS AES ciphersuites as described in RFC3268, "Advanced
14149 Encryption Standard (AES) Ciphersuites for Transport Layer
14150 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
14151 not enabled by default and were not part of the "ALL" ciphersuite
14152 alias because they were not yet official; they could be
14153 explicitly requested by specifying the "AESdraft" ciphersuite
14154 group alias. In the final release of OpenSSL 0.9.7, the group
14155 alias is called "AES" and is part of "ALL".)
14157 *Ben Laurie, Steve Henson, Bodo Moeller*
14159 * New function OCSP_copy_nonce() to copy nonce value (if present) from
14160 request to response.
14164 * Functions for OCSP responders. OCSP_request_onereq_count(),
14165 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
14166 extract information from a certificate request. OCSP_response_create()
14167 creates a response and optionally adds a basic response structure.
14168 OCSP_basic_add1_status() adds a complete single response to a basic
14169 response and returns the OCSP_SINGLERESP structure just added (to allow
14170 extensions to be included for example). OCSP_basic_add1_cert() adds a
14171 certificate to a basic response and OCSP_basic_sign() signs a basic
14172 response with various flags. New helper functions ASN1_TIME_check()
14173 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
14174 (converts ASN1_TIME to GeneralizedTime).
14178 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
14179 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
14180 structure from a certificate. X509_pubkey_digest() digests the public_key
14181 contents: this is used in various key identifiers.
14185 * Make sk_sort() tolerate a NULL argument.
14187 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14189 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14190 passed by the function are trusted implicitly. If any of them signed the
14191 response then it is assumed to be valid and is not verified.
14195 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14196 to data. This was previously part of the PKCS7 ASN1 code. This
14197 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14198 *Steve Henson, reported by Kenneth R. Robinette
14199 <support@securenetterm.com>*
14201 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14202 routines: without these tracing memory leaks is very painful.
14203 Fix leaks in PKCS12 and PKCS7 routines.
14207 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14208 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14209 effectively meant GeneralizedTime would never be used. Now it
14210 is initialised to -1 but X509_time_adj() now has to check the value
14211 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14212 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14213 *Steve Henson, reported by Kenneth R. Robinette
14214 <support@securenetterm.com>*
14216 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14217 result in a zero length in the ASN1_INTEGER structure which was
14218 not consistent with the structure when d2i_ASN1_INTEGER() was used
14219 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14220 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14221 where it did not print out a minus for negative ASN1_INTEGER.
14225 * Add summary printout to ocsp utility. The various functions which
14226 convert status values to strings have been renamed to:
14227 OCSP_response_status_str(), OCSP_cert_status_str() and
14228 OCSP_crl_reason_str() and are no longer static. New options
14229 to verify nonce values and to disable verification. OCSP response
14230 printout format cleaned up.
14234 * Add additional OCSP certificate checks. These are those specified
14235 in RFC2560. This consists of two separate checks: the CA of the
14236 certificate being checked must either be the OCSP signer certificate
14237 or the issuer of the OCSP signer certificate. In the latter case the
14238 OCSP signer certificate must contain the OCSP signing extended key
14239 usage. This check is performed by attempting to match the OCSP
14240 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14241 in the OCSP_CERTID structures of the response.
14245 * Initial OCSP certificate verification added to OCSP_basic_verify()
14246 and related routines. This uses the standard OpenSSL certificate
14247 verify routines to perform initial checks (just CA validity) and
14248 to obtain the certificate chain. Then additional checks will be
14249 performed on the chain. Currently the root CA is checked to see
14250 if it is explicitly trusted for OCSP signing. This is used to set
14251 a root CA as a global signing root: that is any certificate that
14252 chains to that CA is an acceptable OCSP signing certificate.
14256 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14257 extensions from a separate configuration file.
14258 As when reading extensions from the main configuration file,
14259 the '-extensions ...' option may be used for specifying the
14262 *Massimiliano Pala <madwolf@comune.modena.it>*
14264 * New OCSP utility. Allows OCSP requests to be generated or
14265 read. The request can be sent to a responder and the output
14266 parsed, outputted or printed in text form. Not complete yet:
14267 still needs to check the OCSP response validity.
14271 * New subcommands for 'openssl ca':
14272 `openssl ca -status <serial>` prints the status of the cert with
14273 the given serial number (according to the index file).
14274 `openssl ca -updatedb` updates the expiry status of certificates
14277 *Massimiliano Pala <madwolf@comune.modena.it>*
14279 * New '-newreq-nodes' command option to CA.pl. This is like
14280 '-newreq', but calls 'openssl req' with the '-nodes' option
14281 so that the resulting key is not encrypted.
14283 *Damien Miller <djm@mindrot.org>*
14285 * New configuration for the GNU Hurd.
14287 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14289 * Initial code to implement OCSP basic response verify. This
14290 is currently incomplete. Currently just finds the signer's
14291 certificate and verifies the signature on the response.
14295 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14296 value of OPENSSLDIR. This is available via the new '-d' option
14297 to 'openssl version', and is also included in 'openssl version -a'.
14301 * Allowing defining memory allocation callbacks that will be given
14302 file name and line number information in additional arguments
14303 (a `const char*` and an int). The basic functionality remains, as
14304 well as the original possibility to just replace malloc(),
14305 realloc() and free() by functions that do not know about these
14306 additional arguments. To register and find out the current
14307 settings for extended allocation functions, the following
14308 functions are provided:
14310 CRYPTO_set_mem_ex_functions
14311 CRYPTO_set_locked_mem_ex_functions
14312 CRYPTO_get_mem_ex_functions
14313 CRYPTO_get_locked_mem_ex_functions
14315 These work the same way as CRYPTO_set_mem_functions and friends.
14316 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14317 extended allocation function is enabled.
14318 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14319 a conventional allocation function is enabled.
14321 *Richard Levitte, Bodo Moeller*
14323 * Finish off removing the remaining LHASH function pointer casts.
14324 There should no longer be any prototype-casting required when using
14325 the LHASH abstraction, and any casts that remain are "bugs". See
14326 the callback types and macros at the head of lhash.h for details
14327 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14331 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14332 If /dev/[u]random devices are not available or do not return enough
14333 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14335 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14336 /etc/entropy will be queried once each in this sequence, querying stops
14337 when enough entropy was collected without querying more sockets.
14341 * Change the Unix RAND_poll() variant to be able to poll several
14342 random devices, as specified by DEVRANDOM, until a sufficient amount
14343 of data has been collected. We spend at most 10 ms on each file
14344 (select timeout) and read in non-blocking mode. DEVRANDOM now
14345 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14346 (previously it was just the string "/dev/urandom"), so on typical
14347 platforms the 10 ms delay will never occur.
14348 Also separate out the Unix variant to its own file, rand_unix.c.
14349 For VMS, there's a currently-empty rand_vms.c.
14353 * Move OCSP client related routines to ocsp_cl.c. These
14354 provide utility functions which an application needing
14355 to issue a request to an OCSP responder and analyse the
14356 response will typically need: as opposed to those which an
14357 OCSP responder itself would need which will be added later.
14359 OCSP_request_sign() signs an OCSP request with an API similar
14360 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14361 response. OCSP_response_get1_basic() extracts basic response
14362 from response. OCSP_resp_find_status(): finds and extracts status
14363 information from an OCSP_CERTID structure (which will be created
14364 when the request structure is built). These are built from lower
14365 level functions which work on OCSP_SINGLERESP structures but
14366 won't normally be used unless the application wishes to examine
14367 extensions in the OCSP response for example.
14369 Replace nonce routines with a pair of functions.
14370 OCSP_request_add1_nonce() adds a nonce value and optionally
14371 generates a random value. OCSP_check_nonce() checks the
14372 validity of the nonce in an OCSP response.
14376 * Change function OCSP_request_add() to OCSP_request_add0_id().
14377 This doesn't copy the supplied OCSP_CERTID and avoids the
14378 need to free up the newly created id. Change return type
14379 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14380 This can then be used to add extensions to the request.
14381 Deleted OCSP_request_new(), since most of its functionality
14382 is now in OCSP_REQUEST_new() (and the case insensitive name
14383 clash) apart from the ability to set the request name which
14384 will be added elsewhere.
14388 * Update OCSP API. Remove obsolete extensions argument from
14389 various functions. Extensions are now handled using the new
14390 OCSP extension code. New simple OCSP HTTP function which
14391 can be used to send requests and parse the response.
14395 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14396 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14397 uses the special reorder version of SET OF to sort the attributes
14398 and reorder them to match the encoded order. This resolves a long
14399 standing problem: a verify on a PKCS7 structure just after signing
14400 it used to fail because the attribute order did not match the
14401 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14402 it uses the received order. This is necessary to tolerate some broken
14403 software that does not order SET OF. This is handled by encoding
14404 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14405 to produce the required SET OF.
14409 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14410 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14411 files to get correct declarations of the ASN.1 item variables.
14415 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14416 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14417 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14418 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14419 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14420 ASN1_ITEM and no wrapper functions.
14424 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14425 replace the old function pointer based I/O routines. Change most of
14426 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14430 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14431 lines, recognize more "algorithms" that can be deselected, and make
14432 it complain about algorithm deselection that isn't recognised.
14436 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14437 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14438 to use new functions. Add NO_ASN1_OLD which can be set to remove
14439 some old style ASN1 functions: this can be used to determine if old
14440 code will still work when these eventually go away.
14444 * New extension functions for OCSP structures, these follow the
14445 same conventions as certificates and CRLs.
14449 * New function X509V3_add1_i2d(). This automatically encodes and
14450 adds an extension. Its behaviour can be customised with various
14451 flags to append, replace or delete. Various wrappers added for
14452 certificates and CRLs.
14456 * Fix to avoid calling the underlying ASN1 print routine when
14457 an extension cannot be parsed. Correct a typo in the
14458 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14462 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14463 entries for variables.
14467 * Add functionality to `apps/openssl.c` for detecting locking
14468 problems: As the program is single-threaded, all we have
14469 to do is register a locking callback using an array for
14470 storing which locks are currently held by the program.
14474 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14475 SSL_get_ex_data_X509_STORE_idx(), which is used in
14476 ssl_verify_cert_chain() and thus can be called at any time
14477 during TLS/SSL handshakes so that thread-safety is essential.
14478 Unfortunately, the ex_data design is not at all suited
14479 for multi-threaded use, so it probably should be abolished.
14483 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14485 *Broadcom, tweaked and integrated by Geoff Thorpe*
14487 * Move common extension printing code to new function
14488 X509V3_print_extensions(). Reorganise OCSP print routines and
14489 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14493 * New function X509_signature_print() to remove duplication in some
14498 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14499 set (this was treated exactly the same as SET OF previously). This
14500 is used to reorder the STACK representing the structure to match the
14501 encoding. This will be used to get round a problem where a PKCS7
14502 structure which was signed could not be verified because the STACK
14503 order did not reflect the encoded order.
14507 * Reimplement the OCSP ASN1 module using the new code.
14511 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14512 for its ASN1 operations. The old style function pointers still exist
14513 for now but they will eventually go away.
14517 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14518 completely replaces the old ASN1 functionality with a table driven
14519 encoder and decoder which interprets an ASN1_ITEM structure describing
14520 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14521 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14522 has also been converted to the new form.
14526 * Change BN_mod_exp_recp so that negative moduli are tolerated
14527 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14528 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14529 for negative moduli.
14533 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14534 of not touching the result's sign bit.
14538 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14543 * Changed the LHASH code to use prototypes for callbacks, and created
14544 macros to declare and implement thin (optionally static) functions
14545 that provide type-safety and avoid function pointer casting for the
14546 type-specific callbacks.
14550 * Added Kerberos Cipher Suites to be used with TLS, as written in
14552 *Veers Staats <staatsvr@asc.hpc.mil>,
14553 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14555 * Reformat the FAQ so the different questions and answers can be divided
14556 in sections depending on the subject.
14560 * Have the zlib compression code load ZLIB.DLL dynamically under
14565 * New function BN_mod_sqrt for computing square roots modulo a prime
14566 (using the probabilistic Tonelli-Shanks algorithm unless
14567 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14568 be handled deterministically).
14570 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14572 * Make BN_mod_inverse faster by explicitly handling small quotients
14573 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14574 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14578 * New function BN_kronecker.
14582 * Fix BN_gcd so that it works on negative inputs; the result is
14583 positive unless both parameters are zero.
14584 Previously something reasonably close to an infinite loop was
14585 possible because numbers could be growing instead of shrinking
14586 in the implementation of Euclid's algorithm.
14590 * Fix BN_is_word() and BN_is_one() macros to take into account the
14591 sign of the number in question.
14593 Fix BN_is_word(a,w) to work correctly for w == 0.
14595 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14596 because its test if the absolute value of 'a' equals 'w'.
14597 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14598 it exists mostly for use in the implementations of BN_is_zero(),
14599 BN_is_one(), and BN_is_word().
14603 * New function BN_swap.
14607 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14608 the exponentiation functions are more likely to produce reasonable
14609 results on negative inputs.
14613 * Change BN_mod_mul so that the result is always non-negative.
14614 Previously, it could be negative if one of the factors was negative;
14615 I don't think anyone really wanted that behaviour.
14619 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14620 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14621 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14622 and add new functions:
14631 BN_mod_lshift1_quick
14633 BN_mod_lshift_quick
14635 These functions always generate non-negative results.
14637 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14638 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14640 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14641 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14642 be reduced modulo `m`.
14644 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14647 The following entry accidentally appeared in the CHANGES file
14648 distributed with OpenSSL 0.9.7. The modifications described in
14649 it do *not* apply to OpenSSL 0.9.7.
14651 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14652 was actually never needed) and in BN_mul(). The removal in BN_mul()
14653 required a small change in bn_mul_part_recursive() and the addition
14654 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14655 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14656 bn_sub_words() and bn_add_words() except they take arrays with
14662 * In 'openssl passwd', verify passwords read from the terminal
14663 unless the '-salt' option is used (which usually means that
14664 verification would just waste user's time since the resulting
14665 hash is going to be compared with some given password hash)
14666 or the new '-noverify' option is used.
14668 This is an incompatible change, but it does not affect
14669 non-interactive use of 'openssl passwd' (passwords on the command
14670 line, '-stdin' option, '-in ...' option) and thus should not
14671 cause any problems.
14675 * Remove all references to RSAref, since there's no more need for it.
14679 * Make DSO load along a path given through an environment variable
14680 (SHLIB_PATH) with shl_load().
14684 * Constify the ENGINE code as a result of BIGNUM constification.
14685 Also constify the RSA code and most things related to it. In a
14686 few places, most notable in the depth of the ASN.1 code, ugly
14687 casts back to non-const were required (to be solved at a later
14692 * Make it so the openssl application has all engines loaded by default.
14696 * Constify the BIGNUM routines a little more.
14700 * Add the following functions:
14702 ENGINE_load_cswift()
14704 ENGINE_load_atalla()
14705 ENGINE_load_nuron()
14706 ENGINE_load_builtin_engines()
14708 That way, an application can itself choose if external engines that
14709 are built-in in OpenSSL shall ever be used or not. The benefit is
14710 that applications won't have to be linked with libdl or other dso
14711 libraries unless it's really needed.
14713 Changed 'openssl engine' to load all engines on demand.
14714 Changed the engine header files to avoid the duplication of some
14715 declarations (they differed!).
14719 * 'openssl engine' can now list capabilities.
14723 * Better error reporting in 'openssl engine'.
14727 * Never call load_dh_param(NULL) in s_server.
14731 * Add engine application. It can currently list engines by name and
14732 identity, and test if they are actually available.
14736 * Improve RPM specification file by forcing symbolic linking and making
14737 sure the installed documentation is also owned by root.root.
14739 *Damien Miller <djm@mindrot.org>*
14741 * Give the OpenSSL applications more possibilities to make use of
14742 keys (public as well as private) handled by engines.
14746 * Add OCSP code that comes from CertCo.
14750 * Add VMS support for the Rijndael code.
14754 * Added untested support for Nuron crypto accelerator.
14758 * Add support for external cryptographic devices. This code was
14759 previously distributed separately as the "engine" branch.
14761 *Geoff Thorpe, Richard Levitte*
14763 * Rework the filename-translation in the DSO code. It is now possible to
14764 have far greater control over how a "name" is turned into a filename
14765 depending on the operating environment and any oddities about the
14766 different shared library filenames on each system.
14770 * Support threads on FreeBSD-elf in Configure.
14774 * Fix for SHA1 assembly problem with MASM: it produces
14775 warnings about corrupt line number information when assembling
14776 with debugging information. This is caused by the overlapping
14779 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14782 NCONF_get_number() has no error checking at all. As a replacement,
14783 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14784 promoted strongly. The old NCONF_get_number is kept around for
14785 binary backward compatibility.
14786 Make it possible for methods to load from something other than a BIO,
14787 by providing a function pointer that is given a name instead of a BIO.
14788 For example, this could be used to load configuration data from an
14793 * Fix for non blocking accept BIOs. Added new I/O special reason
14794 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14795 with non blocking I/O was not possible because no retry code was
14796 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14801 * Added the beginnings of Rijndael support.
14805 * Fix for bug in DirectoryString mask setting. Add support for
14806 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14807 to allow certificate printing to more controllable, additional
14808 'certopt' option to 'x509' to allow new printing options to be
14813 * Clean old EAY MD5 hack from e_os.h.
14817 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14819 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14820 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14822 *Joe Orton, Steve Henson*
14824 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14826 * Fix additional bug revealed by the NISCC test suite:
14828 Stop bug triggering large recursion when presented with
14829 certain ASN.1 tags ([CVE-2003-0851])
14833 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14835 * Fix various bugs revealed by running the NISCC test suite:
14837 Stop out of bounds reads in the ASN1 code when presented with
14838 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14840 If verify callback ignores invalid public key errors don't try to check
14841 certificate signature with the NULL public key.
14845 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14846 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14851 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14852 extra data after the compression methods not only for TLS 1.0
14853 but also for SSL 3.0 (as required by the specification).
14855 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14857 * Change X509_certificate_type() to mark the key as exported/exportable
14858 when it's 512 *bits* long, not 512 bytes.
14862 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14864 * Countermeasure against the Klima-Pokorny-Rosa extension of
14865 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14866 a protocol version number mismatch like a decryption error
14867 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14871 * Turn on RSA blinding by default in the default implementation
14872 to avoid a timing attack. Applications that don't want it can call
14873 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14874 They would be ill-advised to do so in most cases.
14876 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14878 * Change RSA blinding code so that it works when the PRNG is not
14879 seeded (in this case, the secret RSA exponent is abused as
14880 an unpredictable seed -- if it is not unpredictable, there
14881 is no point in blinding anyway). Make RSA blinding thread-safe
14882 by remembering the creator's thread ID in rsa->blinding and
14883 having all other threads use local one-time blinding factors
14884 (this requires more computation than sharing rsa->blinding, but
14885 avoids excessive locking; and if an RSA object is not shared
14886 between threads, blinding will still be very fast).
14890 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14892 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14893 via timing by performing a MAC computation even if incorrect
14894 block cipher padding has been found. This is a countermeasure
14895 against active attacks where the attacker has to distinguish
14896 between bad padding and a MAC verification error. ([CVE-2003-0078])
14898 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14899 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14900 Martin Vuagnoux (EPFL, Ilion)*
14902 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14904 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14905 memory from its contents. This is done with a counter that will
14906 place alternating values in each byte. This can be used to solve
14907 two issues: 1) the removal of calls to memset() by highly optimizing
14908 compilers, and 2) cleansing with other values than 0, since those can
14909 be read through on certain media, for example a swap space on disk.
14913 * Bugfix: client side session caching did not work with external caching,
14914 because the session->cipher setting was not restored when reloading
14915 from the external cache. This problem was masked, when
14916 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14917 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14921 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14922 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14924 *Zeev Lieber <zeev-l@yahoo.com>*
14926 * Undo an undocumented change introduced in 0.9.6e which caused
14927 repeated calls to OpenSSL_add_all_ciphers() and
14928 OpenSSL_add_all_digests() to be ignored, even after calling
14933 * Change the default configuration reader to deal with last line not
14934 being properly terminated.
14938 * Change X509_NAME_cmp() so it applies the special rules on handling
14939 DN values that are of type PrintableString, as well as RDNs of type
14940 emailAddress where the value has the type ia5String.
14942 *stefank@valicert.com via Richard Levitte*
14944 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14945 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14946 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14947 the bitwise-OR of the two for use by the majority of applications
14948 wanting this behaviour, and update the docs. The documented
14949 behaviour and actual behaviour were inconsistent and had been
14950 changing anyway, so this is more a bug-fix than a behavioural
14953 *Geoff Thorpe, diagnosed by Nadav Har'El*
14955 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14956 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14960 * Fix initialization code race conditions in
14961 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14962 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14963 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14964 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14965 ssl2_get_cipher_by_char(),
14966 ssl3_get_cipher_by_char().
14968 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14970 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14971 the cached sessions are flushed, as the remove_cb() might use ex_data
14972 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14973 (see [openssl.org #212]).
14975 *Geoff Thorpe, Lutz Jaenicke*
14977 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14978 length, instead of the encoding length to d2i_ASN1_OBJECT.
14982 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14984 * [In 0.9.6g-engine release:]
14985 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14987 *Lynn Gazis <lgazis@rainbow.com>*
14989 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14991 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14992 and get fix the header length calculation.
14993 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14994 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14996 * Use proper error handling instead of 'assertions' in buffer
14997 overflow checks added in 0.9.6e. This prevents DoS (the
14998 assertions could call abort()).
15000 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
15002 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
15004 * Add various sanity checks to asn1_get_length() to reject
15005 the ASN1 length bytes if they exceed sizeof(long), will appear
15006 negative or the content length exceeds the length of the
15009 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15011 * Fix cipher selection routines: ciphers without encryption had no flags
15012 for the cipher strength set and where therefore not handled correctly
15013 by the selection routines (PR #130).
15017 * Fix EVP_dsa_sha macro.
15022 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
15023 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
15024 that was added in OpenSSL 0.9.6d.
15026 As the countermeasure turned out to be incompatible with some
15027 broken SSL implementations, the new option is part of SSL_OP_ALL.
15028 SSL_OP_ALL is usually employed when compatibility with weird SSL
15029 implementations is desired (e.g. '-bugs' option to 's_client' and
15030 's_server'), so the new option is automatically set in many
15035 * Changes in security patch:
15037 Changes marked "(CHATS)" were sponsored by the Defense Advanced
15038 Research Projects Agency (DARPA) and Air Force Research Laboratory,
15039 Air Force Materiel Command, USAF, under agreement number
15042 * Add various sanity checks to asn1_get_length() to reject
15043 the ASN1 length bytes if they exceed sizeof(long), will appear
15044 negative or the content length exceeds the length of the
15045 supplied buffer. ([CVE-2002-0659])
15047 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15049 * Assertions for various potential buffer overflows, not known to
15050 happen in practice.
15052 *Ben Laurie (CHATS)*
15054 * Various temporary buffers to hold ASCII versions of integers were
15055 too small for 64 bit platforms. ([CVE-2002-0655])
15056 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
15058 * Remote buffer overflow in SSL3 protocol - an attacker could
15059 supply an oversized session ID to a client. ([CVE-2002-0656])
15061 *Ben Laurie (CHATS)*
15063 * Remote buffer overflow in SSL2 protocol - an attacker could
15064 supply an oversized client master key. ([CVE-2002-0656])
15066 *Ben Laurie (CHATS)*
15068 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
15070 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
15071 encoded as NULL) with id-dsa-with-sha1.
15073 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
15075 * Check various `X509_...()` return values in `apps/req.c`.
15077 *Nils Larsch <nla@trustcenter.de>*
15079 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
15080 an end-of-file condition would erroneously be flagged, when the CRLF
15081 was just at the end of a processed block. The bug was discovered when
15082 processing data through a buffering memory BIO handing the data to a
15083 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
15084 <ptsekov@syntrex.com> and Nedelcho Stanev.
15088 * Implement a countermeasure against a vulnerability recently found
15089 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
15090 before application data chunks to avoid the use of known IVs
15091 with data potentially chosen by the attacker.
15095 * Fix length checks in ssl3_get_client_hello().
15099 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
15100 to prevent ssl3_read_internal() from incorrectly assuming that
15101 ssl3_read_bytes() found application data while handshake
15102 processing was enabled when in fact s->s3->in_read_app_data was
15103 merely automatically cleared during the initial handshake.
15105 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
15107 * Fix object definitions for Private and Enterprise: they were not
15108 recognized in their shortname (=lowercase) representation. Extend
15109 obj_dat.pl to issue an error when using undefined keywords instead
15110 of silently ignoring the problem (Svenning Sorensen
15111 <sss@sss.dnsalias.net>).
15115 * Fix DH_generate_parameters() so that it works for 'non-standard'
15116 generators, i.e. generators other than 2 and 5. (Previously, the
15117 code did not properly initialise the 'add' and 'rem' values to
15118 BN_generate_prime().)
15120 In the new general case, we do not insist that 'generator' is
15121 actually a primitive root: This requirement is rather pointless;
15122 a generator of the order-q subgroup is just as good, if not
15127 * Map new X509 verification errors to alerts. Discovered and submitted by
15128 Tom Wu <tom@arcot.com>.
15132 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
15133 returning non-zero before the data has been completely received
15134 when using non-blocking I/O.
15136 *Bodo Moeller; problem pointed out by John Hughes*
15138 * Some of the ciphers missed the strength entry (SSL_LOW etc).
15140 *Ben Laurie, Lutz Jaenicke*
15142 * Fix bug in SSL_clear(): bad sessions were not removed (found by
15143 Yoram Zahavi <YoramZ@gilian.com>).
15147 * Add information about CygWin 1.3 and on, and preserve proper
15148 configuration for the versions before that.
15150 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
15152 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
15153 check whether we deal with a copy of a session and do not delete from
15154 the cache in this case. Problem reported by "Izhar Shoshani Levi"
15155 <izhar@checkpoint.com>.
15159 * Do not store session data into the internal session cache, if it
15160 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
15161 flag is set). Proposed by Aslam <aslam@funk.com>.
15165 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
15170 * [In 0.9.6d-engine release:]
15171 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
15173 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
15175 * Add the configuration target linux-s390x.
15177 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
15179 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
15180 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
15181 variable as an indication that a ClientHello message has been
15182 received. As the flag value will be lost between multiple
15183 invocations of ssl3_accept when using non-blocking I/O, the
15184 function may not be aware that a handshake has actually taken
15185 place, thus preventing a new session from being added to the
15188 To avoid this problem, we now set s->new_session to 2 instead of
15189 using a local variable.
15191 *Lutz Jaenicke, Bodo Moeller*
15193 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15194 if the SSL_R_LENGTH_MISMATCH error is detected.
15196 *Geoff Thorpe, Bodo Moeller*
15198 * New 'shared_ldflag' column in Configure platform table.
15202 * Fix EVP_CIPHER_mode macro.
15204 *"Dan S. Camper" <dan@bti.net>*
15206 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15207 type, we must throw them away by setting rr->length to 0.
15209 *D P Chang <dpc@qualys.com>*
15211 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15213 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15214 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15215 worked incorrectly for those cases where range = `10..._2` and
15216 `3*range` is two bits longer than range.)
15220 * Only add signing time to PKCS7 structures if it is not already
15225 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15226 OBJ_ld_ce should be OBJ_id_ce.
15227 Also some ip-pda OIDs in crypto/objects/objects.txt were
15228 incorrect (cf. RFC 3039).
15230 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15232 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15233 returns early because it has nothing to do.
15235 *Andy Schneider <andy.schneider@bjss.co.uk>*
15237 * [In 0.9.6c-engine release:]
15238 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15240 *Andy Schneider <andy.schneider@bjss.co.uk>*
15242 * [In 0.9.6c-engine release:]
15243 Add support for Cryptographic Appliance's keyserver technology.
15244 (Use engine 'keyclient')
15246 *Cryptographic Appliances and Geoff Thorpe*
15248 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15249 is called via tools/c89.sh because arguments have to be
15250 rearranged (all '-L' options must appear before the first object
15253 *Richard Shapiro <rshapiro@abinitio.com>*
15255 * [In 0.9.6c-engine release:]
15256 Add support for Broadcom crypto accelerator cards, backported
15259 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15261 * [In 0.9.6c-engine release:]
15262 Add support for SureWare crypto accelerator cards from
15263 Baltimore Technologies. (Use engine 'sureware')
15265 *Baltimore Technologies and Mark Cox*
15267 * [In 0.9.6c-engine release:]
15268 Add support for crypto accelerator cards from Accelerated
15269 Encryption Processing, www.aep.ie. (Use engine 'aep')
15271 *AEP Inc. and Mark Cox*
15273 * Add a configuration entry for gcc on UnixWare.
15275 *Gary Benson <gbenson@redhat.com>*
15277 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15278 messages are stored in a single piece (fixed-length part and
15279 variable-length part combined) and fix various bugs found on the way.
15283 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15284 instead. BIO_gethostbyname() does not know what timeouts are
15285 appropriate, so entries would stay in cache even when they have
15287 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15289 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15290 faced with a pathologically small ClientHello fragment that does
15291 not contain client_version: Instead of aborting with an error,
15292 simply choose the highest available protocol version (i.e.,
15293 TLS 1.0 unless it is disabled). In practice, ClientHello
15294 messages are never sent like this, but this change gives us
15295 strictly correct behaviour at least for TLS.
15299 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15300 never resets s->method to s->ctx->method when called from within
15301 one of the SSL handshake functions.
15303 *Bodo Moeller; problem pointed out by Niko Baric*
15305 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15306 (sent using the client's version number) if client_version is
15307 smaller than the protocol version in use. Also change
15308 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15309 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15310 the client will at least see that alert.
15314 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15319 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15320 client receives HelloRequest while in a handshake.
15322 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15324 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15325 should end in 'break', not 'goto end' which circumvents various
15326 cleanups done in state SSL_ST_OK. But session related stuff
15327 must be disabled for SSL_ST_OK in the case that we just sent a
15330 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15331 before just sending a HelloRequest.
15333 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15335 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15336 reveal whether illegal block cipher padding was found or a MAC
15337 verification error occurred. (Neither SSLerr() codes nor alerts
15338 are directly visible to potential attackers, but the information
15339 may leak via logfiles.)
15341 Similar changes are not required for the SSL 2.0 implementation
15342 because the number of padding bytes is sent in clear for SSL 2.0,
15343 and the extra bytes are just ignored. However ssl/s2_pkt.c
15344 failed to verify that the purported number of padding bytes is in
15349 * Add OpenUNIX-8 support including shared libraries
15350 (Boyd Lynn Gerber <gerberb@zenez.com>).
15354 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15355 'wristwatch attack' using huge encoding parameters (cf.
15356 James H. Manger's CRYPTO 2001 paper). Note that the
15357 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15358 encoding parameters and hence was not vulnerable.
15362 * BN_sqr() bug fix.
15364 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15366 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15367 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15368 followed by modular reduction.
15370 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15372 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15373 equivalent based on BN_pseudo_rand() instead of BN_rand().
15377 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15378 This function was broken, as the check for a new client hello message
15379 to handle SGC did not allow these large messages.
15380 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15384 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15388 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15389 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15393 * Rework the configuration and shared library support for Tru64 Unix.
15394 The configuration part makes use of modern compiler features and
15395 still retains old compiler behavior for those that run older versions
15396 of the OS. The shared library support part includes a variant that
15397 uses the RPATH feature, and is available through the special
15398 configuration target "alpha-cc-rpath", which will never be selected
15401 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15403 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15404 with the same message size as in ssl3_get_certificate_request().
15405 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15406 messages might inadvertently be reject as too long.
15408 *Petr Lampa <lampa@fee.vutbr.cz>*
15410 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15414 * Modified SSL library such that the verify_callback that has been set
15415 specifically for an SSL object with SSL_set_verify() is actually being
15416 used. Before the change, a verify_callback set with this function was
15417 ignored and the verify_callback() set in the SSL_CTX at the time of
15418 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15419 to allow the necessary settings.
15423 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15424 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15425 done automatically (in contradiction to the requirements of the C
15426 standard). This made problems when used from OpenSSH.
15430 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15431 dh->length and always used
15433 BN_rand_range(priv_key, dh->p).
15435 BN_rand_range() is not necessary for Diffie-Hellman, and this
15436 specific range makes Diffie-Hellman unnecessarily inefficient if
15437 dh->length (recommended exponent length) is much smaller than the
15438 length of dh->p. We could use BN_rand_range() if the order of
15439 the subgroup was stored in the DH structure, but we only have
15444 BN_rand(priv_key, l, ...)
15446 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15453 RSA_eay_public_encrypt
15454 RSA_eay_private_decrypt
15455 RSA_eay_private_encrypt (signing)
15456 RSA_eay_public_decrypt (signature verification)
15458 (default implementations for RSA_public_encrypt,
15459 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15460 always reject numbers >= n.
15464 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15465 to synchronize access to 'locking_thread'. This is necessary on
15466 systems where access to 'locking_thread' (an 'unsigned long'
15467 variable) is not atomic.
15471 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15472 *before* setting the 'crypto_lock_rand' flag. The previous code had
15473 a race condition if 0 is a valid thread ID.
15475 *Travis Vitek <vitek@roguewave.com>*
15477 * Add support for shared libraries under Irix.
15479 *Albert Chin-A-Young <china@thewrittenword.com>*
15481 * Add configuration option to build on Linux on both big-endian and
15482 little-endian MIPS.
15484 *Ralf Baechle <ralf@uni-koblenz.de>*
15486 * Add the possibility to create shared libraries on HP-UX.
15490 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15492 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15493 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15494 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15495 PRNG state recovery was possible based on the output of
15496 one PRNG request appropriately sized to gain knowledge on
15497 'md' followed by enough consecutive 1-byte PRNG requests
15498 to traverse all of 'state'.
15500 1. When updating 'md_local' (the current thread's copy of 'md')
15501 during PRNG output generation, hash all of the previous
15502 'md_local' value, not just the half used for PRNG output.
15504 2. Make the number of bytes from 'state' included into the hash
15505 independent from the number of PRNG bytes requested.
15507 The first measure alone would be sufficient to avoid
15508 Markku-Juhani's attack. (Actually it had never occurred
15509 to me that the half of 'md_local' used for chaining was the
15510 half from which PRNG output bytes were taken -- I had always
15511 assumed that the secret half would be used.) The second
15512 measure makes sure that additional data from 'state' is never
15513 mixed into 'md_local' in small portions; this heuristically
15514 further strengthens the PRNG.
15518 * Fix crypto/bn/asm/mips3.s.
15522 * When only the key is given to "enc", the IV is undefined. Print out
15523 an error message in this case.
15527 * Handle special case when X509_NAME is empty in X509 printing routines.
15531 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15532 positive and less than q.
15536 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15537 used: it isn't thread safe and the add_lock_callback should handle
15540 *Paul Rose <Paul.Rose@bridge.com>*
15542 * Verify that incoming data obeys the block size in
15543 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15549 *Ulf Möller, Bodo Möller*
15551 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15552 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15553 when fixing the server behaviour for backwards-compatible 'client
15554 hello' messages. (Note that the attack is impractical against
15555 SSL 3.0 and TLS 1.0 anyway because length and version checking
15556 means that the probability of guessing a valid ciphertext is
15557 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15560 Before 0.9.5, the countermeasure (hide the error by generating a
15561 random 'decryption result') did not work properly because
15562 ERR_clear_error() was missing, meaning that SSL_get_error() would
15563 detect the supposedly ignored error.
15565 Both problems are now fixed.
15569 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15570 (previously it was 1024).
15574 * Fix for compatibility mode trust settings: ignore trust settings
15575 unless some valid trust or reject settings are present.
15579 * Fix for blowfish EVP: its a variable length cipher.
15583 * Fix various bugs related to DSA S/MIME verification. Handle missing
15584 parameters in DSA public key structures and return an error in the
15585 DSA routines if parameters are absent.
15589 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15590 in the current directory if neither $RANDFILE nor $HOME was set.
15591 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15592 caused some confusion to Windows users who haven't defined $HOME.
15593 Thus RAND_file_name() is changed again: e_os.h can define a
15594 DEFAULT_HOME, which will be used if $HOME is not set.
15595 For Windows, we use "C:"; on other platforms, we still require
15596 environment variables.
15598 * Move 'if (!initialized) RAND_poll()' into regions protected by
15599 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15600 having multiple threads call RAND_poll() concurrently.
15604 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15605 combination of a flag and a thread ID variable.
15606 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15607 flag), *other* threads can enter ssleay_add_bytes without obeying
15608 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15609 that they do not hold after the first thread unsets add_do_not_lock).
15613 * Change bctest again: '-x' expressions are not available in all
15614 versions of 'test'.
15618 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15620 * Fix a couple of memory leaks in PKCS7_dataDecode()
15622 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15624 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15625 the default extension for executables, if any. Also, make the perl
15626 scripts that use symlink() to test if it really exists and use "cp"
15627 if it doesn't. All this made OpenSSL compilable and installable in
15632 * Fix for asn1_GetSequence() for indefinite length constructed data.
15633 If SEQUENCE is length is indefinite just set c->slen to the total
15634 amount of data available.
15636 *Steve Henson, reported by shige@FreeBSD.org*
15638 *This change does not apply to 0.9.7.*
15640 * Change bctest to avoid here-documents inside command substitution
15641 (workaround for FreeBSD /bin/sh bug).
15642 For compatibility with Ultrix, avoid shell functions (introduced
15643 in the bctest version that searches along $PATH).
15647 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15648 with des_encrypt() defined on some operating systems, like Solaris
15653 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15654 On the Importance of Eliminating Errors in Cryptographic
15655 Computations, J. Cryptology 14 (2001) 2, 101-119,
15656 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15660 * MIPS assembler BIGNUM division bug fix.
15664 * Disabled incorrect Alpha assembler code.
15668 * Fix PKCS#7 decode routines so they correctly update the length
15669 after reading an EOC for the EXPLICIT tag.
15673 *This change does not apply to 0.9.7.*
15675 * Fix bug in PKCS#12 key generation routines. This was triggered
15676 if a 3DES key was generated with a 0 initial byte. Include
15677 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15678 (but broken) behaviour.
15682 * Enhance bctest to search for a working bc along $PATH and print
15685 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15687 * Fix memory leaks in err.c: free err_data string if necessary;
15688 don't write to the wrong index in ERR_set_error_data.
15692 * Implement ssl23_peek (analogous to ssl23_read), which previously
15697 * Replace rdtsc with `_emit` statements for VC++ version 5.
15699 *Jeremy Cooper <jeremy@baymoo.org>*
15701 * Make it possible to reuse SSLv2 sessions.
15705 * In copy_email() check for >= 0 as a return value for
15706 X509_NAME_get_index_by_NID() since 0 is a valid index.
15708 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15710 * Avoid coredump with unsupported or invalid public keys by checking if
15711 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15712 PKCS7_verify() fails with non detached data.
15716 * Don't use getenv in library functions when run as setuid/setgid.
15717 New function OPENSSL_issetugid().
15721 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15722 due to incorrect handling of multi-threading:
15724 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15726 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15728 3. Count how many times MemCheck_off() has been called so that
15729 nested use can be treated correctly. This also avoids
15730 inband-signalling in the previous code (which relied on the
15731 assumption that thread ID 0 is impossible).
15735 * Add "-rand" option also to s_client and s_server.
15739 * Fix CPU detection on Irix 6.x.
15740 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15741 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15743 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15748 *This change does not apply to 0.9.7.*
15750 * Use the cached encoding of an X509_NAME structure rather than
15751 copying it. This is apparently the reason for the libsafe "errors"
15752 but the code is actually correct.
15756 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15757 Bleichenbacher's DSA attack.
15758 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15759 to be set and top=0 forces the highest bit to be set; top=-1 is new
15760 and leaves the highest bit random.
15762 *Ulf Moeller, Bodo Moeller*
15764 * In the `NCONF_...`-based implementations for `CONF_...` queries
15765 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15766 a temporary CONF structure with the data component set to NULL
15767 (which gives segmentation faults in lh_retrieve).
15768 Instead, use NULL for the CONF pointer in CONF_get_string and
15769 CONF_get_number (which may use environment variables) and directly
15770 return NULL from CONF_get_section.
15774 * Fix potential buffer overrun for EBCDIC.
15778 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15779 keyUsage if basicConstraints absent for a CA.
15783 * Make SMIME_write_PKCS7() write mail header values with a format that
15784 is more generally accepted (no spaces before the semicolon), since
15785 some programs can't parse those values properly otherwise. Also make
15786 sure BIO's that break lines after each write do not create invalid
15791 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15792 macros previously used would not encode an empty SEQUENCE OF
15793 and break the signature.
15797 *This change does not apply to 0.9.7.*
15799 * Zero the premaster secret after deriving the master secret in
15804 * Add some EVP_add_digest_alias registrations (as found in
15805 OpenSSL_add_all_digests()) to SSL_library_init()
15806 aka OpenSSL_add_ssl_algorithms(). This provides improved
15807 compatibility with peers using X.509 certificates
15808 with unconventional AlgorithmIdentifier OIDs.
15812 * Fix for Irix with NO_ASM.
15814 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15816 * ./config script fixes.
15818 *Ulf Moeller, Richard Levitte*
15820 * Fix 'openssl passwd -1'.
15824 * Change PKCS12_key_gen_asc() so it can cope with non null
15825 terminated strings whose length is passed in the passlen
15826 parameter, for example from PEM callbacks. This was done
15827 by adding an extra length parameter to asc2uni().
15829 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15831 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15832 call failed, free the DSA structure.
15836 * Fix to uni2asc() to cope with zero length Unicode strings.
15837 These are present in some PKCS#12 files.
15841 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15842 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15843 when writing a 32767 byte record.
15845 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15847 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15848 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15850 (RSA objects have a reference count access to which is protected
15851 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15852 so they are meant to be shared between threads.)
15853 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15854 "Reddie, Steven" <Steven.Reddie@ca.com>*
15856 * Fix a deadlock in CRYPTO_mem_leaks().
15860 * Use better test patterns in bntest.
15864 * rand_win.c fix for Borland C.
15868 * BN_rshift bugfix for n == 0.
15872 * Add a 'bctest' script that checks for some known 'bc' bugs
15873 so that 'make test' does not abort just because 'bc' is broken.
15877 * Store verify_result within SSL_SESSION also for client side to
15878 avoid potential security hole. (Re-used sessions on the client side
15879 always resulted in verify_result==X509_V_OK, not using the original
15880 result of the server certificate verification.)
15884 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15885 SSL3_RT_APPLICATION_DATA, return 0.
15886 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15891 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15892 releases, have been re-implemented by renaming the previous
15893 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15894 and ssl3_read_internal, respectively, and adding 'peek' parameters
15895 to them. The new ssl[23]_{read,peek} functions are calls to
15896 ssl[23]_read_internal with the 'peek' flag set appropriately.
15897 A 'peek' parameter has also been added to ssl3_read_bytes, which
15898 does the actual work for ssl3_read_internal.
15902 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15903 the method-specific "init()" handler. Also clean up ex_data after
15904 calling the method-specific "finish()" handler. Previously, this was
15905 happening the other way round.
15909 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15910 The previous value, 12, was not always sufficient for BN_mod_exp().
15914 * Make sure that shared libraries get the internal name engine with
15915 the full version number and not just 0. This should mark the
15916 shared libraries as not backward compatible. Of course, this should
15917 be changed again when we can guarantee backward binary compatibility.
15921 * Fix typo in get_cert_by_subject() in by_dir.c
15923 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15925 * Rework the system to generate shared libraries:
15927 - Make note of the expected extension for the shared libraries and
15928 if there is a need for symbolic links from for example libcrypto.so.0
15929 to libcrypto.so.0.9.7. There is extended info in Configure for
15932 - Make as few rebuilds of the shared libraries as possible.
15934 - Still avoid linking the OpenSSL programs with the shared libraries.
15936 - When installing, install the shared libraries separately from the
15941 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15943 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15944 and not in SSL_clear because the latter is also used by the
15945 accept/connect functions; previously, the settings made by
15946 SSL_set_read_ahead would be lost during the handshake.
15948 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15950 * Correct util/mkdef.pl to be selective about disabled algorithms.
15951 Previously, it would create entries for disabled algorithms no
15956 * Added several new manual pages for SSL_* function.
15960 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15962 * In ssl23_get_client_hello, generate an error message when faced
15963 with an initial SSL 3.0/TLS record that is too small to contain the
15964 first two bytes of the ClientHello message, i.e. client_version.
15965 (Note that this is a pathologic case that probably has never happened
15966 in real life.) The previous approach was to use the version number
15967 from the record header as a substitute; but our protocol choice
15968 should not depend on that one because it is not authenticated
15969 by the Finished messages.
15973 * More robust randomness gathering functions for Windows.
15975 *Jeffrey Altman <jaltman@columbia.edu>*
15977 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15978 not set then we don't setup the error code for issuer check errors
15979 to avoid possibly overwriting other errors which the callback does
15980 handle. If an application does set the flag then we assume it knows
15981 what it is doing and can handle the new informational codes
15986 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15987 a general "ANY" type, as such it should be able to decode anything
15988 including tagged types. However it didn't check the class so it would
15989 wrongly interpret tagged types in the same way as their universal
15990 counterpart and unknown types were just rejected. Changed so that the
15991 tagged and unknown types are handled in the same way as a SEQUENCE:
15992 that is the encoding is stored intact. There is also a new type
15993 "V_ASN1_OTHER" which is used when the class is not universal, in this
15994 case we have no idea what the actual type is so we just lump them all
15999 * On VMS, stdout may very well lead to a file that is written to
16000 in a record-oriented fashion. That means that every write() will
16001 write a separate record, which will be read separately by the
16002 programs trying to read from it. This can be very confusing.
16004 The solution is to put a BIO filter in the way that will buffer
16005 text until a linefeed is reached, and then write everything a
16006 line at a time, so every record written will be an actual line,
16007 not chunks of lines and not (usually doesn't happen, but I've
16008 seen it once) several lines in one record. BIO_f_linebuffer() is
16011 Currently, it's a VMS-only method, because that's where it has
16012 been tested well enough.
16016 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
16017 it can return incorrect results.
16018 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
16019 but it was in 0.9.6-beta[12].)
16023 * Disable the check for content being present when verifying detached
16024 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
16025 include zero length content when signing messages.
16029 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
16030 BIO_ctrl (for BIO pairs).
16034 * Add DSO method for VMS.
16038 * Bug fix: Montgomery multiplication could produce results with the
16043 * Add RPM specification openssl.spec and modify it to build three
16044 packages. The default package contains applications, application
16045 documentation and run-time libraries. The devel package contains
16046 include files, static libraries and function documentation. The
16047 doc package contains the contents of the doc directory. The original
16048 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
16052 * Add a large number of documentation files for many SSL routines.
16054 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16056 * Add a configuration entry for Sony News 4.
16058 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
16060 * Don't set the two most significant bits to one when generating a
16061 random number < q in the DSA library.
16065 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
16066 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
16067 the underlying transport is blocking) if a handshake took place.
16068 (The default behaviour is needed by applications such as s_client
16069 and s_server that use select() to determine when to use SSL_read;
16070 but for applications that know in advance when to expect data, it
16071 just makes things more complicated.)
16075 * Add RAND_egd_bytes(), which gives control over the number of bytes read
16080 * Add a few more EBCDIC conditionals that make `req` and `x509`
16081 work better on such systems.
16083 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16085 * Add two demo programs for PKCS12_parse() and PKCS12_create().
16086 Update PKCS12_parse() so it copies the friendlyName and the
16087 keyid to the certificates aux info.
16091 * Fix bug in PKCS7_verify() which caused an infinite loop
16092 if there was more than one signature.
16094 *Sven Uszpelkat <su@celocom.de>*
16096 * Major change in util/mkdef.pl to include extra information
16097 about each symbol, as well as presenting variables as well
16098 as functions. This change means that there's n more need
16099 to rebuild the .num files when some algorithms are excluded.
16103 * Allow the verify time to be set by an application,
16104 rather than always using the current time.
16108 * Phase 2 verify code reorganisation. The certificate
16109 verify code now looks up an issuer certificate by a
16110 number of criteria: subject name, authority key id
16111 and key usage. It also verifies self signed certificates
16112 by the same criteria. The main comparison function is
16113 X509_check_issued() which performs these checks.
16115 Lot of changes were necessary in order to support this
16116 without completely rewriting the lookup code.
16118 Authority and subject key identifier are now cached.
16120 The LHASH 'certs' is X509_STORE has now been replaced
16121 by a STACK_OF(X509_OBJECT). This is mainly because an
16122 LHASH can't store or retrieve multiple objects with
16123 the same hash value.
16125 As a result various functions (which were all internal
16126 use only) have changed to handle the new X509_STORE
16127 structure. This will break anything that messed round
16128 with X509_STORE internally.
16130 The functions X509_STORE_add_cert() now checks for an
16131 exact match, rather than just subject name.
16133 The X509_STORE API doesn't directly support the retrieval
16134 of multiple certificates matching a given criteria, however
16135 this can be worked round by performing a lookup first
16136 (which will fill the cache with candidate certificates)
16137 and then examining the cache for matches. This is probably
16138 the best we can do without throwing out X509_LOOKUP
16139 entirely (maybe later...).
16141 The X509_VERIFY_CTX structure has been enhanced considerably.
16143 All certificate lookup operations now go via a get_issuer()
16144 callback. Although this currently uses an X509_STORE it
16145 can be replaced by custom lookups. This is a simple way
16146 to bypass the X509_STORE hackery necessary to make this
16147 work and makes it possible to use more efficient techniques
16148 in future. A very simple version which uses a simple
16149 STACK for its trusted certificate store is also provided
16150 using X509_STORE_CTX_trusted_stack().
16152 The verify_cb() and verify() callbacks now have equivalents
16153 in the X509_STORE_CTX structure.
16155 X509_STORE_CTX also has a 'flags' field which can be used
16156 to customise the verify behaviour.
16160 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
16161 excludes S/MIME capabilities.
16165 * When a certificate request is read in keep a copy of the
16166 original encoding of the signed data and use it when outputting
16167 again. Signatures then use the original encoding rather than
16168 a decoded, encoded version which may cause problems if the
16169 request is improperly encoded.
16173 * For consistency with other BIO_puts implementations, call
16174 buffer_write(b, ...) directly in buffer_puts instead of calling
16177 In BIO_puts, increment b->num_write as in BIO_write.
16179 *Peter.Sylvester@EdelWeb.fr*
16181 * Fix BN_mul_word for the case where the word is 0. (We have to use
16182 BN_zero, we may not return a BIGNUM with an array consisting of
16183 words set to zero.)
16187 * Avoid calling abort() from within the library when problems are
16188 detected, except if preprocessor symbols have been defined
16189 (such as REF_CHECK, BN_DEBUG etc.).
16193 * New openssl application 'rsautl'. This utility can be
16194 used for low-level RSA operations. DER public key
16195 BIO/fp routines also added.
16199 * New Configure entry and patches for compiling on QNX 4.
16201 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16203 * A demo state-machine implementation was sponsored by
16204 Nuron (<http://www.nuron.com/>) and is now available in
16205 demos/state_machine.
16209 * New options added to the 'dgst' utility for signature
16210 generation and verification.
16214 * Unrecognized PKCS#7 content types are now handled via a
16215 catch all ASN1_TYPE structure. This allows unsupported
16216 types to be stored as a "blob" and an application can
16217 encode and decode it manually.
16221 * Fix various signed/unsigned issues to make a_strex.c
16222 compile under VC++.
16224 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16226 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16227 length if passed a buffer. ASN1_INTEGER_to_BN failed
16228 if passed a NULL BN and its argument was negative.
16230 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16232 * Modification to PKCS#7 encoding routines to output definite
16233 length encoding. Since currently the whole structures are in
16234 memory there's not real point in using indefinite length
16235 constructed encoding. However if OpenSSL is compiled with
16236 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16240 * Added BIO_vprintf() and BIO_vsnprintf().
16244 * Added more prefixes to parse for in the strings written
16245 through a logging bio, to cover all the levels that are available
16246 through syslog. The prefixes are now:
16248 PANIC, EMERG, EMR => LOG_EMERG
16249 ALERT, ALR => LOG_ALERT
16250 CRIT, CRI => LOG_CRIT
16251 ERROR, ERR => LOG_ERR
16252 WARNING, WARN, WAR => LOG_WARNING
16253 NOTICE, NOTE, NOT => LOG_NOTICE
16254 INFO, INF => LOG_INFO
16255 DEBUG, DBG => LOG_DEBUG
16257 and as before, if none of those prefixes are present at the
16258 beginning of the string, LOG_ERR is chosen.
16260 On Win32, the `LOG_*` levels are mapped according to this:
16262 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16263 LOG_WARNING => EVENTLOG_WARNING_TYPE
16264 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16268 * Made it possible to reconfigure with just the configuration
16269 argument "reconf" or "reconfigure". The command line arguments
16270 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16271 and are retrieved from there when reconfiguring.
16277 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16279 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16283 * The obj_dat.pl script was messing up the sorting of object
16284 names. The reason was that it compared the quoted version
16285 of strings as a result "OCSP" > "OCSP Signing" because
16286 " > SPACE. Changed script to store unquoted versions of
16287 names and add quotes on output. It was also omitting some
16288 names from the lookup table if they were given a default
16289 value (that is if SN is missing it is given the same
16290 value as LN and vice versa), these are now added on the
16291 grounds that if an object has a name we should be able to
16292 look it up. Finally added warning output when duplicate
16293 short or long names are found.
16297 * Changes needed for Tandem NSK.
16299 *Scott Uroff <scott@xypro.com>*
16301 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16302 RSA_padding_check_SSLv23(), special padding was never detected
16303 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16304 version rollback attacks was not effective.
16306 In s23_clnt.c, don't use special rollback-attack detection padding
16307 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16308 client; similarly, in s23_srvr.c, don't do the rollback check if
16309 SSL 2.0 is the only protocol enabled in the server.
16313 * Make it possible to get hexdumps of unprintable data with 'openssl
16314 asn1parse'. By implication, the functions ASN1_parse_dump() and
16315 BIO_dump_indent() are added.
16319 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16320 these print out strings and name structures based on various
16321 flags including RFC2253 support and proper handling of
16322 multibyte characters. Added options to the 'x509' utility
16323 to allow the various flags to be set.
16327 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16328 Also change the functions X509_cmp_current_time() and
16329 X509_gmtime_adj() work with an ASN1_TIME structure,
16330 this will enable certificates using GeneralizedTime in validity
16331 dates to be checked.
16335 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16336 negative public key encodings) on by default,
16337 NO_NEG_PUBKEY_BUG can be set to disable it.
16341 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16342 content octets. An i2c_ASN1_OBJECT is unnecessary because
16343 the encoding can be trivially obtained from the structure.
16347 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16348 not read locks (`CRYPTO_r_[un]lock`).
16352 * A first attempt at creating official support for shared
16353 libraries through configuration. I've kept it so the
16354 default is static libraries only, and the OpenSSL programs
16355 are always statically linked for now, but there are
16356 preparations for dynamic linking in place.
16357 This has been tested on Linux and Tru64.
16361 * Randomness polling function for Win9x, as described in:
16362 Peter Gutmann, Software Generation of Practically Strong
16367 * Fix so PRNG is seeded in req if using an already existing
16372 * New options to smime application. -inform and -outform
16373 allow alternative formats for the S/MIME message including
16374 PEM and DER. The -content option allows the content to be
16375 specified separately. This should allow things like Netscape
16376 form signing output easier to verify.
16380 * Fix the ASN1 encoding of tags using the 'long form'.
16384 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16385 STRING types. These convert content octets to and from the
16386 underlying type. The actual tag and length octets are
16387 already assumed to have been read in and checked. These
16388 are needed because all other string types have virtually
16389 identical handling apart from the tag. By having versions
16390 of the ASN1 functions that just operate on content octets
16391 IMPLICIT tagging can be handled properly. It also allows
16392 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16393 and ASN1_INTEGER are identical apart from the tag.
16397 * Change the handling of OID objects as follows:
16399 - New object identifiers are inserted in objects.txt, following
16400 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16401 - objects.pl is used to process obj_mac.num and create a new
16403 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16406 This is currently kind of a hack, and the perl code in objects.pl
16407 isn't very elegant, but it works as I intended. The simplest way
16408 to check that it worked correctly is to look in obj_dat.h and
16409 check the array nid_objs and make sure the objects haven't moved
16410 around (this is important!). Additions are OK, as well as
16411 consistent name changes.
16415 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16419 * Addition of the command line parameter '-rand file' to 'openssl req'.
16420 The given file adds to whatever has already been seeded into the
16421 random pool through the RANDFILE configuration file option or
16422 environment variable, or the default random state file.
16426 * mkstack.pl now sorts each macro group into lexical order.
16427 Previously the output order depended on the order the files
16428 appeared in the directory, resulting in needless rewriting
16433 * Patches to make OpenSSL compile under Win32 again. Mostly
16434 work arounds for the VC++ problem that it treats func() as
16435 func(void). Also stripped out the parts of mkdef.pl that
16436 added extra typesafe functions: these no longer exist.
16440 * Reorganisation of the stack code. The macros are now all
16441 collected in safestack.h . Each macro is defined in terms of
16442 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16443 DEBUG_SAFESTACK is now handled in terms of function casts,
16444 this has the advantage of retaining type safety without the
16445 use of additional functions. If DEBUG_SAFESTACK is not defined
16446 then the non typesafe macros are used instead. Also modified the
16447 mkstack.pl script to handle the new form. Needs testing to see
16448 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16449 the default if no major problems. Similar behaviour for ASN1_SET_OF
16450 and PKCS12_STACK_OF.
16454 * When some versions of IIS use the 'NET' form of private key the
16455 key derivation algorithm is different. Normally MD5(password) is
16456 used as a 128 bit RC4 key. In the modified case
16457 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16458 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16459 as the old Netscape_RSA functions except they have an additional
16460 'sgckey' parameter which uses the modified algorithm. Also added
16461 an -sgckey command line option to the rsa utility. Thanks to
16462 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16463 algorithm to openssl-dev.
16467 * The evp_local.h macros were using 'c.##kname' which resulted in
16468 invalid expansion on some systems (SCO 5.0.5 for example).
16469 Corrected to 'c.kname'.
16471 *Phillip Porch <root@theporch.com>*
16473 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16474 a STACK of email addresses from a certificate or request, these look
16475 in the subject name and the subject alternative name extensions and
16476 omit any duplicate addresses.
16480 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16481 This makes DSA verification about 2 % faster.
16485 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16486 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16487 plus overhead for 1024 bit moduli).
16488 This makes exponentiations about 0.5 % faster for 1024 bit
16489 exponents (as measured by "openssl speed rsa2048").
16493 * Rename memory handling macros to avoid conflicts with other
16495 Malloc => OPENSSL_malloc
16496 Malloc_locked => OPENSSL_malloc_locked
16497 Realloc => OPENSSL_realloc
16498 Free => OPENSSL_free
16502 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16503 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16507 * CygWin32 support.
16509 *John Jarvie <jjarvie@newsguy.com>*
16511 * The type-safe stack code has been rejigged. It is now only compiled
16512 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16513 by default all type-specific stack functions are "#define"d back to
16514 standard stack functions. This results in more streamlined output
16515 but retains the type-safety checking possibilities of the original
16520 * The STACK code has been cleaned up, and certain type declarations
16521 that didn't make a lot of sense have been brought in line. This has
16522 also involved a cleanup of sorts in safestack.h to more correctly
16523 map type-safe stack functions onto their plain stack counterparts.
16524 This work has also resulted in a variety of "const"ifications of
16525 lots of the code, especially `_cmp` operations which should normally
16526 be prototyped with "const" parameters anyway.
16530 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16531 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16532 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16533 where all of 'md' is used each time the PRNG is used, but 'state'
16534 is used only indexed by a cyclic counter. As entropy may not be
16535 well distributed from the beginning, 'md' is important as a
16536 chaining variable. However, the output function chains only half
16537 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16538 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16539 in all of 'state' being rewritten, with the new values depending
16540 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16544 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16545 the handshake is continued after ssl_verify_cert_chain();
16546 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16547 can lead to 'unexplainable' connection aborts later.
16549 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16551 * Major EVP API cipher revision.
16552 Add hooks for extra EVP features. This allows various cipher
16553 parameters to be set in the EVP interface. Support added for variable
16554 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16555 setting of RC2 and RC5 parameters.
16557 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16560 Remove lots of duplicated code from the EVP library. For example *every*
16561 cipher init() function handles the 'iv' in the same way according to the
16562 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16563 for CFB and OFB modes they zero ctx->num.
16565 New functionality allows removal of S/MIME code RC2 hack.
16567 Most of the routines have the same form and so can be declared in terms
16570 By shifting this to the top level EVP_CipherInit() it can be removed from
16571 all individual ciphers. If the cipher wants to handle IVs or keys
16572 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16575 Change lots of functions like EVP_EncryptUpdate() to now return a
16576 value: although software versions of the algorithms cannot fail
16577 any installed hardware versions can.
16581 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16582 this option is set, tolerate broken clients that send the negotiated
16583 protocol version number instead of the requested protocol version
16588 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16589 i.e. non-zero for export ciphersuites, zero otherwise.
16590 Previous versions had this flag inverted, inconsistent with
16591 rsa_tmp_cb (..._TMP_RSA_CB).
16593 *Bodo Moeller; problem reported by Amit Chopra*
16595 * Add missing DSA library text string. Work around for some IIS
16596 key files with invalid SEQUENCE encoding.
16600 * Add a document (doc/standards.txt) that list all kinds of standards
16601 and so on that are implemented in OpenSSL.
16605 * Enhance c_rehash script. Old version would mishandle certificates
16606 with the same subject name hash and wouldn't handle CRLs at all.
16607 Added -fingerprint option to crl utility, to support new c_rehash
16612 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16616 * Fix for SSL server purpose checking. Server checking was
16617 rejecting certificates which had extended key usage present
16618 but no ssl client purpose.
16620 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16622 * Make PKCS#12 code work with no password. The PKCS#12 spec
16623 is a little unclear about how a blank password is handled.
16624 Since the password in encoded as a BMPString with terminating
16625 double NULL a zero length password would end up as just the
16626 double NULL. However no password at all is different and is
16627 handled differently in the PKCS#12 key generation code. NS
16628 treats a blank password as zero length. MSIE treats it as no
16629 password on export: but it will try both on import. We now do
16630 the same: PKCS12_parse() tries zero length and no password if
16631 the password is set to "" or NULL (NULL is now a valid password:
16632 it wasn't before) as does the pkcs12 application.
16636 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16637 perror when PEM_read_bio_X509_REQ fails, the error message must
16638 be obtained from the error queue.
16642 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16643 it in ERR_remove_state if appropriate, and change ERR_get_state
16644 accordingly to avoid race conditions (this is necessary because
16645 thread_hash is no longer constant once set).
16649 * Bugfix for linux-elf makefile.one.
16653 * RSA_get_default_method() will now cause a default
16654 RSA_METHOD to be chosen if one doesn't exist already.
16655 Previously this was only set during a call to RSA_new()
16656 or RSA_new_method(NULL) meaning it was possible for
16657 RSA_get_default_method() to return NULL.
16661 * Added native name translation to the existing DSO code
16662 that will convert (if the flag to do so is set) filenames
16663 that are sufficiently small and have no path information
16664 into a canonical native form. Eg. "blah" converted to
16665 "libblah.so" or "blah.dll" etc.
16669 * New function ERR_error_string_n(e, buf, len) which is like
16670 ERR_error_string(e, buf), but writes at most 'len' bytes
16671 including the 0 terminator. For ERR_error_string_n, 'buf'
16674 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16676 * CONF library reworked to become more general. A new CONF
16677 configuration file reader "class" is implemented as well as a
16678 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16679 old `CONF_*` functions are still there, but are reimplemented to
16680 work in terms of the new functions. Also, a set of functions
16681 to handle the internal storage of the configuration data is
16682 provided to make it easier to write new configuration file
16683 reader "classes" (I can definitely see something reading a
16684 configuration file in XML format, for example), called `_CONF_*`,
16685 or "the configuration storage API"...
16687 The new configuration file reading functions are:
16689 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16690 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16692 NCONF_default, NCONF_WIN32
16694 NCONF_dump_fp, NCONF_dump_bio
16696 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16697 NCONF_new creates a new CONF object. This works in the same way
16698 as other interfaces in OpenSSL, like the BIO interface.
16699 `NCONF_dump_*` dump the internal storage of the configuration file,
16700 which is useful for debugging. All other functions take the same
16701 arguments as the old `CONF_*` functions with the exception of the
16702 first that must be a `CONF *` instead of a `LHASH *`.
16704 To make it easier to use the new classes with the old `CONF_*` functions,
16705 the function CONF_set_default_method is provided.
16709 * Add '-tls1' option to 'openssl ciphers', which was already
16710 mentioned in the documentation but had not been implemented.
16711 (This option is not yet really useful because even the additional
16712 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16716 * Initial DSO code added into libcrypto for letting OpenSSL (and
16717 OpenSSL-based applications) load shared libraries and bind to
16718 them in a portable way.
16720 *Geoff Thorpe, with contributions from Richard Levitte*
16722 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16724 * Make sure _lrotl and _lrotr are only used with MSVC.
16726 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16727 (the default implementation of RAND_status).
16729 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16730 to '-clrext' (= clear extensions), as intended and documented.
16731 *Bodo Moeller; inconsistency pointed out by Michael Attili
16732 <attili@amaxo.com>*
16734 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16735 was larger than the MD block size.
16737 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16739 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16740 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16741 using the passed key: if the passed key was a private key the result
16742 of X509_print(), for example, would be to print out all the private key
16747 * des_quad_cksum() byte order bug fix.
16748 *Ulf Möller, using the problem description in krb4-0.9.7, where
16749 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16751 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16754 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16756 * For easily testing in shell scripts whether some command
16757 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16758 returns with exit code 0 iff no command of the given name is available.
16759 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16760 the output goes to stdout and nothing is printed to stderr.
16761 Additional arguments are always ignored.
16763 Since for each cipher there is a command of the same name,
16764 the 'no-cipher' compilation switches can be tested this way.
16766 ('openssl no-XXX' is not able to detect pseudo-commands such
16767 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16771 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16775 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16776 is set; it will be thrown away anyway because each handshake creates
16778 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16779 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16780 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16781 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16785 * New s_client option -ign_eof: EOF at stdin is ignored, and
16786 'Q' and 'R' lose their special meanings (quit/renegotiate).
16787 This is part of what -quiet does; unlike -quiet, -ign_eof
16788 does not suppress any output.
16792 * Add compatibility options to the purpose and trust code. The
16793 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16794 accepts a certificate or CA, this was the previous behaviour,
16795 with all the associated security issues.
16797 X509_TRUST_COMPAT is the old trust behaviour: only and
16798 automatically trust self signed roots in certificate store. A
16799 new trust setting X509_TRUST_DEFAULT is used to specify that
16800 a purpose has no associated trust setting and it should instead
16801 use the value in the default purpose.
16805 * Fix the PKCS#8 DSA private key code so it decodes keys again
16806 and fix a memory leak.
16810 * In util/mkerr.pl (which implements 'make errors'), preserve
16811 reason strings from the previous version of the .c file, as
16812 the default to have only downcase letters (and digits) in
16813 automatically generated reasons codes is not always appropriate.
16817 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16818 using strerror. Previously, ERR_reason_error_string() returned
16819 library names as reason strings for SYSerr; but SYSerr is a special
16820 case where small numbers are errno values, not library numbers.
16824 * Add '-dsaparam' option to 'openssl dhparam' application. This
16825 converts DSA parameters into DH parameters. (When creating parameters,
16826 DSA_generate_parameters is used.)
16830 * Include 'length' (recommended exponent length) in C code generated
16831 by 'openssl dhparam -C'.
16835 * The second argument to set_label in perlasm was already being used
16836 so couldn't be used as a "file scope" flag. Moved to third argument
16841 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16842 instead of RAND_bytes for encryption IVs and salts.
16846 * Include RAND_status() into RAND_METHOD instead of implementing
16847 it only for md_rand.c Otherwise replacing the PRNG by calling
16848 RAND_set_rand_method would be impossible.
16852 * Don't let DSA_generate_key() enter an infinite loop if the random
16853 number generation fails.
16857 * New 'rand' application for creating pseudo-random output.
16861 * Added configuration support for Linux/IA64
16863 *Rolf Haberrecker <rolf@suse.de>*
16865 * Assembler module support for Mingw32.
16869 * Shared library support for HPUX (in shlib/).
16871 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16873 * Shared library support for Solaris gcc.
16875 *Lutz Behnke <behnke@trustcenter.de>*
16877 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16879 * PKCS7_encrypt() was adding text MIME headers twice because they
16880 were added manually and by SMIME_crlf_copy().
16884 * In bntest.c don't call BN_rand with zero bits argument.
16886 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16888 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16889 case was implemented. This caused BN_div_recp() to fail occasionally.
16893 * Add an optional second argument to the set_label() in the perl
16894 assembly language builder. If this argument exists and is set
16895 to 1 it signals that the assembler should use a symbol whose
16896 scope is the entire file, not just the current function. This
16897 is needed with MASM which uses the format label:: for this scope.
16899 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16901 * Change the ASN1 types so they are typedefs by default. Before
16902 almost all types were #define'd to ASN1_STRING which was causing
16903 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16908 * Change names of new functions to the new get1/get0 naming
16909 convention: After 'get1', the caller owns a reference count
16910 and has to call `..._free`; 'get0' returns a pointer to some
16911 data structure without incrementing reference counters.
16912 (Some of the existing 'get' functions increment a reference
16913 counter, some don't.)
16914 Similarly, 'set1' and 'add1' functions increase reference
16915 counters or duplicate objects.
16919 * Allow for the possibility of temp RSA key generation failure:
16920 the code used to assume it always worked and crashed on failure.
16924 * Fix potential buffer overrun problem in BIO_printf().
16925 *Ulf Möller, using public domain code by Patrick Powell; problem
16926 pointed out by David Sacerdote <das33@cornell.edu>*
16928 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16929 RAND_egd() and RAND_status(). In the command line application,
16930 the EGD socket can be specified like a seed file using RANDFILE
16935 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16936 Some CAs (e.g. Verisign) distribute certificates in this form.
16940 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16941 list to exclude them. This means that no special compilation option
16942 is needed to use anonymous DH: it just needs to be included in the
16947 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16948 EVP_MD_type. The old functionality is available in a new macro called
16949 EVP_MD_md(). Change code that uses it and update docs.
16953 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16954 where the `void *` argument is replaced by a function pointer argument.
16955 Previously `void *` was abused to point to functions, which works on
16956 many platforms, but is not correct. As these functions are usually
16957 called by macros defined in OpenSSL header files, most source code
16958 should work without changes.
16962 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16963 sections with information on -D... compiler switches used for
16964 compiling the library so that applications can see them. To enable
16965 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16966 must be defined. E.g.,
16967 #define OPENSSL_ALGORITHM_DEFINES
16968 #include <openssl/opensslconf.h>
16969 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16971 *Richard Levitte, Ulf and Bodo Möller*
16973 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16978 * Change the 'other' type in certificate aux info to a STACK_OF
16979 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16980 the required ASN1 format: arbitrary types determined by an OID.
16984 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16985 argument to 'req'. This is not because the function is newer or
16986 better than others it just uses the work 'NEW' in the certificate
16987 request header lines. Some software needs this.
16991 * Reorganise password command line arguments: now passwords can be
16992 obtained from various sources. Delete the PEM_cb function and make
16993 it the default behaviour: i.e. if the callback is NULL and the
16994 usrdata argument is not NULL interpret it as a null terminated pass
16995 phrase. If usrdata and the callback are NULL then the pass phrase
16996 is prompted for as usual.
17000 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
17001 the support is automatically enabled. The resulting binaries will
17002 autodetect the card and use it if present.
17004 *Ben Laurie and Compaq Inc.*
17006 * Work around for Netscape hang bug. This sends certificate request
17007 and server done in one record. Since this is perfectly legal in the
17008 SSL/TLS protocol it isn't a "bug" option and is on by default. See
17009 the bugs/SSLv3 entry for more info.
17013 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
17017 * Add -rand argument to smime and pkcs12 applications and read/write
17022 * New 'passwd' tool for crypt(3) and apr1 password hashes.
17026 * Add command line password options to the remaining applications.
17030 * Bug fix for BN_div_recp() for numerators with an even number of
17035 * More tests in bntest.c, and changed test_bn output.
17039 * ./config recognizes MacOS X now.
17043 * Bug fix for BN_div() when the first words of num and divisor are
17044 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
17048 * Add support for various broken PKCS#8 formats, and command line
17049 options to produce them.
17053 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
17054 get temporary BIGNUMs from a BN_CTX.
17058 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
17063 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
17064 include a #define from the old name to the new. The original intent
17065 was that statically linked binaries could for example just call
17066 SSLeay_add_all_ciphers() to just add ciphers to the table and not
17067 link with digests. This never worked because SSLeay_add_all_digests()
17068 and SSLeay_add_all_ciphers() were in the same source file so calling
17069 one would link with the other. They are now in separate source files.
17073 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
17077 * Use a less unusual form of the Miller-Rabin primality test (it used
17078 a binary algorithm for exponentiation integrated into the Miller-Rabin
17079 loop, our standard modexp algorithms are faster).
17083 * Support for the EBCDIC character set completed.
17085 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
17087 * Source code cleanups: use const where appropriate, eliminate casts,
17088 use `void *` instead of `char *` in lhash.
17092 * Bugfix: ssl3_send_server_key_exchange was not restartable
17093 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
17094 this the server could overwrite ephemeral keys that the client
17099 * Turn DSA_is_prime into a macro that calls BN_is_prime,
17100 using 50 iterations of the Rabin-Miller test.
17102 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
17103 iterations of the Rabin-Miller test as required by the appendix
17104 to FIPS PUB 186[-1]) instead of DSA_is_prime.
17105 As BN_is_prime_fasttest includes trial division, DSA parameter
17106 generation becomes much faster.
17108 This implies a change for the callback functions in DSA_is_prime
17109 and DSA_generate_parameters: The callback function is called once
17110 for each positive witness in the Rabin-Miller test, not just
17111 occasionally in the inner loop; and the parameters to the
17112 callback function now provide an iteration count for the outer
17113 loop rather than for the current invocation of the inner loop.
17114 DSA_generate_parameters additionally can call the callback
17115 function with an 'iteration count' of -1, meaning that a
17116 candidate has passed the trial division test (when q is generated
17117 from an application-provided seed, trial division is skipped).
17121 * New function BN_is_prime_fasttest that optionally does trial
17122 division before starting the Rabin-Miller test and has
17123 an additional BN_CTX * argument (whereas BN_is_prime always
17124 has to allocate at least one BN_CTX).
17125 'callback(1, -1, cb_arg)' is called when a number has passed the
17126 trial division stage.
17130 * Fix for bug in CRL encoding. The validity dates weren't being handled
17135 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
17139 * New function BN_pseudo_rand().
17143 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
17144 bignum version of BN_from_montgomery() with the working code from
17145 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
17150 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
17151 made it impossible to use the same SSL_SESSION data structure in
17152 SSL2 clients in multiple threads.
17156 * The return value of RAND_load_file() no longer counts bytes obtained
17157 by stat(). RAND_load_file(..., -1) is new and uses the complete file
17158 to seed the PRNG (previously an explicit byte count was required).
17160 *Ulf Möller, Bodo Möller*
17162 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
17163 used `char *` instead of `void *` and had casts all over the place.
17167 * Make BN_generate_prime() return NULL on error if ret!=NULL.
17171 * Retain source code compatibility for BN_prime_checks macro:
17172 BN_is_prime(..., BN_prime_checks, ...) now uses
17173 BN_prime_checks_for_size to determine the appropriate number of
17174 Rabin-Miller iterations.
17178 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
17179 DH_CHECK_P_NOT_SAFE_PRIME.
17180 (Check if this is true? OpenPGP calls them "strong".)
17184 * Merge the functionality of "dh" and "gendh" programs into a new program
17185 "dhparam". The old programs are retained for now but will handle DH keys
17186 (instead of parameters) in future.
17190 * Make the ciphers, s_server and s_client programs check the return values
17191 when a new cipher list is set.
17195 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17196 ciphers. Before when the 56bit ciphers were enabled the sorting was
17199 The syntax for the cipher sorting has been extended to support sorting by
17200 cipher-strength (using the strength_bits hard coded in the tables).
17201 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17203 Fix a bug in the cipher-command parser: when supplying a cipher command
17204 string with an "undefined" symbol (neither command nor alphanumeric
17205 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17206 an error is flagged.
17208 Due to the strength-sorting extension, the code of the
17209 ssl_create_cipher_list() function was completely rearranged. I hope that
17210 the readability was also increased :-)
17212 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17214 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17215 for the first serial number and places 2 in the serial number file. This
17216 avoids problems when the root CA is created with serial number zero and
17217 the first user certificate has the same issuer name and serial number
17222 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17223 the new code. Add documentation for this stuff.
17227 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17228 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17229 structures and behave in an analogous way to the X509v3 functions:
17230 they shouldn't be called directly but wrapper functions should be used
17233 So we also now have some wrapper functions that call the X509at functions
17234 when passed certificate requests. (TO DO: similar things can be done with
17235 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17236 things. Some of these need some d2i or i2d and print functionality
17237 because they handle more complex structures.)
17241 * Add missing #ifndefs that caused missing symbols when building libssl
17242 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17243 NO_RSA in `ssl/s2*.c`.
17245 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17247 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17248 has a return value which indicates the quality of the random data
17249 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17250 error queue. New function RAND_pseudo_bytes() generates output that is
17251 guaranteed to be unique but not unpredictable. RAND_add is like
17252 RAND_seed, but takes an extra argument for an entropy estimate
17253 (RAND_seed always assumes full entropy).
17257 * Do more iterations of Rabin-Miller probable prime test (specifically,
17258 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17259 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17260 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17261 false-positive rate of at most 2^-80 for random input.
17265 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17269 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17270 in the 0.9.5 release), this returns the chain
17271 from an X509_CTX structure with a dup of the stack and all
17272 the X509 reference counts upped: so the stack will exist
17273 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17276 Also make SSL_SESSION_print() print out the verify return
17281 * Add manpage for the pkcs12 command. Also change the default
17282 behaviour so MAC iteration counts are used unless the new
17283 -nomaciter option is used. This improves file security and
17284 only older versions of MSIE (4.0 for example) need it.
17288 * Honor the no-xxx Configure options when creating .DEF files.
17292 * Add PKCS#10 attributes to field table: challengePassword,
17293 unstructuredName and unstructuredAddress. These are taken from
17294 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17295 international characters are used.
17297 More changes to X509_ATTRIBUTE code: allow the setting of types
17298 based on strings. Remove the 'loc' parameter when adding
17299 attributes because these will be a SET OF encoding which is sorted
17304 * Initial changes to the 'req' utility to allow request generation
17305 automation. This will allow an application to just generate a template
17306 file containing all the field values and have req construct the
17309 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17310 used all over the place including certificate requests and PKCS#7
17311 structures. They are currently handled manually where necessary with
17312 some primitive wrappers for PKCS#7. The new functions behave in a
17313 manner analogous to the X509 extension functions: they allow
17314 attributes to be looked up by NID and added.
17316 Later something similar to the X509V3 code would be desirable to
17317 automatically handle the encoding, decoding and printing of the
17318 more complex types. The string types like challengePassword can
17319 be handled by the string table functions.
17321 Also modified the multi byte string table handling. Now there is
17322 a 'global mask' which masks out certain types. The table itself
17323 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17324 is useful when for example there is only one permissible type
17325 (as in countryName) and using the mask might result in no valid
17330 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17331 SSL_get_peer_finished to allow applications to obtain the latest
17332 Finished messages sent to the peer or expected from the peer,
17333 respectively. (SSL_get_peer_finished is usually the Finished message
17334 actually received from the peer, otherwise the protocol will be aborted.)
17336 As the Finished message are message digests of the complete handshake
17337 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17338 be used for external authentication procedures when the authentication
17339 provided by SSL/TLS is not desired or is not enough.
17343 * Enhanced support for Alpha Linux is added. Now ./config checks if
17344 the host supports BWX extension and if Compaq C is present on the
17345 $PATH. Just exploiting of the BWX extension results in 20-30%
17346 performance kick for some algorithms, e.g. DES and RC4 to mention
17347 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17352 * Add support for MS "fast SGC". This is arguably a violation of the
17353 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17354 weak crypto and after checking the certificate is SGC a second one
17355 with strong crypto. MS SGC stops the first handshake after receiving
17356 the server certificate message and sends a second client hello. Since
17357 a server will typically do all the time consuming operations before
17358 expecting any further messages from the client (server key exchange
17359 is the most expensive) there is little difference between the two.
17361 To get OpenSSL to support MS SGC we have to permit a second client
17362 hello message after we have sent server done. In addition we have to
17363 reset the MAC if we do get this second client hello.
17367 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17368 if a DER encoded private key is RSA or DSA traditional format. Changed
17369 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17370 format DER encoded private key. Newer code should use PKCS#8 format which
17371 has the key type encoded in the ASN1 structure. Added DER private key
17372 support to pkcs8 application.
17376 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17377 ciphersuites has been selected (as required by the SSL 3/TLS 1
17378 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17379 is set, we interpret this as a request to violate the specification
17380 (the worst that can happen is a handshake failure, and 'correct'
17381 behaviour would result in a handshake failure anyway).
17385 * In SSL_CTX_add_session, take into account that there might be multiple
17386 SSL_SESSION structures with the same session ID (e.g. when two threads
17387 concurrently obtain them from an external cache).
17388 The internal cache can handle only one SSL_SESSION with a given ID,
17389 so if there's a conflict, we now throw out the old one to achieve
17394 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17395 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17396 some routines that use cipher OIDs: some ciphers do not have OIDs
17397 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17402 * Simplify the trust setting structure and code. Now we just have
17403 two sequences of OIDs for trusted and rejected settings. These will
17404 typically have values the same as the extended key usage extension
17405 and any application specific purposes.
17407 The trust checking code now has a default behaviour: it will just
17408 check for an object with the same NID as the passed id. Functions can
17409 be provided to override either the default behaviour or the behaviour
17410 for a given id. SSL client, server and email already have functions
17411 in place for compatibility: they check the NID and also return "trusted"
17412 if the certificate is self signed.
17416 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17417 traditional format into an EVP_PKEY structure.
17421 * Add a password callback function PEM_cb() which either prompts for
17422 a password if usr_data is NULL or otherwise assumes it is a null
17423 terminated password. Allow passwords to be passed on command line
17424 environment or config files in a few more utilities.
17428 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17429 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17430 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17431 Update documentation.
17435 * Support for ASN1 "NULL" type. This could be handled before by using
17436 ASN1_TYPE but there wasn't any function that would try to read a NULL
17437 and produce an error if it couldn't. For compatibility we also have
17438 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17439 don't allocate anything because they don't need to.
17443 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17446 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17448 * Rebuild of the memory allocation routines used by OpenSSL code and
17449 possibly others as well. The purpose is to make an interface that
17450 provide hooks so anyone can build a separate set of allocation and
17451 deallocation routines to be used by OpenSSL, for example memory
17452 pool implementations, or something else, which was previously hard
17453 since Malloc(), Realloc() and Free() were defined as macros having
17454 the values malloc, realloc and free, respectively (except for Win32
17455 compilations). The same is provided for memory debugging code.
17456 OpenSSL already comes with functionality to find memory leaks, but
17457 this gives people a chance to debug other memory problems.
17459 With these changes, a new set of functions and macros have appeared:
17461 CRYPTO_set_mem_debug_functions() [F]
17462 CRYPTO_get_mem_debug_functions() [F]
17463 CRYPTO_dbg_set_options() [F]
17464 CRYPTO_dbg_get_options() [F]
17465 CRYPTO_malloc_debug_init() [M]
17467 The memory debug functions are NULL by default, unless the library
17468 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17469 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17470 gives the standard debugging functions that come with OpenSSL) or
17471 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17472 provided by the library user) must be used. When the standard
17473 debugging functions are used, CRYPTO_dbg_set_options can be used to
17474 request additional information:
17475 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17476 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17478 Also, things like CRYPTO_set_mem_functions will always give the
17479 expected result (the new set of functions is used for allocation
17480 and deallocation) at all times, regardless of platform and compiler
17483 To finish it up, some functions that were never use in any other
17484 way than through macros have a new API and new semantic:
17486 CRYPTO_dbg_malloc()
17487 CRYPTO_dbg_realloc()
17490 All macros of value have retained their old syntax.
17492 *Richard Levitte and Bodo Moeller*
17494 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17495 ordering of SMIMECapabilities wasn't in "strength order" and there
17496 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17501 * Some ASN1 types with illegal zero length encoding (INTEGER,
17502 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17504 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17506 * Merge in my S/MIME library for OpenSSL. This provides a simple
17507 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17508 functionality to handle multipart/signed properly) and a utility
17509 called 'smime' to call all this stuff. This is based on code I
17510 originally wrote for Celo who have kindly allowed it to be
17511 included in OpenSSL.
17515 * Add variants des_set_key_checked and des_set_key_unchecked of
17516 des_set_key (aka des_key_sched). Global variable des_check_key
17517 decides which of these is called by des_set_key; this way
17518 des_check_key behaves as it always did, but applications and
17519 the library itself, which was buggy for des_check_key == 1,
17520 have a cleaner way to pick the version they need.
17524 * New function PKCS12_newpass() which changes the password of a
17529 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17530 dynamic mix. In both cases the ids can be used as an index into the
17531 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17532 functions so they accept a list of the field values and the
17533 application doesn't need to directly manipulate the X509_TRUST
17538 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17543 * Modify the way the V3 extension code looks up extensions. This now
17544 works in a similar way to the object code: we have some "standard"
17545 extensions in a static table which is searched with OBJ_bsearch()
17546 and the application can add dynamic ones if needed. The file
17547 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17548 updated whenever a new extension is added to the core code and kept
17549 in ext_nid order. There is a simple program 'tabtest.c' which checks
17550 this. New extensions are not added too often so this file can readily
17551 be maintained manually.
17553 There are two big advantages in doing things this way. The extensions
17554 can be looked up immediately and no longer need to be "added" using
17555 X509V3_add_standard_extensions(): this function now does nothing.
17556 Side note: I get *lots* of email saying the extension code doesn't
17557 work because people forget to call this function.
17558 Also no dynamic allocation is done unless new extensions are added:
17559 so if we don't add custom extensions there is no need to call
17560 X509V3_EXT_cleanup().
17564 * Modify enc utility's salting as follows: make salting the default. Add a
17565 magic header, so unsalted files fail gracefully instead of just decrypting
17566 to garbage. This is because not salting is a big security hole, so people
17567 should be discouraged from doing it.
17571 * Fixes and enhancements to the 'x509' utility. It allowed a message
17572 digest to be passed on the command line but it only used this
17573 parameter when signing a certificate. Modified so all relevant
17574 operations are affected by the digest parameter including the
17575 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17576 DSA key was used because it didn't fix the digest.
17580 * Initial certificate chain verify code. Currently tests the untrusted
17581 certificates for consistency with the verify purpose (which is set
17582 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17584 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17585 this is because it will reject chains with invalid extensions whereas
17586 every previous version of OpenSSL and SSLeay made no checks at all.
17588 Trust code: checks the root CA for the relevant trust settings. Trust
17589 settings have an initial value consistent with the verify purpose: e.g.
17590 if the verify purpose is for SSL client use it expects the CA to be
17591 trusted for SSL client use. However the default value can be changed to
17592 permit custom trust settings: one example of this would be to only trust
17593 certificates from a specific "secure" set of CAs.
17595 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17596 which should be used for version portability: especially since the
17597 verify structure is likely to change more often now.
17599 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17600 to set them. If not set then assume SSL clients will verify SSL servers
17603 Two new options to the verify program: -untrusted allows a set of
17604 untrusted certificates to be passed in and -purpose which sets the
17605 intended purpose of the certificate. If a purpose is set then the
17606 new chain verify code is used to check extension consistency.
17610 * Support for the authority information access extension.
17614 * Modify RSA and DSA PEM read routines to transparently handle
17615 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17616 public keys in a format compatible with certificate
17617 SubjectPublicKeyInfo structures. Unfortunately there were already
17618 functions called *_PublicKey_* which used various odd formats so
17619 these are retained for compatibility: however the DSA variants were
17620 never in a public release so they have been deleted. Changed dsa/rsa
17621 utilities to handle the new format: note no releases ever handled public
17622 keys so we should be OK.
17624 The primary motivation for this change is to avoid the same fiasco
17625 that dogs private keys: there are several incompatible private key
17626 formats some of which are standard and some OpenSSL specific and
17627 require various evil hacks to allow partial transparent handling and
17628 even then it doesn't work with DER formats. Given the option anything
17629 other than PKCS#8 should be dumped: but the other formats have to
17630 stay in the name of compatibility.
17632 With public keys and the benefit of hindsight one standard format
17633 is used which works with EVP_PKEY, RSA or DSA structures: though
17634 it clearly returns an error if you try to read the wrong kind of key.
17636 Added a -pubkey option to the 'x509' utility to output the public key.
17637 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17638 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17639 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17640 that do the same as the `EVP_PKEY_assign_*()` except they up the
17641 reference count of the added key (they don't "swallow" the
17646 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17647 CRLs would fail if the file contained no certificates or no CRLs:
17648 added a new function to read in both types and return the number
17649 read: this means that if none are read it will be an error. The
17650 DER versions of the certificate and CRL reader would always fail
17651 because it isn't possible to mix certificates and CRLs in DER format
17652 without choking one or the other routine. Changed this to just read
17653 a certificate: this is the best we can do. Also modified the code
17654 in `apps/verify.c` to take notice of return codes: it was previously
17655 attempting to read in certificates from NULL pointers and ignoring
17656 any errors: this is one reason why the cert and CRL reader seemed
17657 to work. It doesn't check return codes from the default certificate
17658 routines: these may well fail if the certificates aren't installed.
17662 * Code to support otherName option in GeneralName.
17666 * First update to verify code. Change the verify utility
17667 so it warns if it is passed a self signed certificate:
17668 for consistency with the normal behaviour. X509_verify
17669 has been modified to it will now verify a self signed
17670 certificate if *exactly* the same certificate appears
17671 in the store: it was previously impossible to trust a
17672 single self signed certificate. This means that:
17673 openssl verify ss.pem
17674 now gives a warning about a self signed certificate but
17675 openssl verify -CAfile ss.pem ss.pem
17680 * For servers, store verify_result in SSL_SESSION data structure
17681 (and add it to external session representation).
17682 This is needed when client certificate verifications fails,
17683 but an application-provided verification callback (set by
17684 SSL_CTX_set_cert_verify_callback) allows accepting the session
17685 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17686 but returns 1): When the session is reused, we have to set
17687 ssl->verify_result to the appropriate error code to avoid
17690 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17692 * Fix a bug in the new PKCS#7 code: it didn't consider the
17693 case in PKCS7_dataInit() where the signed PKCS7 structure
17694 didn't contain any existing data because it was being created.
17696 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17698 * Add a salt to the key derivation routines in enc.c. This
17699 forms the first 8 bytes of the encrypted file. Also add a
17700 -S option to allow a salt to be input on the command line.
17704 * New function X509_cmp(). Oddly enough there wasn't a function
17705 to compare two certificates. We do this by working out the SHA1
17706 hash and comparing that. X509_cmp() will be needed by the trust
17711 * SSL_get1_session() is like SSL_get_session(), but increments
17712 the reference count in the SSL_SESSION returned.
17714 *Geoff Thorpe <geoff@eu.c2.net>*
17716 * Fix for 'req': it was adding a null to request attributes.
17717 Also change the X509_LOOKUP and X509_INFO code to handle
17718 certificate auxiliary information.
17722 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17727 * Add the possibility to add extra information to the memory leak
17728 detecting output, to form tracebacks, showing from where each
17729 allocation was originated: CRYPTO_push_info("constant string") adds
17730 the string plus current file name and line number to a per-thread
17731 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17732 is like calling CYRPTO_pop_info() until the stack is empty.
17733 Also updated memory leak detection code to be multi-thread-safe.
17737 * Add options -text and -noout to pkcs7 utility and delete the
17738 encryption options which never did anything. Update docs.
17742 * Add options to some of the utilities to allow the pass phrase
17743 to be included on either the command line (not recommended on
17744 OSes like Unix) or read from the environment. Update the
17745 manpages and fix a few bugs.
17749 * Add a few manpages for some of the openssl commands.
17753 * Fix the -revoke option in ca. It was freeing up memory twice,
17754 leaking and not finding already revoked certificates.
17758 * Extensive changes to support certificate auxiliary information.
17759 This involves the use of X509_CERT_AUX structure and X509_AUX
17760 functions. An X509_AUX function such as PEM_read_X509_AUX()
17761 can still read in a certificate file in the usual way but it
17762 will also read in any additional "auxiliary information". By
17763 doing things this way a fair degree of compatibility can be
17764 retained: existing certificates can have this information added
17765 using the new 'x509' options.
17767 Current auxiliary information includes an "alias" and some trust
17768 settings. The trust settings will ultimately be used in enhanced
17769 certificate chain verification routines: currently a certificate
17770 can only be trusted if it is self signed and then it is trusted
17775 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17776 The problem was that one of the replacement routines had not been working
17777 since SSLeay releases. For now the offending routine has been replaced
17778 with non-optimised assembler. Even so, this now gives around 95%
17779 performance improvement for 1024 bit RSA signs.
17783 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17784 handling. Most clients have the effective key size in bits equal to
17785 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17786 A few however don't do this and instead use the size of the decrypted key
17787 to determine the RC2 key length and the AlgorithmIdentifier to determine
17788 the effective key length. In this case the effective key length can still
17789 be 40 bits but the key length can be 168 bits for example. This is fixed
17790 by manually forcing an RC2 key into the EVP_PKEY structure because the
17791 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17792 the key length and effective key length are equal.
17796 * Add a bunch of functions that should simplify the creation of
17797 X509_NAME structures. Now you should be able to do:
17798 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17799 and have it automatically work out the correct field type and fill in
17800 the structures. The more adventurous can try:
17801 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17802 and it will (hopefully) work out the correct multibyte encoding.
17806 * Change the 'req' utility to use the new field handling and multibyte
17807 copy routines. Before the DN field creation was handled in an ad hoc
17808 way in req, ca, and x509 which was rather broken and didn't support
17809 BMPStrings or UTF8Strings. Since some software doesn't implement
17810 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17811 using the dirstring_type option. See the new comment in the default
17812 openssl.cnf for more info.
17816 * Make crypto/rand/md_rand.c more robust:
17817 - Assure unique random numbers after fork().
17818 - Make sure that concurrent threads access the global counter and
17819 md serializably so that we never lose entropy in them
17820 or use exactly the same state in multiple threads.
17821 Access to the large state is not always serializable because
17822 the additional locking could be a performance killer, and
17823 md should be large enough anyway.
17827 * New file `apps/app_rand.c` with commonly needed functionality
17828 for handling the random seed file.
17830 Use the random seed file in some applications that previously did not:
17832 dsaparam -genkey (which also ignored its '-rand' option),
17835 x509 (when signing).
17836 Except on systems with /dev/urandom, it is crucial to have a random
17837 seed file at least for key creation, DSA signing, and for DH exchanges;
17838 for RSA signatures we could do without one.
17840 gendh and gendsa (unlike genrsa) used to read only the first byte
17841 of each file listed in the '-rand' option. The function as previously
17842 found in genrsa is now in app_rand.c and is used by all programs
17843 that support '-rand'.
17847 * In RAND_write_file, use mode 0600 for creating files;
17848 don't just chmod when it may be too late.
17852 * Report an error from X509_STORE_load_locations
17853 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17857 * New function ASN1_mbstring_copy() this copies a string in either
17858 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17859 into an ASN1_STRING type. A mask of permissible types is passed
17860 and it chooses the "minimal" type to use or an error if not type
17865 * Add function equivalents to the various macros in asn1.h. The old
17866 macros are retained with an `M_` prefix. Code inside the library can
17867 use the `M_` macros. External code (including the openssl utility)
17868 should *NOT* in order to be "shared library friendly".
17872 * Add various functions that can check a certificate's extensions
17873 to see if it usable for various purposes such as SSL client,
17874 server or S/MIME and CAs of these types. This is currently
17875 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17876 verification. Also added a -purpose flag to x509 utility to
17877 print out all the purposes.
17881 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17886 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17887 for, obtain and decode and extension and obtain its critical flag.
17888 This allows all the necessary extension code to be handled in a
17889 single function call.
17893 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17894 platforms. See crypto/rc4/rc4_enc.c for further details.
17898 * New -noout option to asn1parse. This causes no output to be produced
17899 its main use is when combined with -strparse and -out to extract data
17900 from a file (which may not be in ASN.1 format).
17904 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17905 when producing the local key id.
17907 *Richard Levitte <levitte@stacken.kth.se>*
17909 * New option -dhparam in s_server. This allows a DH parameter file to be
17910 stated explicitly. If it is not stated then it tries the first server
17911 certificate file. The previous behaviour hard coded the filename
17916 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17917 a public key to be input or output. For example:
17918 openssl rsa -in key.pem -pubout -out pubkey.pem
17919 Also added necessary DSA public key functions to handle this.
17923 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17924 in the message. This was handled by allowing
17925 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17927 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17929 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17930 to the end of the strings whereas this didn't. This would cause problems
17931 if strings read with d2i_ASN1_bytes() were later modified.
17933 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17935 * Fix for base64 decode bug. When a base64 bio reads only one line of
17936 data and it contains EOF it will end up returning an error. This is
17937 caused by input 46 bytes long. The cause is due to the way base64
17938 BIOs find the start of base64 encoded data. They do this by trying a
17939 trial decode on each line until they find one that works. When they
17940 do a flag is set and it starts again knowing it can pass all the
17941 data directly through the decoder. Unfortunately it doesn't reset
17942 the context it uses. This means that if EOF is reached an attempt
17943 is made to pass two EOFs through the context and this causes the
17944 resulting error. This can also cause other problems as well. As is
17945 usual with these problems it takes *ages* to find and the fix is
17946 trivial: move one line.
17948 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17950 * Ugly workaround to get s_client and s_server working under Windows. The
17951 old code wouldn't work because it needed to select() on sockets and the
17952 tty (for keypresses and to see if data could be written). Win32 only
17953 supports select() on sockets so we select() with a 1s timeout on the
17954 sockets and then see if any characters are waiting to be read, if none
17955 are present then we retry, we also assume we can always write data to
17956 the tty. This isn't nice because the code then blocks until we've
17957 received a complete line of data and it is effectively polling the
17958 keyboard at 1s intervals: however it's quite a bit better than not
17959 working at all :-) A dedicated Windows application might handle this
17960 with an event loop for example.
17964 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17965 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17966 will be called when RSA_sign() and RSA_verify() are used. This is useful
17967 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17968 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17969 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17970 This necessitated the support of an extra signature type NID_md5_sha1
17971 for SSL signatures and modifications to the SSL library to use it instead
17972 of calling RSA_public_decrypt() and RSA_private_encrypt().
17976 * Add new -verify -CAfile and -CApath options to the crl program, these
17977 will lookup a CRL issuers certificate and verify the signature in a
17978 similar way to the verify program. Tidy up the crl program so it
17979 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17980 less strict. It will now permit CRL extensions even if it is not
17981 a V2 CRL: this will allow it to tolerate some broken CRLs.
17985 * Initialize all non-automatic variables each time one of the openssl
17986 sub-programs is started (this is necessary as they may be started
17987 multiple times from the "OpenSSL>" prompt).
17989 *Lennart Bang, Bodo Moeller*
17991 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17992 removing all other RSA functionality (this is what NO_RSA does). This
17993 is so (for example) those in the US can disable those operations covered
17994 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17999 * Non-copying interface to BIO pairs.
18000 (still largely untested)
18004 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
18005 ASCII string. This was handled independently in various places before.
18009 * New functions UTF8_getc() and UTF8_putc() that parse and generate
18010 UTF8 strings a character at a time.
18014 * Use client_version from client hello to select the protocol
18015 (s23_srvr.c) and for RSA client key exchange verification
18016 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
18020 * Add various utility functions to handle SPKACs, these were previously
18021 handled by poking round in the structure internals. Added new function
18022 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
18023 print, verify and generate SPKACs. Based on an original idea from
18024 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
18028 * RIPEMD160 is operational on all platforms and is back in 'make test'.
18032 * Allow the config file extension section to be overwritten on the
18033 command line. Based on an original idea from Massimiliano Pala
18034 <madwolf@comune.modena.it>. The new option is called -extensions
18035 and can be applied to ca, req and x509. Also -reqexts to override
18036 the request extensions in req and -crlexts to override the crl extensions
18041 * Add new feature to the SPKAC handling in ca. Now you can include
18042 the same field multiple times by preceding it by "XXXX." for example:
18045 this is the same syntax as used in the req config file.
18049 * Allow certificate extensions to be added to certificate requests. These
18050 are specified in a 'req_extensions' option of the req section of the
18051 config file. They can be printed out with the -text option to req but
18052 are otherwise ignored at present.
18056 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
18057 data read consists of only the final block it would not decrypted because
18058 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
18059 A misplaced 'break' also meant the decrypted final block might not be
18060 copied until the next read.
18064 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
18065 a few extra parameters to the DH structure: these will be useful if
18066 for example we want the value of 'q' or implement X9.42 DH.
18070 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
18071 provides hooks that allow the default DSA functions or functions on a
18072 "per key" basis to be replaced. This allows hardware acceleration and
18073 hardware key storage to be handled without major modification to the
18074 library. Also added low-level modexp hooks and CRYPTO_EX structure and
18075 associated functions.
18079 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
18080 as "read only": it can't be written to and the buffer it points to will
18081 not be freed. Reading from a read only BIO is much more efficient than
18082 a normal memory BIO. This was added because there are several times when
18083 an area of memory needs to be read from a BIO. The previous method was
18084 to create a memory BIO and write the data to it, this results in two
18085 copies of the data and an O(n^2) reading algorithm. There is a new
18086 function BIO_new_mem_buf() which creates a read only memory BIO from
18087 an area of memory. Also modified the PKCS#7 routines to use read only
18092 * Bugfix: ssl23_get_client_hello did not work properly when called in
18093 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
18094 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
18095 but a retry condition occurred while trying to read the rest.
18099 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
18100 NID_pkcs7_encrypted by default: this was wrong since this should almost
18101 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
18102 the encrypted data type: this is a more sensible place to put it and it
18103 allows the PKCS#12 code to be tidied up that duplicated this
18108 * Changed obj_dat.pl script so it takes its input and output files on
18109 the command line. This should avoid shell escape redirection problems
18114 * Initial support for certificate extension requests, these are included
18115 in things like Xenroll certificate requests. Included functions to allow
18116 extensions to be obtained and added.
18120 * -crlf option to s_client and s_server for sending newlines as
18121 CRLF (as required by many protocols).
18125 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
18127 * Install libRSAglue.a when OpenSSL is built with RSAref.
18129 *Ralf S. Engelschall*
18131 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
18133 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
18135 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
18140 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
18141 DH parameters/keys (q is lost during that conversion, but the resulting
18142 DH parameters contain its length).
18144 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
18145 much faster than DH_generate_parameters (which creates parameters
18146 where `p = 2*q + 1`), and also the smaller q makes DH computations
18147 much more efficient (160-bit exponentiation instead of 1024-bit
18148 exponentiation); so this provides a convenient way to support DHE
18149 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
18150 utter importance to use
18151 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18153 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18154 when such DH parameters are used, because otherwise small subgroup
18155 attacks may become possible!
18159 * Avoid memory leak in i2d_DHparams.
18163 * Allow the -k option to be used more than once in the enc program:
18164 this allows the same encrypted message to be read by multiple recipients.
18168 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
18169 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
18170 it will always use the numerical form of the OID, even if it has a short
18175 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
18176 method only got called if p,q,dmp1,dmq1,iqmp components were present,
18177 otherwise bn_mod_exp was called. In the case of hardware keys for example
18178 no private key components need be present and it might store extra data
18179 in the RSA structure, which cannot be accessed from bn_mod_exp.
18180 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
18181 private key operations.
18185 * Added support for SPARC Linux.
18189 * pem_password_cb function type incompatibly changed from
18190 typedef int pem_password_cb(char *buf, int size, int rwflag);
18192 ....(char *buf, int size, int rwflag, void *userdata);
18193 so that applications can pass data to their callbacks:
18194 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18195 additional void * argument, which is just handed through whenever
18196 the password callback is called.
18198 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18200 New function SSL_CTX_set_default_passwd_cb_userdata.
18202 Compatibility note: As many C implementations push function arguments
18203 onto the stack in reverse order, the new library version is likely to
18204 interoperate with programs that have been compiled with the old
18205 pem_password_cb definition (PEM_whatever takes some data that
18206 happens to be on the stack as its last argument, and the callback
18207 just ignores this garbage); but there is no guarantee whatsoever that
18210 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18211 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18212 problems not only on Windows, but also on some Unix platforms.
18213 To avoid problematic command lines, these definitions are now in an
18214 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18215 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18219 * MIPS III/IV assembler module is reimplemented.
18223 * More DES library cleanups: remove references to srand/rand and
18224 delete an unused file.
18228 * Add support for the free Netwide assembler (NASM) under Win32,
18229 since not many people have MASM (ml) and it can be hard to obtain.
18230 This is currently experimental but it seems to work OK and pass all
18231 the tests. Check out INSTALL.W32 for info.
18235 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18236 without temporary keys kept an extra copy of the server key,
18237 and connections with temporary keys did not free everything in case
18242 * New function RSA_check_key and new openssl rsa option -check
18243 for verifying the consistency of RSA keys.
18245 *Ulf Moeller, Bodo Moeller*
18247 * Various changes to make Win32 compile work:
18248 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18249 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18250 comparison" warnings.
18251 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18255 * Add a debugging option to PKCS#5 v2 key generation function: when
18256 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18257 derived keys are printed to stderr.
18261 * Copy the flags in ASN1_STRING_dup().
18263 *Roman E. Pavlov <pre@mo.msk.ru>*
18265 * The x509 application mishandled signing requests containing DSA
18266 keys when the signing key was also DSA and the parameters didn't match.
18268 It was supposed to omit the parameters when they matched the signing key:
18269 the verifying software was then supposed to automatically use the CA's
18270 parameters if they were absent from the end user certificate.
18272 Omitting parameters is no longer recommended. The test was also
18273 the wrong way round! This was probably due to unusual behaviour in
18274 EVP_cmp_parameters() which returns 1 if the parameters match.
18275 This meant that parameters were omitted when they *didn't* match and
18276 the certificate was useless. Certificates signed with 'ca' didn't have
18279 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18281 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18282 The interface is as follows:
18283 Applications can use
18284 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18285 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18286 "off" is now the default.
18287 The library internally uses
18288 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18289 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18290 to disable memory-checking temporarily.
18292 Some inconsistent states that previously were possible (and were
18293 even the default) are now avoided.
18295 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18296 with each memory chunk allocated; this is occasionally more helpful
18297 than just having a counter.
18299 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18301 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18306 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18307 which largely parallels "options", but is for changing API behaviour,
18308 whereas "options" are about protocol behaviour.
18309 Initial "mode" flags are:
18311 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18312 a single record has been written.
18313 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18314 retries use the same buffer location.
18315 (But all of the contents must be
18320 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18323 * Fix problems with no-hmac etc.
18325 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18327 * New functions RSA_get_default_method(), RSA_set_method() and
18328 RSA_get_method(). These allows replacement of RSA_METHODs without having
18329 to mess around with the internals of an RSA structure.
18333 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18334 Also really enable memory leak checks in openssl.c and in some
18337 *Chad C. Mulligan, Bodo Moeller*
18339 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18340 up the length of negative integers. This has now been simplified to just
18341 store the length when it is first determined and use it later, rather
18342 than trying to keep track of where data is copied and updating it to
18344 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18346 * Add a new function PKCS7_signatureVerify. This allows the verification
18347 of a PKCS#7 signature but with the signing certificate passed to the
18348 function itself. This contrasts with PKCS7_dataVerify which assumes the
18349 certificate is present in the PKCS#7 structure. This isn't always the
18350 case: certificates can be omitted from a PKCS#7 structure and be
18351 distributed by "out of band" means (such as a certificate database).
18355 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18356 function prototypes in pem.h, also change util/mkdef.pl to add the
18357 necessary function names.
18361 * mk1mf.pl (used by Windows builds) did not properly read the
18362 options set by Configure in the top level Makefile, and Configure
18363 was not even able to write more than one option correctly.
18364 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18368 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18369 file to be loaded from a BIO or FILE pointer. The BIO version will
18370 for example allow memory BIOs to contain config info.
18374 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18375 Whoever hopes to achieve shared-library compatibility across versions
18376 must use this, not the compile-time macro.
18377 (Exercise 0.9.4: Which is the minimum library version required by
18379 Note: All this applies only to multi-threaded programs, others don't
18384 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18385 through a BIO pair triggered the default case, i.e.
18386 SSLerr(...,SSL_R_UNKNOWN_STATE).
18390 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18391 can use the SSL library even if none of the specific BIOs is
18396 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18397 for the encoded length.
18399 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18401 * Add initial documentation of the X509V3 functions.
18405 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18406 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18407 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18408 secure PKCS#8 private key format with a high iteration count.
18412 * Fix determination of Perl interpreter: A perl or perl5
18413 *directory* in $PATH was also accepted as the interpreter.
18415 *Ralf S. Engelschall*
18417 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18418 wrong with it but it was very old and did things like calling
18419 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18420 unusual formatting.
18424 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18425 to use the new extension code.
18429 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18430 with macros. This should make it easier to change their form, add extra
18431 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18436 * Add to configuration table a new entry that can specify an alternative
18437 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18438 according to Mark Crispin <MRC@Panda.COM>.
18442 * DES CBC did not update the IV. Weird.
18446 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18447 Changing the behaviour of the former might break existing programs --
18448 where IV updating is needed, des_ncbc_encrypt can be used.
18451 * When bntest is run from "make test" it drives bc to check its
18452 calculations, as well as internally checking them. If an internal check
18453 fails, it needs to cause bc to give a non-zero result or make test carries
18454 on without noticing the failure. Fixed.
18458 * DES library cleanups.
18462 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18463 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18464 ciphers. NOTE: although the key derivation function has been verified
18465 against some published test vectors it has not been extensively tested
18466 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18471 * Instead of "mkdir -p", which is not fully portable, use new
18472 Perl script "util/mkdir-p.pl".
18476 * Rewrite the way password based encryption (PBE) is handled. It used to
18477 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18478 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18479 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18480 the 'parameter' field of the AlgorithmIdentifier is passed to the
18481 underlying key generation function so it must do its own ASN1 parsing.
18482 This has also changed the EVP_PBE_CipherInit() function which now has a
18483 'parameter' argument instead of literal salt and iteration count values
18484 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18488 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18489 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18490 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18491 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18492 value was just used as a "magic string" and not used directly its
18493 value doesn't matter.
18497 * Introduce some semblance of const correctness to BN. Shame C doesn't
18502 * "linux-sparc64" configuration (ultrapenguin).
18504 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18505 "linux-sparc" configuration.
18507 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18509 * config now generates no-xxx options for missing ciphers.
18513 * Support the EBCDIC character set (work in progress).
18514 File ebcdic.c not yet included because it has a different license.
18516 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18518 * Support BS2000/OSD-POSIX.
18520 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18522 * Make callbacks for key generation use `void *` instead of `char *`.
18526 * Make S/MIME samples compile (not yet tested).
18530 * Additional typesafe stacks.
18534 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18538 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18540 * New configuration variant "sco5-gcc".
18542 * Updated some demos.
18544 *Sean O Riordain, Wade Scholine*
18546 * Add missing BIO_free at exit of pkcs12 application.
18550 * Fix memory leak in conf.c.
18554 * Updates for Win32 to assembler version of MD5.
18558 * Set #! path to perl in `apps/der_chop` to where we found it
18559 instead of using a fixed path.
18563 * SHA library changes for irix64-mips4-cc.
18567 * Improvements for VMS support.
18571 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18573 * Bignum library bug fix. IRIX 6 passes "make test" now!
18574 This also avoids the problems with SC4.2 and unpatched SC5.
18576 *Andy Polyakov <appro@fy.chalmers.se>*
18578 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18579 These are required because of the typesafe stack would otherwise break
18580 existing code. If old code used a structure member which used to be STACK
18581 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18582 sk_num or sk_value it would produce an error because the num, data members
18583 are not present in STACK_OF. Now it just produces a warning. sk_set
18584 replaces the old method of assigning a value to sk_value
18585 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18586 that does this will no longer work (and should use sk_set instead) but
18587 this could be regarded as a "questionable" behaviour anyway.
18591 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18592 correctly handle encrypted S/MIME data.
18596 * Change type of various DES function arguments from des_cblock
18597 (which means, in function argument declarations, pointer to char)
18598 to des_cblock * (meaning pointer to array with 8 char elements),
18599 which allows the compiler to do more typechecking; it was like
18600 that back in SSLeay, but with lots of ugly casts.
18602 Introduce new type const_des_cblock.
18606 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18607 problems: find RecipientInfo structure that matches recipient certificate
18608 and initialise the ASN1 structures properly based on passed cipher.
18612 * Belatedly make the BN tests actually check the results.
18616 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18617 to and from BNs: it was completely broken. New compilation option
18618 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18619 key elements as negative integers.
18623 * Reorganize and speed up MD5.
18625 *Andy Polyakov <appro@fy.chalmers.se>*
18629 *Richard Levitte <richard@levitte.org>*
18631 * New option -out to asn1parse to allow the parsed structure to be
18632 output to a file. This is most useful when combined with the -strparse
18633 option to examine the output of things like OCTET STRINGS.
18637 * Make SSL library a little more fool-proof by not requiring any longer
18638 that `SSL_set_{accept,connect}_state` be called before
18639 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18640 in many applications because usually everything *appeared* to work as
18641 intended anyway -- now it really works as intended).
18645 * Move openssl.cnf out of lib/.
18649 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18650 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18651 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18653 *Ralf S. Engelschall*
18655 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18656 handle PKCS#7 enveloped data properly.
18658 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18660 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18661 copying pointers. The cert_st handling is changed by this in
18662 various ways (and thus what used to be known as ctx->default_cert
18663 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18664 any longer when s->cert does not give us what we need).
18665 ssl_cert_instantiate becomes obsolete by this change.
18666 As soon as we've got the new code right (possibly it already is?),
18667 we have solved a couple of bugs of the earlier code where s->cert
18668 was used as if it could not have been shared with other SSL structures.
18670 Note that using the SSL API in certain dirty ways now will result
18671 in different behaviour than observed with earlier library versions:
18672 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18673 does not influence s as it used to.
18675 In order to clean up things more thoroughly, inside SSL_SESSION
18676 we don't use CERT any longer, but a new structure SESS_CERT
18677 that holds per-session data (if available); currently, this is
18678 the peer's certificate chain and, for clients, the server's certificate
18679 and temporary key. CERT holds only those values that can have
18680 meaningful defaults in an SSL_CTX.
18684 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18685 from the internal representation. Various PKCS#7 fixes: remove some
18686 evil casts and set the enc_dig_alg field properly based on the signing
18691 * Allow PKCS#12 password to be set from the command line or the
18692 environment. Let 'ca' get its config file name from the environment
18693 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18698 * Allow certificate policies extension to use an IA5STRING for the
18699 organization field. This is contrary to the PKIX definition but
18700 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18705 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18706 without disallowing inline assembler and the like for non-pedantic builds.
18710 * Support Borland C++ builder.
18712 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18718 * SHA-1 cleanups and performance enhancements.
18720 *Andy Polyakov <appro@fy.chalmers.se>*
18722 * Sparc v8plus assembler for the bignum library.
18724 *Andy Polyakov <appro@fy.chalmers.se>*
18726 * Accept any -xxx and +xxx compiler options in Configure.
18730 * Update HPUX configuration.
18734 * Add missing `sk_<type>_unshift()` function to safestack.h
18736 *Ralf S. Engelschall*
18738 * New function SSL_CTX_use_certificate_chain_file that sets the
18739 "extra_cert"s in addition to the certificate. (This makes sense
18740 only for "PEM" format files, as chains as a whole are not
18745 * Support verify_depth from the SSL API.
18746 x509_vfy.c had what can be considered an off-by-one-error:
18747 Its depth (which was not part of the external interface)
18748 was actually counting the number of certificates in a chain;
18749 now it really counts the depth.
18753 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18754 instead of X509err, which often resulted in confusing error
18755 messages since the error codes are not globally unique
18756 (e.g. an alleged error in ssl3_accept when a certificate
18757 didn't match the private key).
18759 * New function SSL_CTX_set_session_id_context that allows to set a default
18760 value (so that you don't need SSL_set_session_id_context for each
18761 connection using the SSL_CTX).
18765 * OAEP decoding bug fix.
18769 * Support INSTALL_PREFIX for package builders, as proposed by
18774 * New Configure options "threads" and "no-threads". For systems
18775 where the proper compiler options are known (currently Solaris
18776 and Linux), "threads" is the default.
18780 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18784 * Install various scripts to $(OPENSSLDIR)/misc, not to
18785 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18786 such as /usr/local/bin.
18790 * "make linux-shared" to build shared libraries.
18792 *Niels Poppe <niels@netbox.org>*
18794 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18798 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18799 extension adding in x509 utility.
18803 * Remove NOPROTO sections and error code comments.
18807 * Partial rewrite of the DEF file generator to now parse the ANSI
18812 * New Configure options --prefix=DIR and --openssldir=DIR.
18816 * Complete rewrite of the error code script(s). It is all now handled
18817 by one script at the top level which handles error code gathering,
18818 header rewriting and C source file generation. It should be much better
18819 than the old method: it now uses a modified version of Ulf's parser to
18820 read the ANSI prototypes in all header files (thus the old K&R definitions
18821 aren't needed for error creation any more) and do a better job of
18822 translating function codes into names. The old 'ASN1 error code embedded
18823 in a comment' is no longer necessary and it doesn't use .err files which
18824 have now been deleted. Also the error code call doesn't have to appear all
18825 on one line (which resulted in some large lines...).
18829 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18833 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18834 0 (which usually indicates a closed connection), but continue reading.
18838 * Fix some race conditions.
18842 * Add support for CRL distribution points extension. Add Certificate
18843 Policies and CRL distribution points documentation.
18847 * Move the autogenerated header file parts to crypto/opensslconf.h.
18851 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18852 8 of keying material. Merlin has also confirmed interop with this fix
18853 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18855 *Merlin Hughes <merlin@baltimore.ie>*
18857 * Fix lots of warnings.
18859 *Richard Levitte <levitte@stacken.kth.se>*
18861 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18862 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18864 *Richard Levitte <levitte@stacken.kth.se>*
18866 * Fix problems with sizeof(long) == 8.
18868 *Andy Polyakov <appro@fy.chalmers.se>*
18870 * Change functions to ANSI C.
18874 * Fix typos in error codes.
18876 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18878 * Remove defunct assembler files from Configure.
18882 * SPARC v8 assembler BIGNUM implementation.
18884 *Andy Polyakov <appro@fy.chalmers.se>*
18886 * Support for Certificate Policies extension: both print and set.
18887 Various additions to support the r2i method this uses.
18891 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18892 return a const string when you are expecting an allocated buffer.
18896 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18897 types DirectoryString and DisplayText.
18901 * Add code to allow r2i extensions to access the configuration database,
18902 add an LHASH database driver and add several ctx helper functions.
18906 * Fix an evil bug in bn_expand2() which caused various BN functions to
18907 fail when they extended the size of a BIGNUM.
18911 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18912 support typesafe stack.
18916 * Fix typo in SSL_[gs]et_options().
18918 *Nils Frostberg <nils@medcom.se>*
18920 * Delete various functions and files that belonged to the (now obsolete)
18921 old X509V3 handling code.
18925 * New Configure option "rsaref".
18929 * Don't auto-generate pem.h.
18933 * Introduce type-safe ASN.1 SETs.
18937 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18939 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18941 * Introduce type-safe STACKs. This will almost certainly break lots of code
18942 that links with OpenSSL (well at least cause lots of warnings), but fear
18943 not: the conversion is trivial, and it eliminates loads of evil casts. A
18944 few STACKed things have been converted already. Feel free to convert more.
18945 In the fullness of time, I'll do away with the STACK type altogether.
18949 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18950 specified in `<certfile>` by updating the entry in the index.txt file.
18951 This way one no longer has to edit the index.txt file manually for
18952 revoking a certificate. The -revoke option does the gory details now.
18954 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18956 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18957 `-text` option at all and this way the `-noout -text` combination was
18958 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18960 *Ralf S. Engelschall*
18962 * Make sure a corresponding plain text error message exists for the
18963 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18964 verify callback function determined that a certificate was revoked.
18966 *Ralf S. Engelschall*
18968 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18969 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18970 all available ciphers including rc5, which was forgotten until now.
18971 In order to let the testing shell script know which algorithms
18972 are available, a new (up to now undocumented) command
18973 `openssl list-cipher-commands` is used.
18977 * Bugfix: s_client occasionally would sleep in select() when
18978 it should have checked SSL_pending() first.
18982 * New functions DSA_do_sign and DSA_do_verify to provide access to
18983 the raw DSA values prior to ASN.1 encoding.
18987 * Tweaks to Configure
18989 *Niels Poppe <niels@netbox.org>*
18991 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18996 * New variables $(RANLIB) and $(PERL) in the Makefiles.
19000 * New config option to avoid instructions that are illegal on the 80386.
19001 The default code is faster, but requires at least a 486.
19005 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
19006 SSL2_SERVER_VERSION (not used at all) macros, which are now the
19007 same as SSL2_VERSION anyway.
19011 * New "-showcerts" option for s_client.
19015 * Still more PKCS#12 integration. Add pkcs12 application to openssl
19016 application. Various cleanups and fixes.
19020 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
19021 modify error routines to work internally. Add error codes and PBE init
19022 to library startup routines.
19026 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
19027 packing functions to asn1 and evp. Changed function names and error
19028 codes along the way.
19032 * PKCS12 integration: and so it begins... First of several patches to
19033 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
19034 objects to objects.h
19038 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
19039 and display support for Thawte strong extranet extension.
19043 * Add LinuxPPC support.
19045 *Jeff Dubrule <igor@pobox.org>*
19047 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
19048 bn_div_words in alpha.s.
19050 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
19052 * Make sure the RSA OAEP test is skipped under -DRSAref because
19053 OAEP isn't supported when OpenSSL is built with RSAref.
19055 *Ulf Moeller <ulf@fitug.de>*
19057 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
19058 so they no longer are missing under -DNOPROTO.
19060 *Soren S. Jorvang <soren@t.dk>*
19062 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
19064 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
19065 doesn't work when the session is reused. Coming soon!
19069 * Fix a security hole, that allows sessions to be reused in the wrong
19070 context thus bypassing client cert protection! All software that uses
19071 client certs and session caches in multiple contexts NEEDS PATCHING to
19072 allow session reuse! A fuller solution is in the works.
19074 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
19076 * Some more source tree cleanups (removed obsolete files
19077 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
19078 permission on "config" script to be executable) and a fix for the INSTALL
19081 *Ulf Moeller <ulf@fitug.de>*
19083 * Remove some legacy and erroneous uses of malloc, free instead of
19086 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
19088 * Make rsa_oaep_test return non-zero on error.
19090 *Ulf Moeller <ulf@fitug.de>*
19092 * Add support for native Solaris shared libraries. Configure
19093 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
19094 if someone would make that last step automatic.
19096 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
19098 * ctx_size was not built with the right compiler during "make links". Fixed.
19102 * Change the meaning of 'ALL' in the cipher list. It now means "everything
19103 except NULL ciphers". This means the default cipher list will no longer
19104 enable NULL ciphers. They need to be specifically enabled e.g. with
19105 the string "DEFAULT:eNULL".
19109 * Fix to RSA private encryption routines: if p < q then it would
19110 occasionally produce an invalid result. This will only happen with
19111 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
19115 * Be less restrictive and allow also `perl util/perlpath.pl
19116 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
19117 because this way one can also use an interpreter named `perl5` (which is
19118 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
19119 installed as `perl`).
19121 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19123 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
19125 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19127 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
19128 advapi32.lib to Win32 build and change the pem test comparison
19129 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
19130 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
19131 and crypto/des/ede_cbcm_enc.c.
19135 * DES quad checksum was broken on big-endian architectures. Fixed.
19139 * Comment out two functions in bio.h that aren't implemented. Fix up the
19140 Win32 test batch file so it (might) work again. The Win32 test batch file
19141 is horrible: I feel ill....
19145 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
19146 in e_os.h. Audit of header files to check ANSI and non ANSI
19147 sections: 10 functions were absent from non ANSI section and not exported
19148 from Windows DLLs. Fixed up libeay.num for new functions.
19152 * Make `openssl version` output lines consistent.
19154 *Ralf S. Engelschall*
19156 * Fix Win32 symbol export lists for BIO functions: Added
19157 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
19158 to ms/libeay{16,32}.def.
19160 *Ralf S. Engelschall*
19162 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
19163 fine under Unix and passes some trivial tests I've now added. But the
19164 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
19165 added to make sure no one expects that this stuff really works in the
19166 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
19167 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
19170 *Ralf S. Engelschall*
19172 * Fix the generation of two part addresses in perl.
19174 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
19176 * Add config entry for Linux on MIPS.
19178 *John Tobey <jtobey@channel1.com>*
19180 * Make links whenever Configure is run, unless we are on Windoze.
19184 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
19185 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19190 * Add a useful kludge to allow package maintainers to specify compiler and
19191 other platforms details on the command line without having to patch the
19192 Configure script every time: One now can use
19193 `perl Configure <id>:<details>`,
19194 i.e. platform ids are allowed to have details appended
19195 to them (separated by colons). This is treated as there would be a static
19196 pre-configured entry in Configure's %table under key `<id>` with value
19197 `<details>` and `perl Configure <id>` is called. So, when you want to
19198 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19199 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19200 now, which overrides the FreeBSD-elf entry on-the-fly.
19202 *Ralf S. Engelschall*
19204 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19208 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19209 on the `perl Configure ...` command line. This way one can compile
19210 OpenSSL libraries with Position Independent Code (PIC) which is needed
19211 for linking it into DSOs.
19213 *Ralf S. Engelschall*
19215 * Remarkably, export ciphers were totally broken and no-one had noticed!
19220 * Cleaned up the LICENSE document: The official contact for any license
19221 questions now is the OpenSSL core team under openssl-core@openssl.org.
19222 And add a paragraph about the dual-license situation to make sure people
19223 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19224 to the OpenSSL toolkit.
19226 *Ralf S. Engelschall*
19228 * General source tree makefile cleanups: Made `making xxx in yyy...`
19229 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19230 Additionally cleaned up the `make links` target: Remove unnecessary
19231 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19232 to speed processing and no longer clutter the display with confusing
19233 stuff. Instead only the actually done links are displayed.
19235 *Ralf S. Engelschall*
19237 * Permit null encryption ciphersuites, used for authentication only. It used
19238 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19239 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19244 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19245 signed attributes when verifying signatures (this would break them),
19246 the detached data encoding was wrong and public keys obtained using
19247 X509_get_pubkey() weren't freed.
19251 * Add text documentation for the BUFFER functions. Also added a work around
19252 to a Win95 console bug. This was triggered by the password read stuff: the
19253 last character typed gets carried over to the next fread(). If you were
19254 generating a new cert request using 'req' for example then the last
19255 character of the passphrase would be CR which would then enter the first
19260 * Added the new 'Includes OpenSSL Cryptography Software' button as
19261 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19262 button and can be used by applications based on OpenSSL to show the
19263 relationship to the OpenSSL project.
19265 *Ralf S. Engelschall*
19267 * Remove confusing variables in function signatures in files
19268 ssl/ssl_lib.c and ssl/ssl.h.
19270 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19272 * Don't install bss_file.c under PREFIX/include/
19274 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19276 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19277 functions that return function pointers and has support for NT specific
19278 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19279 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19280 unsigned to signed types: this was killing the Win32 compile.
19284 * Add new certificate file to stack functions,
19285 SSL_add_dir_cert_subjects_to_stack() and
19286 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19287 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19288 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19289 This means that Apache-SSL and similar packages don't have to mess around
19290 to add as many CAs as they want to the preferred list.
19294 * Experiment with doxygen documentation. Currently only partially applied to
19296 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19297 openssl.doxy as the configuration file.
19301 * Get rid of remaining C++-style comments which strict C compilers hate.
19303 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19305 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19306 compiled in by default: it has problems with large keys.
19310 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19311 DH private keys and/or callback functions which directly correspond to
19312 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19313 is needed for applications which have to configure certificates on a
19314 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19316 For the RSA certificate situation is makes no difference, but
19317 for the DSA certificate situation this fixes the "no shared cipher"
19318 problem where the OpenSSL cipher selection procedure failed because the
19319 temporary keys were not overtaken from the context and the API provided
19320 no way to reconfigure them.
19321 The new functions now let applications reconfigure the stuff and they
19322 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19323 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19324 non-public-API function ssl_cert_instantiate() is used as a helper
19325 function and also to reduce code redundancy inside ssl_rsa.c.
19327 *Ralf S. Engelschall*
19329 * Move s_server -dcert and -dkey options out of the undocumented feature
19330 area because they are useful for the DSA situation and should be
19331 recognized by the users.
19333 *Ralf S. Engelschall*
19335 * Fix the cipher decision scheme for export ciphers: the export bits are
19336 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19337 SSL_EXP_MASK. So, the original variable has to be used instead of the
19338 already masked variable.
19340 *Richard Levitte <levitte@stacken.kth.se>*
19342 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19344 *Richard Levitte <levitte@stacken.kth.se>*
19346 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19347 from `int` to `unsigned int` because it is a length and initialized by
19348 EVP_DigestFinal() which expects an `unsigned int *`.
19350 *Richard Levitte <levitte@stacken.kth.se>*
19352 * Don't hard-code path to Perl interpreter on shebang line of Configure
19353 script. Instead use the usual Shell->Perl transition trick.
19355 *Ralf S. Engelschall*
19357 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19358 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19359 -noout -modulus` as it's already the case for `openssl rsa -noout
19360 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19361 currently the public key is printed (a decision which was already done by
19362 `openssl dsa -modulus` in the past) which serves a similar purpose.
19363 Additionally the NO_RSA no longer completely removes the whole -modulus
19364 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19367 *Ralf S. Engelschall*
19369 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19370 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19372 *Arne Ansper <arne@ats.cyber.ee>*
19374 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19375 to be added. Now both 'req' and 'ca' can use new objects defined in the
19380 * Add cool BIO that does syslog (or event log on NT).
19382 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19384 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19385 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19386 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19387 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19391 * Add preliminary config info for new extension code.
19395 * Make RSA_NO_PADDING really use no padding.
19397 *Ulf Moeller <ulf@fitug.de>*
19399 * Generate errors when private/public key check is done.
19403 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19404 for some CRL extensions and new objects added.
19408 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19409 key usage extension and fuller support for authority key id.
19413 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19414 padding method for RSA, which is recommended for new applications in PKCS
19415 #1 v2.0 (RFC 2437, October 1998).
19416 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19417 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19418 against Bleichbacher's attack on RSA.
19419 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19422 * Updates to the new SSL compression code
19424 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19426 * Fix so that the version number in the master secret, when passed
19427 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19428 (because the server will not accept higher), that the version number
19429 is 0x03,0x01, not 0x03,0x00
19431 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19433 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19434 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19435 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19439 * Support for RAW extensions where an arbitrary extension can be
19440 created by including its DER encoding. See `apps/openssl.cnf` for
19445 * Make sure latest Perl versions don't interpret some generated C array
19446 code as Perl array code in the crypto/err/err_genc.pl script.
19448 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19450 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19451 not many people have the assembler. Various Win32 compilation fixes and
19452 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19453 build instructions.
19457 * Modify configure script 'Configure' to automatically create crypto/date.h
19458 file under Win32 and also build pem.h from pem.org. New script
19459 util/mkfiles.pl to create the MINFO file on environments that can't do a
19460 'make files': perl util/mkfiles.pl >MINFO should work.
19464 * Major rework of DES function declarations, in the pursuit of correctness
19465 and purity. As a result, many evil casts evaporated, and some weirdness,
19466 too. You may find this causes warnings in your code. Zapping your evil
19467 casts will probably fix them. Mostly.
19471 * Fix for a typo in asn1.h. Bug fix to object creation script
19472 obj_dat.pl. It considered a zero in an object definition to mean
19473 "end of object": none of the objects in objects.h have any zeros
19474 so it wasn't spotted.
19476 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19478 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19479 Masking (CBCM). In the absence of test vectors, the best I have been able
19480 to do is check that the decrypt undoes the encrypt, so far. Send me test
19481 vectors if you have them.
19485 * Correct calculation of key length for export ciphers (too much space was
19486 allocated for null ciphers). This has not been tested!
19490 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19491 message is now correct (it understands "crypto" and "ssl" on its
19492 command line). There is also now an "update" option. This will update
19493 the util/ssleay.num and util/libeay.num files with any new functions.
19495 perl util/mkdef.pl crypto ssl update
19496 it will update them.
19500 * Overhauled the Perl interface:
19501 - ported BN stuff to OpenSSL's different BN library
19502 - made the perl/ source tree CVS-aware
19503 - renamed the package from SSLeay to OpenSSL (the files still contain
19504 their history because I've copied them in the repository)
19505 - removed obsolete files (the test scripts will be replaced
19506 by better Test::Harness variants in the future)
19508 *Ralf S. Engelschall*
19510 * First cut for a very conservative source tree cleanup:
19511 1. merge various obsolete readme texts into doc/ssleay.txt
19512 where we collect the old documents and readme texts.
19513 2. remove the first part of files where I'm already sure that we no
19514 longer need them because of three reasons: either they are just temporary
19515 files which were left by Eric or they are preserved original files where
19516 I've verified that the diff is also available in the CVS via "cvs diff
19517 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19518 the crypto/md/ stuff).
19520 *Ralf S. Engelschall*
19522 * More extension code. Incomplete support for subject and issuer alt
19523 name, issuer and authority key id. Change the i2v function parameters
19524 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19525 what that's for :-) Fix to ASN1 macro which messed up
19526 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19530 * Preliminary support for ENUMERATED type. This is largely copied from the
19535 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19537 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19539 * Make sure `make rehash` target really finds the `openssl` program.
19541 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19543 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19544 like to hear about it if this slows down other processors.
19548 * Add CygWin32 platform information to Configure script.
19550 *Alan Batie <batie@aahz.jf.intel.com>*
19552 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19554 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19556 * New program nseq to manipulate netscape certificate sequences
19560 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19565 * Fixes to BN code. Previously the default was to define BN_RECURSION
19566 but the BN code had some problems that would cause failures when
19567 doing certificate verification and some other functions.
19569 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19571 * Add ASN1 and PEM code to support netscape certificate sequences.
19575 * Add ASN1 and PEM code to support netscape certificate sequences.
19579 * Add several PKIX and private extended key usage OIDs.
19583 * Modify the 'ca' program to handle the new extension code. Modify
19584 openssl.cnf for new extension format, add comments.
19588 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19589 and add a sample to openssl.cnf so req -x509 now adds appropriate
19594 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19595 error code, add initial support to X509_print() and x509 application.
19599 * Takes a deep breath and start adding X509 V3 extension support code. Add
19600 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19601 stuff is currently isolated and isn't even compiled yet.
19605 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19606 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19607 Removed the versions check from X509 routines when loading extensions:
19608 this allows certain broken certificates that don't set the version
19609 properly to be processed.
19613 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19614 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19615 can still be regenerated with "make depend".
19619 * Spelling mistake in C version of CAST-128.
19621 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19623 * Changes to the error generation code. The perl script err-code.pl
19624 now reads in the old error codes and retains the old numbers, only
19625 adding new ones if necessary. It also only changes the .err files if new
19626 codes are added. The makefiles have been modified to only insert errors
19627 when needed (to avoid needlessly modifying header files). This is done
19628 by only inserting errors if the .err file is newer than the auto generated
19629 C file. To rebuild all the error codes from scratch (the old behaviour)
19630 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19631 or delete all the .err files.
19635 * CAST-128 was incorrectly implemented for short keys. The C version has
19636 been fixed, but is untested. The assembler versions are also fixed, but
19637 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19638 to regenerate it if needed.
19639 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19640 Hagino <itojun@kame.net>*
19642 * File was opened incorrectly in randfile.c.
19644 *Ulf Möller <ulf@fitug.de>*
19646 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19647 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19648 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19649 al: it's just almost always a UTCTime. Note this patch adds new error
19650 codes so do a "make errors" if there are problems.
19654 * Correct Linux 1 recognition in config.
19656 *Ulf Möller <ulf@fitug.de>*
19658 * Remove pointless MD5 hash when using DSA keys in ca.
19660 *Anonymous <nobody@replay.com>*
19662 * Generate an error if given an empty string as a cert directory. Also
19663 generate an error if handed NULL (previously returned 0 to indicate an
19664 error, but didn't set one).
19666 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19668 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19672 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19673 parameters. This was causing a warning which killed off the Win32 compile.
19677 * Remove C++ style comments from crypto/bn/bn_local.h.
19679 *Neil Costigan <neil.costigan@celocom.com>*
19681 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19682 based on a text string, looking up short and long names and finally
19683 "dot" format. The "dot" format stuff didn't work. Added new function
19684 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19685 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19686 OID is not part of the table.
19690 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19691 X509_LOOKUP_by_alias().
19695 * Sort openssl functions by name.
19699 * Get the `gendsa` command working and add it to the `list` command. Remove
19700 encryption from sample DSA keys (in case anyone is interested the password
19705 * Make *all* `*_free` functions accept a NULL pointer.
19707 *Frans Heymans <fheymans@isaserver.be>*
19709 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19712 *Anonymous <nobody@replay.com>*
19714 * s_server should send the CAfile as acceptable CAs, not its own cert.
19716 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19718 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19720 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19722 * Temp key "for export" tests were wrong in s3_srvr.c.
19724 *Anonymous <nobody@replay.com>*
19726 * Add prototype for temp key callback functions
19727 SSL_CTX_set_tmp_{rsa,dh}_callback().
19731 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19732 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19736 * X509_name_add_entry() freed the wrong thing after an error.
19738 *Arne Ansper <arne@ats.cyber.ee>*
19740 * rsa_eay.c would attempt to free a NULL context.
19742 *Arne Ansper <arne@ats.cyber.ee>*
19744 * BIO_s_socket() had a broken should_retry() on Windoze.
19746 *Arne Ansper <arne@ats.cyber.ee>*
19748 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19750 *Arne Ansper <arne@ats.cyber.ee>*
19752 * Make sure the already existing X509_STORE->depth variable is initialized
19753 in X509_STORE_new(), but document the fact that this variable is still
19754 unused in the certificate verification process.
19756 *Ralf S. Engelschall*
19758 * Fix the various library and `apps/` files to free up pkeys obtained from
19759 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19763 * Fix reference counting in X509_PUBKEY_get(). This makes
19764 demos/maurice/example2.c work, amongst others, probably.
19766 *Steve Henson and Ben Laurie*
19768 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19769 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19770 are no longer created. This way we have a single and consistent command
19771 line interface `openssl <command>`, similar to `cvs <command>`.
19773 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19775 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19776 BIT STRING wrapper always have zero unused bits.
19780 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19784 * Make the top-level INSTALL documentation easier to understand.
19788 * Makefiles updated to exit if an error occurs in a sub-directory
19789 make (including if user presses ^C) [Paul Sutton]
19791 * Make Montgomery context stuff explicit in RSA data structure.
19795 * Fix build order of pem and err to allow for generated pem.h.
19799 * Fix renumbering bug in X509_NAME_delete_entry().
19803 * Enhanced the err-ins.pl script so it makes the error library number
19804 global and can add a library name. This is needed for external ASN1 and
19805 other error libraries.
19809 * Fixed sk_insert which never worked properly.
19813 * Fix ASN1 macros so they can handle indefinite length constructed
19814 EXPLICIT tags. Some non standard certificates use these: they can now
19819 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19820 into a single doc/ssleay.txt bundle. This way the information is still
19821 preserved but no longer messes up this directory. Now it's new room for
19822 the new set of documentation files.
19824 *Ralf S. Engelschall*
19826 * SETs were incorrectly DER encoded. This was a major pain, because they
19827 shared code with SEQUENCEs, which aren't coded the same. This means that
19828 almost everything to do with SETs or SEQUENCEs has either changed name or
19829 number of arguments.
19831 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19833 * Fix test data to work with the above.
19837 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19838 was already fixed by Eric for 0.9.1 it seems.
19840 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19842 * Autodetect FreeBSD3.
19846 * Fix various bugs in Configure. This affects the following platforms:
19850 unixware-2.0-pentium
19855 * Eliminate generated files from CVS. Reorder tests to regenerate files
19856 before they are needed.
19860 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19864 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19866 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19867 changed SSLeay to OpenSSL in version strings.
19869 *Ralf S. Engelschall*
19871 * Some fixups to the top-level documents.
19875 * Fixed the nasty bug where rsaref.h was not found under compile-time
19876 because the symlink to include/ was missing.
19878 *Ralf S. Engelschall*
19880 * Incorporated the popular no-RSA/DSA-only patches
19881 which allow to compile an RSA-free SSLeay.
19883 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19885 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19886 when "ssleay" is still not found.
19888 *Ralf S. Engelschall*
19890 * Added more platforms to Configure: Cray T3E, HPUX 11,
19892 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19894 * Updated the README file.
19896 *Ralf S. Engelschall*
19898 * Added various .cvsignore files in the CVS repository subdirs
19899 to make a "cvs update" really silent.
19901 *Ralf S. Engelschall*
19903 * Recompiled the error-definition header files and added
19904 missing symbols to the Win32 linker tables.
19906 *Ralf S. Engelschall*
19908 * Cleaned up the top-level documents;
19909 o new files: CHANGES and LICENSE
19910 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19911 o merged COPYRIGHT into LICENSE
19912 o removed obsolete TODO file
19913 o renamed MICROSOFT to INSTALL.W32
19915 *Ralf S. Engelschall*
19917 * Removed dummy files from the 0.9.1b source tree:
19918 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19919 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19920 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19921 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19922 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19924 *Ralf S. Engelschall*
19926 * Added various platform portability fixes.
19930 * The Genesis of the OpenSSL rpject:
19931 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19932 Young and Tim J. Hudson created while they were working for C2Net until
19935 *The OpenSSL Project*
19937 ### Changes between 0.9.0b and 0.9.1b [not released]
19939 * Updated a few CA certificates under certs/
19943 * Changed some BIGNUM api stuff.
19947 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19948 DGUX x86, Linux Alpha, etc.
19952 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19953 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19958 * Add -strparse option to asn1pars program which parses nested
19961 *Dr Stephen Henson <shenson@bigfoot.com>*
19963 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19967 * DSA fix for "ca" program.
19971 * Added "-genkey" option to "dsaparam" program.
19975 * Added RIPE MD160 (rmd160) message digest.
19979 * Added -a (all) option to "ssleay version" command.
19983 * Added PLATFORM define which is the id given to Configure.
19987 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19991 * Extended the ASN.1 parser routines.
19995 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19999 * Added a BN_CTX to the BN library.
20003 * Fixed the weak key values in DES library
20007 * Changed API in EVP library for cipher aliases.
20011 * Added support for RC2/64bit cipher.
20015 * Converted the lhash library to the crypto/mem.c functions.
20019 * Added more recognized ASN.1 object ids.
20023 * Added more RSA padding checks for SSL/TLS.
20027 * Added BIO proxy/filter functionality.
20031 * Added extra_certs to SSL_CTX which can be used
20032 send extra CA certificates to the client in the CA cert chain sending
20033 process. It can be configured with SSL_CTX_add_extra_chain_cert().
20037 * Now Fortezza is denied in the authentication phase because
20038 this is key exchange mechanism is not supported by SSLeay at all.
20042 * Additional PKCS1 checks.
20046 * Support the string "TLSv1" for all TLS v1 ciphers.
20050 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
20051 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
20055 * Fixed a few memory leaks.
20059 * Fixed various code and comment typos.
20063 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
20064 bytes sent in the client random.
20066 *Edward Bishop <ebishop@spyglass.com>*
20070 [CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817
20071 [CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446
20072 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
20073 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
20074 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
20075 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
20076 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
20077 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
20078 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
20079 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
20080 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
20081 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
20082 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
20083 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
20084 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
20085 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
20086 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
20087 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
20088 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
20089 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
20090 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
20091 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
20092 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
20093 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
20094 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
20095 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
20096 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
20097 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
20098 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
20099 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
20100 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
20101 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
20102 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
20103 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
20104 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
20105 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
20106 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
20107 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
20108 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
20109 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
20110 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
20111 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
20112 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
20113 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
20114 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
20115 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
20116 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
20117 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
20118 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
20119 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
20120 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
20121 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
20122 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
20123 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
20124 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
20125 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
20126 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
20127 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
20128 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
20129 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
20130 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
20131 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
20132 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
20133 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
20134 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
20135 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
20136 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
20137 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
20138 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
20139 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
20140 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
20141 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
20142 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
20143 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
20144 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
20145 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
20146 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
20147 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
20148 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
20149 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
20150 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
20151 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
20152 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
20153 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
20154 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
20155 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
20156 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
20157 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
20158 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
20159 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
20160 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
20161 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
20162 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
20163 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
20164 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
20165 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
20166 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
20167 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
20168 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
20169 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
20170 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
20171 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
20172 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
20173 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
20174 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
20175 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
20176 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
20177 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
20178 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
20179 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
20180 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
20181 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
20182 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
20183 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20184 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20185 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20186 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20187 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20188 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20189 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20190 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20191 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20192 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20193 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20194 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20195 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20196 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20197 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20198 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20199 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20200 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20201 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20202 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20203 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20204 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20205 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20206 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20207 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20208 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20209 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20210 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20211 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20212 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20213 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20214 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20215 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20216 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20217 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20218 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20219 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20220 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20221 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20222 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20223 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20224 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20225 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20226 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20227 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20228 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20229 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20230 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20231 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20232 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20233 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20234 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20235 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20236 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20237 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20238 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20239 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20240 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20241 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20242 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20243 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20244 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20245 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20246 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20247 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20248 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20249 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20250 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20251 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655