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 * Changed the default salt length used by PBES2 KDF's (PBKDF2 and scrypt)
29 from 8 bytes to 16 bytes.
30 The PKCS5 (RFC 8018) standard uses a 64 bit salt length for PBE, and
31 recommends a minimum of 64 bits for PBES2. For FIPS compliance PBKDF2
32 requires a salt length of 128 bits. This affects OpenSSL command line
33 applications such as "genrsa" and "pkcs8" and API's such as
34 PEM_write_bio_PrivateKey() that are reliant on the default value.
35 The additional commandline option 'saltlen' has been added to the
36 OpenSSL command line applications for "pkcs8" and "enc" to allow the
37 salt length to be set to a non default value.
41 * Changed the default value of the `ess_cert_id_alg` configuration
42 option which is used to calculate the TSA's public key certificate
43 identifier. The default algorithm is updated to be sha256 instead
48 * Added optimization for SM2 algorithm on aarch64. It uses a huge precomputed
49 table for point multiplication of the base point, which increases the size of
50 libcrypto from 4.4 MB to 4.9 MB. A new configure option `no-sm2-precomp` has
51 been added to disable the precomputed table.
55 * Added client side support for QUIC
57 *Hugo Landau, Matt Caswell, Paul Dale, Tomáš Mráz, Richard Levitte*
59 * Added multiple tutorials on the OpenSSL library and in particular
60 on writing various clients (using TLS and QUIC protocols) with libssl.
64 * Added secp384r1 implementation using Solinas' reduction to improve
65 speed of the NIST P-384 elliptic curve. To enable the implementation
66 the build option `enable-ec_nistp_64_gcc_128` must be used.
70 * Improved RFC7468 compliance of the asn1parse command.
74 * Added SHA256/192 algorithm support.
78 * Added support for securely getting root CA certificate update in
83 * Improved contention on global write locks by using more read locks where
88 * Improved performance of OSSL_PARAM lookups in performance critical
93 * Added the SSL_get0_group_name() function to provide access to the
94 name of the group used for the TLS key exchange.
98 * Provide a new configure option `no-http` that can be used to disable the
99 HTTP support. Provide new configure options `no-apps` and `no-docs` to
100 disable building the openssl command line application and the documentation.
104 * Provide a new configure option `no-ecx` that can be used to disable the
105 X25519, X448, and EdDSA support.
109 * When multiple OSSL_KDF_PARAM_INFO parameters are passed to
110 the EVP_KDF_CTX_set_params() function they are now concatenated not just
111 for the HKDF algorithm but also for SSKDF and X9.63 KDF algorithms.
115 * Added OSSL_FUNC_keymgmt_im/export_types_ex() provider functions that get
116 the provider context as a parameter.
120 * TLS round-trip time calculation was added by a Brigham Young University
121 Capstone team partnering with Sandia National Laboratories. A new function
122 in ssl_lib titled SSL_get_handshake_rtt will calculate and retrieve this
127 * Added the "-quic" option to s_client to enable connectivity to QUIC servers.
128 QUIC requires the use of ALPN, so this must be specified via the "-alpn"
129 option. Use of the "advanced" s_client command command via the "-adv" option
134 * Added an "advanced" command mode to s_client. Use this with the "-adv"
135 option. The old "basic" command mode recognises certain letters that must
136 always appear at the start of a line and cannot be escaped. The advanced
137 command mode enables commands to be entered anywhere and there is an
138 escaping mechanism. After starting s_client with "-adv" type "{help}"
139 to show a list of available commands.
143 * Add Raw Public Key (RFC7250) support. Authentication is supported
144 by matching keys against either local policy (TLSA records synthesised
145 from the expected keys) or DANE (TLSA records obtained by the
146 application from DNS). TLSA records will also match the same key in
147 the server certificate, should RPK use not happen to be negotiated.
151 * Added support for modular exponentiation and CRT offloading for the
156 * Added further assembler code for the RISC-V architecture.
160 * Added EC_GROUP_to_params() which creates an OSSL_PARAM array
161 from a given EC_GROUP.
165 * Improved support for non-default library contexts and property queries
166 when parsing PKCS#12 files.
170 * Implemented support for all five instances of EdDSA from RFC8032:
171 Ed25519, Ed25519ctx, Ed25519ph, Ed448, and Ed448ph.
172 The streaming is not yet supported for the HashEdDSA variants
173 (Ed25519ph and Ed448ph).
177 * Added SM4 optimization for ARM processors using ASIMD and AES HW
182 * Implemented SM4-XTS support.
186 * Added platform-agnostic OSSL_sleep() function.
190 * Implemented deterministic ECDSA signatures (RFC6979) support.
194 * Implemented AES-GCM-SIV (RFC8452) support.
198 * Added support for pluggable (provider-based) TLS signature algorithms.
199 This enables TLS 1.3 authentication operations with algorithms embedded
200 in providers not included by default in OpenSSL. In combination with
201 the already available pluggable KEM and X.509 support, this enables
202 for example suitable providers to deliver post-quantum or quantum-safe
203 cryptography to OpenSSL users.
207 * Added support for pluggable (provider-based) CMS signature algorithms.
208 This enables CMS sign and verify operations with algorithms embedded
209 in providers not included by default in OpenSSL.
213 * Added support for Hybrid Public Key Encryption (HPKE) as defined
214 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
215 Message Layer Security (MLS) and other IETF specifications.
216 HPKE can also be used by other applications that require
217 encrypting "to" an ECDH public key. External APIs are defined in
218 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
222 * Implemented HPKE DHKEM support in providers used by HPKE (RFC9180)
227 * Add support for certificate compression (RFC8879), including
228 library support for Brotli and Zstandard compression.
232 * Add the ability to add custom attributes to PKCS12 files. Add a new API
233 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
234 for a user specified callback and optional argument.
235 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
236 added to the existing STACK_OF attrs.
240 * Major refactor of the libssl record layer.
244 * Add a mac salt length option for the pkcs12 command.
248 * Add more SRTP protection profiles from RFC8723 and RFC8269.
252 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
254 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
256 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
257 supported and enabled.
261 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
262 to the list of ciphersuites providing Perfect Forward Secrecy as
263 required by SECLEVEL >= 3.
265 *Dmitry Belyavskiy, Nicola Tuveri*
267 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
268 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
269 SSL_get0_iana_groups() function-like macro, retrieves the list of
270 supported groups sent by the peer.
271 The function SSL_client_hello_get_extension_order() populates
272 a caller-supplied array with the list of extension types present in the
273 ClientHello, in order of appearance.
277 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
278 to make it possible to use empty passphrase strings.
282 * The PKCS12_parse() function now supports MAC-less PKCS12 files.
286 * Added ASYNC_set_mem_functions() and ASYNC_get_mem_functions() calls to be able
287 to change functions used for allocating the memory of asynchronous call stack.
291 * Added support for signed BIGNUMs in the OSSL_PARAM APIs.
295 * A failure exit code is returned when using the openssl x509 command to check
296 certificate attributes and the checks fail.
300 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
301 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
302 of 160 bits and above and less than 224 bits were previously accepted by
303 default but are now no longer allowed. By default TLS compression was
304 already disabled in previous OpenSSL versions. At security level 2 it cannot
309 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
314 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
315 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
316 will need to load the legacy crypto provider.
320 * CCM8 cipher suites in TLS have been downgraded to security level zero
321 because they use a short authentication tag which lowers their strength.
325 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
330 * Add X.509 certificate codeSigning purpose and related checks on key usage and
331 extended key usage of the leaf certificate according to the CA/Browser Forum.
335 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
336 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
337 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
338 X.509 version 3 if the certificate information includes X.509 extensions.
342 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
343 such as adding a trace facility for debugging certificate chain building.
347 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
348 in particular supporting requests for central key generation, generalized
349 polling, and various types of genm/genp exchanges defined in CMP Updates.
353 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
354 like correcting the TLS and proxy support and adding tracing for debugging.
358 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
362 * `CMS_add0_cert()` and `CMS_add1_cert()` no longer throw an error if
363 a certificate to be added is already present. `CMS_sign_ex()` and
364 `CMS_sign()` now ignore any duplicate certificates in their `certs` argument
365 and no longer throw an error for them.
369 * Fixed and extended `util/check-format.pl` for checking adherence to the
370 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
371 The checks are meanwhile more complete and yield fewer false positives.
375 * Added BIO_s_dgram_pair() and BIO_s_dgram_mem() that provide memory-based
376 BIOs with datagram semantics and support for BIO_sendmmsg() and BIO_recvmmsg()
377 calls. They can be used as the transport BIOs for QUIC.
379 *Hugo Landau, Matt Caswell and Tomáš Mráz*
381 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
382 sending and receiving multiple messages in a single call. An implementation
383 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
387 * Support for loading root certificates from the Windows certificate store
388 has been added. The support is in the form of a store which recognises the
389 URI string of `org.openssl.winstore://`. This URI scheme currently takes no
390 arguments. This store is built by default and can be disabled using the new
391 compile-time option `no-winstore`. This store is not currently used by
392 default and must be loaded explicitly using the above store URI. It is
393 expected to be loaded by default in the future.
397 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
398 kernel versions that support KTLS have a known bug in CCM processing. That
399 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
400 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
405 * Added `-ktls` option to `s_server` and `s_client` commands to enable the
410 * Zerocopy KTLS sendfile() support on Linux.
414 * The OBJ_ calls are now thread safe using a global lock.
418 * New parameter `-digest` for openssl cms command allowing signing
419 pre-computed digests and new CMS API functions supporting that
424 * OPENSSL_malloc() and other allocation functions now raise errors on
425 allocation failures. The callers do not need to explicitly raise errors
426 unless they want to for tracing purposes.
430 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
431 decryption as a protection against Bleichenbacher-like attacks.
432 The RSA decryption API will now return a randomly generated deterministic
433 message instead of an error in case it detects an error when checking
434 padding during PKCS#1 v1.5 decryption. This is a general protection against
435 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
437 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
438 on the RSA decryption context.
442 * Added support for Brainpool curves in TLS-1.3.
444 *Bernd Edlinger and Matt Caswell*
446 * Added OpenBSD specific build targets.
450 * Support for Argon2d, Argon2i, Argon2id KDFs has been added along with
451 a basic thread pool implementation for select platforms.
458 ### Changes between 3.1.2 and 3.1.3 [xx XXX xxxx]
460 * Fix POLY1305 MAC implementation corrupting XMM registers on Windows.
462 The POLY1305 MAC (message authentication code) implementation in OpenSSL
463 does not save the contents of non-volatile XMM registers on Windows 64
464 platform when calculating the MAC of data larger than 64 bytes. Before
465 returning to the caller all the XMM registers are set to zero rather than
466 restoring their previous content. The vulnerable code is used only on newer
467 x86_64 processors supporting the AVX512-IFMA instructions.
469 The consequences of this kind of internal application state corruption can
470 be various - from no consequences, if the calling application does not
471 depend on the contents of non-volatile XMM registers at all, to the worst
472 consequences, where the attacker could get complete control of the
473 application process. However given the contents of the registers are just
474 zeroized so the attacker cannot put arbitrary values inside, the most likely
475 consequence, if any, would be an incorrect result of some application
476 dependent calculations or a crash leading to a denial of service.
482 ### Changes between 3.1.1 and 3.1.2 [1 Aug 2023]
484 * Fix excessive time spent checking DH q parameter value.
486 The function DH_check() performs various checks on DH parameters. After
487 fixing CVE-2023-3446 it was discovered that a large q parameter value can
488 also trigger an overly long computation during some of these checks.
489 A correct q value, if present, cannot be larger than the modulus p
490 parameter, thus it is unnecessary to perform these checks if q is larger
493 If DH_check() is called with such q parameter value,
494 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally
495 intensive checks are skipped.
501 * Fix DH_check() excessive time with over sized modulus.
503 The function DH_check() performs various checks on DH parameters. One of
504 those checks confirms that the modulus ("p" parameter) is not too large.
505 Trying to use a very large modulus is slow and OpenSSL will not normally use
506 a modulus which is over 10,000 bits in length.
508 However the DH_check() function checks numerous aspects of the key or
509 parameters that have been supplied. Some of those checks use the supplied
510 modulus value even if it has already been found to be too large.
512 A new limit has been added to DH_check of 32,768 bits. Supplying a
513 key/parameters with a modulus over this size will simply cause DH_check() to
520 * Do not ignore empty associated data entries with AES-SIV.
522 The AES-SIV algorithm allows for authentication of multiple associated
523 data entries along with the encryption. To authenticate empty data the
524 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
525 with NULL pointer as the output buffer and 0 as the input buffer length.
526 The AES-SIV implementation in OpenSSL just returns success for such call
527 instead of performing the associated data authentication operation.
528 The empty data thus will not be authenticated. ([CVE-2023-2975])
530 Thanks to Juerg Wullschleger (Google) for discovering the issue.
532 The fix changes the authentication tag value and the ciphertext for
533 applications that use empty associated data entries with AES-SIV.
534 To decrypt data encrypted with previous versions of OpenSSL the application
535 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
540 * When building with the `enable-fips` option and using the resulting
541 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
542 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
543 not operate with truncated digests (FIPS 140-3 IG G.R).
547 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
549 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
550 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
552 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
553 numeric text form. For gigantic sub-identifiers, this would take a very
554 long time, the time complexity being O(n^2) where n is the size of that
555 sub-identifier. ([CVE-2023-2650])
557 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
558 IDENTIFIER to canonical numeric text form if the size of that OBJECT
559 IDENTIFIER is 586 bytes or less, and fail otherwise.
561 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
562 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
563 most 128 sub-identifiers, and that the maximum value that each sub-
564 identifier may have is 2^32-1 (4294967295 decimal).
566 For each byte of every sub-identifier, only the 7 lower bits are part of
567 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
568 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
573 * Multiple algorithm implementation fixes for ARM BE platforms.
577 * Added a -pedantic option to fipsinstall that adjusts the various
578 settings to ensure strict FIPS compliance rather than backwards
583 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
584 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
585 trigger a crash of an application using AES-XTS decryption if the memory
586 just after the buffer being decrypted is not mapped.
587 Thanks to Anton Romanov (Amazon) for discovering the issue.
592 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
593 The previous fix for this timing side channel turned out to cause
594 a severe 2-3x performance regression in the typical use case
595 compared to 3.0.7. The new fix uses existing constant time
596 code paths, and restores the previous performance level while
597 fully eliminating all existing timing side channels.
598 The fix was developed by Bernd Edlinger with testing support
603 * Add FIPS provider configuration option to disallow the use of
604 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
605 The option '-no_drbg_truncated_digests' can optionally be
606 supplied to 'openssl fipsinstall'.
610 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
611 that it does not enable policy checking. Thanks to David Benjamin for
612 discovering this issue.
617 * Fixed an issue where invalid certificate policies in leaf certificates are
618 silently ignored by OpenSSL and other certificate policy checks are skipped
619 for that certificate. A malicious CA could use this to deliberately assert
620 invalid certificate policies in order to circumvent policy checking on the
621 certificate altogether.
626 * Limited the number of nodes created in a policy tree to mitigate
627 against CVE-2023-0464. The default limit is set to 1000 nodes, which
628 should be sufficient for most installations. If required, the limit
629 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
630 time define to a desired maximum number of nodes or zero to allow
636 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
638 * Add FIPS provider configuration option to enforce the
639 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
640 The option '-ems-check' can optionally be supplied to
641 'openssl fipsinstall'.
645 * The FIPS provider includes a few non-approved algorithms for
646 backward compatibility purposes and the "fips=yes" property query
647 must be used for all algorithm fetches to ensure FIPS compliance.
649 The algorithms that are included but not approved are Triple DES ECB,
650 Triple DES CBC and EdDSA.
654 * Added support for KMAC in KBKDF.
658 * RNDR and RNDRRS support in provider functions to provide
659 random number generation for Arm CPUs (aarch64).
663 * s_client and s_server apps now explicitly say when the TLS version
664 does not include the renegotiation mechanism. This avoids confusion
665 between that scenario versus when the TLS version includes secure
666 renegotiation but the peer lacks support for it.
670 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
672 *Tomasz Kantecki, Andrey Matyukov*
674 * The various OBJ_* functions have been made thread safe.
678 * Parallel dual-prime 1536/2048-bit modular exponentiation for
679 AVX512_IFMA capable processors.
681 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
683 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
684 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
685 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
686 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
687 `OPENSSL_NO_DEPRECATED_3_1`.
689 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
690 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
691 definitions for these functions regardless of whether
692 `OPENSSL_NO_DEPRECATED_3_1` is defined.
694 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
695 functions regardless of whether they are using them. It is recommended that
696 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
700 * When generating safe-prime DH parameters set the recommended private key
701 length equivalent to minimum key lengths as in RFC 7919.
705 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
706 maximum size that is smaller or equal to the digest length to comply with
707 FIPS 186-4 section 5. This is implemented by a new option
708 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
709 `rsa_pss_saltlen` parameter, which is now the default. Signature
710 verification is not affected by this change and continues to work as before.
717 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
718 listed here are only a brief description.
719 The migration guide contains more detailed information related to new features,
720 breaking changes, and mappings for the large list of deprecated functions.
722 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
724 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
726 * Fixed NULL dereference during PKCS7 data verification.
728 A NULL pointer can be dereferenced when signatures are being
729 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
730 algorithm used for the signature is known to the OpenSSL library but
731 the implementation of the hash algorithm is not available the digest
732 initialization will fail. There is a missing check for the return
733 value from the initialization function which later leads to invalid
734 usage of the digest API most likely leading to a crash.
737 PKCS7 data is processed by the SMIME library calls and also by the
738 time stamp (TS) library calls. The TLS implementation in OpenSSL does
739 not call these functions however third party applications would be
740 affected if they call these functions to verify signatures on untrusted
745 * Fixed X.400 address type confusion in X.509 GeneralName.
747 There is a type confusion vulnerability relating to X.400 address processing
748 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
749 but the public structure definition for GENERAL_NAME incorrectly specified
750 the type of the x400Address field as ASN1_TYPE. This field is subsequently
751 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
754 When CRL checking is enabled (i.e. the application sets the
755 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
756 pass arbitrary pointers to a memcmp call, enabling them to read memory
757 contents or enact a denial of service.
762 * Fixed NULL dereference validating DSA public key.
764 An invalid pointer dereference on read can be triggered when an
765 application tries to check a malformed DSA public key by the
766 EVP_PKEY_public_check() function. This will most likely lead
767 to an application crash. This function can be called on public
768 keys supplied from untrusted sources which could allow an attacker
769 to cause a denial of service attack.
771 The TLS implementation in OpenSSL does not call this function
772 but applications might call the function if there are additional
773 security requirements imposed by standards such as FIPS 140-3.
776 *Shane Lontis, Tomáš Mráz*
778 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
780 An invalid pointer dereference on read can be triggered when an
781 application tries to load malformed PKCS7 data with the
782 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
784 The result of the dereference is an application crash which could
785 lead to a denial of service attack. The TLS implementation in OpenSSL
786 does not call this function however third party applications might
787 call these functions on untrusted data.
792 * Fixed Use-after-free following BIO_new_NDEF.
794 The public API function BIO_new_NDEF is a helper function used for
795 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
796 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
797 be called directly by end user applications.
799 The function receives a BIO from the caller, prepends a new BIO_f_asn1
800 filter BIO onto the front of it to form a BIO chain, and then returns
801 the new head of the BIO chain to the caller. Under certain conditions,
802 for example if a CMS recipient public key is invalid, the new filter BIO
803 is freed and the function returns a NULL result indicating a failure.
804 However, in this case, the BIO chain is not properly cleaned up and the
805 BIO passed by the caller still retains internal pointers to the previously
806 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
807 then a use-after-free will occur. This will most likely result in a crash.
810 *Viktor Dukhovni, Matt Caswell*
812 * Fixed Double free after calling PEM_read_bio_ex.
814 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
815 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
816 data. If the function succeeds then the "name_out", "header" and "data"
817 arguments are populated with pointers to buffers containing the relevant
818 decoded data. The caller is responsible for freeing those buffers. It is
819 possible to construct a PEM file that results in 0 bytes of payload data.
820 In this case PEM_read_bio_ex() will return a failure code but will populate
821 the header argument with a pointer to a buffer that has already been freed.
822 If the caller also frees this buffer then a double free will occur. This
823 will most likely lead to a crash.
825 The functions PEM_read_bio() and PEM_read() are simple wrappers around
826 PEM_read_bio_ex() and therefore these functions are also directly affected.
828 These functions are also called indirectly by a number of other OpenSSL
829 functions including PEM_X509_INFO_read_bio_ex() and
830 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
831 internal uses of these functions are not vulnerable because the caller does
832 not free the header argument if PEM_read_bio_ex() returns a failure code.
835 *Kurt Roeckx, Matt Caswell*
837 * Fixed Timing Oracle in RSA Decryption.
839 A timing based side channel exists in the OpenSSL RSA Decryption
840 implementation which could be sufficient to recover a plaintext across
841 a network in a Bleichenbacher style attack. To achieve a successful
842 decryption an attacker would have to be able to send a very large number
843 of trial messages for decryption. The vulnerability affects all RSA padding
844 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
847 *Dmitry Belyavsky, Hubert Kario*
849 * Fixed X.509 Name Constraints Read Buffer Overflow.
851 A read buffer overrun can be triggered in X.509 certificate verification,
852 specifically in name constraint checking. The read buffer overrun might
853 result in a crash which could lead to a denial of service attack.
854 In a TLS client, this can be triggered by connecting to a malicious
855 server. In a TLS server, this can be triggered if the server requests
856 client authentication and a malicious client connects.
861 * Fixed X.509 Policy Constraints Double Locking security issue.
863 If an X.509 certificate contains a malformed policy constraint and
864 policy processing is enabled, then a write lock will be taken twice
865 recursively. On some operating systems (most widely: Windows) this
866 results in a denial of service when the affected process hangs. Policy
867 processing being enabled on a publicly facing server is not considered
868 to be a common setup.
873 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
874 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
875 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
876 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
877 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
878 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
879 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
880 for legacy EC and SM2 keys is also changed similarly to honor the
881 equivalent conversion format flag as specified in the underlying
882 `EC_KEY` object being exported to a provider, when this function is
883 called through `EVP_PKEY_export()`.
887 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
889 * Fixed two buffer overflows in punycode decoding functions.
891 A buffer overrun can be triggered in X.509 certificate verification,
892 specifically in name constraint checking. Note that this occurs after
893 certificate chain signature verification and requires either a CA to
894 have signed the malicious certificate or for the application to continue
895 certificate verification despite failure to construct a path to a trusted
898 In a TLS client, this can be triggered by connecting to a malicious
899 server. In a TLS server, this can be triggered if the server requests
900 client authentication and a malicious client connects.
902 An attacker can craft a malicious email address to overflow
903 an arbitrary number of bytes containing the `.` character (decimal 46)
904 on the stack. This buffer overflow could result in a crash (causing a
908 An attacker can craft a malicious email address to overflow four
909 attacker-controlled bytes on the stack. This buffer overflow could
910 result in a crash (causing a denial of service) or potentially remote code
911 execution depending on stack layout for any given platform/compiler.
916 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
917 parameters in OpenSSL code.
918 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
919 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
920 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
921 Using these invalid names may cause algorithms to use slower methods
922 that ignore the CRT parameters.
926 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
931 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
932 data to be signed before signing the certificate.
936 * Added RIPEMD160 to the default provider.
940 * Ensured that the key share group sent or accepted for the key exchange
941 is allowed for the protocol version.
945 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
947 * OpenSSL supports creating a custom cipher via the legacy
948 EVP_CIPHER_meth_new() function and associated function calls. This function
949 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
950 to use the new provider mechanism in order to implement custom ciphers.
952 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
953 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
954 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
955 and decryption initialisation functions). Instead of using the custom cipher
956 directly it incorrectly tries to fetch an equivalent cipher from the
957 available providers. An equivalent cipher is found based on the NID passed to
958 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
959 given cipher. However it is possible for an application to incorrectly pass
960 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
961 is used in this way the OpenSSL encryption/decryption initialisation function
962 will match the NULL cipher as being equivalent and will fetch this from the
963 available providers. This will succeed if the default provider has been
964 loaded (or if a third party provider has been loaded that offers this
965 cipher). Using the NULL cipher means that the plaintext is emitted as the
968 Applications are only affected by this issue if they call
969 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
970 encryption/decryption initialisation function. Applications that only use
971 SSL/TLS are not impacted by this issue.
976 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
981 * Fixed the linux-mips64 Configure target which was missing the
982 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
987 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
992 * Correctly handle a retransmitted ClientHello in DTLS
996 * Fixed detection of ktls support in cross-compile environment on Linux
1000 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
1005 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
1006 report correct results in some cases
1010 * Fix UWP builds by defining VirtualLock
1014 * For known safe primes use the minimum key length according to RFC 7919.
1015 Longer private key sizes unnecessarily raise the cycles needed to compute the
1016 shared secret without any increase of the real security. This fixes a
1017 regression from 1.1.1 where these shorter keys were generated for the known
1022 * Added the loongarch64 target
1026 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
1027 only passed to the FIPS provider and not to the default or legacy provider.
1031 * Fixed reported performance degradation on aarch64. Restored the
1032 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
1033 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
1034 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
1035 The new algorithm is still used for 32 bit targets.
1039 * Added a missing header for memcmp that caused compilation failure on some
1044 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
1046 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
1047 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
1048 This issue makes the RSA implementation with 2048 bit private keys
1049 incorrect on such machines and memory corruption will happen during
1050 the computation. As a consequence of the memory corruption an attacker
1051 may be able to trigger a remote code execution on the machine performing
1054 SSL/TLS servers or other servers using 2048 bit RSA private keys running
1055 on machines supporting AVX512IFMA instructions of the X86_64 architecture
1056 are affected by this issue.
1061 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
1062 implementation would not encrypt the entirety of the data under some
1063 circumstances. This could reveal sixteen bytes of data that was
1064 preexisting in the memory that wasn't written. In the special case of
1065 "in place" encryption, sixteen bytes of the plaintext would be revealed.
1067 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
1068 they are both unaffected.
1071 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
1073 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
1075 * In addition to the c_rehash shell command injection identified in
1076 CVE-2022-1292, further bugs where the c_rehash script does not
1077 properly sanitise shell metacharacters to prevent command injection have been
1080 When the CVE-2022-1292 was fixed it was not discovered that there
1081 are other places in the script where the file names of certificates
1082 being hashed were possibly passed to a command executed through the shell.
1084 This script is distributed by some operating systems in a manner where
1085 it is automatically executed. On such operating systems, an attacker
1086 could execute arbitrary commands with the privileges of the script.
1088 Use of the c_rehash script is considered obsolete and should be replaced
1089 by the OpenSSL rehash command line tool.
1092 *Daniel Fiala, Tomáš Mráz*
1094 * Case insensitive string comparison no longer uses locales. It has instead
1095 been directly implemented.
1099 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
1101 * Case insensitive string comparison is reimplemented via new locale-agnostic
1102 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
1103 comparison. The previous implementation had problems when the Turkish locale
1108 * Fixed a bug in the c_rehash script which was not properly sanitising shell
1109 metacharacters to prevent command injection. This script is distributed by
1110 some operating systems in a manner where it is automatically executed. On
1111 such operating systems, an attacker could execute arbitrary commands with the
1112 privileges of the script.
1114 Use of the c_rehash script is considered obsolete and should be replaced
1115 by the OpenSSL rehash command line tool.
1120 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
1121 certificate on an OCSP response. The bug caused the function in the case
1122 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
1123 response (meaning a successful verification) even in the case where the
1124 response signing certificate fails to verify.
1126 It is anticipated that most users of `OCSP_basic_verify` will not use the
1127 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
1128 a negative value (indicating a fatal error) in the case of a certificate
1129 verification failure. The normal expected return value in this case would be
1132 This issue also impacts the command line OpenSSL "ocsp" application. When
1133 verifying an ocsp response with the "-no_cert_checks" option the command line
1134 application will report that the verification is successful even though it
1135 has in fact failed. In this case the incorrect successful response will also
1136 be accompanied by error messages showing the failure and contradicting the
1137 apparently successful result.
1142 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
1143 AAD data as the MAC key. This made the MAC key trivially predictable.
1145 An attacker could exploit this issue by performing a man-in-the-middle attack
1146 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
1147 that the modified data would still pass the MAC integrity check.
1149 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
1150 endpoint will always be rejected by the recipient and the connection will
1151 fail at that point. Many application protocols require data to be sent from
1152 the client to the server first. Therefore, in such a case, only an OpenSSL
1153 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
1155 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
1156 sent in both directions. In this case both clients and servers could be
1157 affected, regardless of the application protocol.
1159 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
1160 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
1161 the handshake when using this ciphersuite.
1163 The confidentiality of data is not impacted by this issue, i.e. an attacker
1164 cannot decrypt data that has been encrypted using this ciphersuite - they can
1167 In order for this attack to work both endpoints must legitimately negotiate
1168 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
1169 OpenSSL 3.0, and is not available within the default provider or the default
1170 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
1171 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
1172 following must have occurred:
1174 1) OpenSSL must have been compiled with the (non-default) compile time option
1175 enable-weak-ssl-ciphers
1177 2) OpenSSL must have had the legacy provider explicitly loaded (either
1178 through application code or via configuration)
1180 3) The ciphersuite must have been explicitly added to the ciphersuite list
1182 4) The libssl security level must have been set to 0 (default is 1)
1184 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
1186 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
1187 others that both endpoints have in common
1192 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
1193 occupied by the removed hash table entries.
1195 This function is used when decoding certificates or keys. If a long lived
1196 process periodically decodes certificates or keys its memory usage will
1197 expand without bounds and the process might be terminated by the operating
1198 system causing a denial of service. Also traversing the empty hash table
1199 entries will take increasingly more time.
1201 Typically such long lived processes might be TLS clients or TLS servers
1202 configured to accept client certificate authentication.
1205 *Hugo Landau, Aliaksei Levin*
1207 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
1208 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
1209 statistics are no longer supported. For compatibility, these statistics are
1210 still listed in the output but are now always reported as zero.
1214 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
1216 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
1217 for non-prime moduli.
1219 Internally this function is used when parsing certificates that contain
1220 elliptic curve public keys in compressed form or explicit elliptic curve
1221 parameters with a base point encoded in compressed form.
1223 It is possible to trigger the infinite loop by crafting a certificate that
1224 has invalid explicit curve parameters.
1226 Since certificate parsing happens prior to verification of the certificate
1227 signature, any process that parses an externally supplied certificate may thus
1228 be subject to a denial of service attack. The infinite loop can also be
1229 reached when parsing crafted private keys as they can contain explicit
1230 elliptic curve parameters.
1232 Thus vulnerable situations include:
1234 - TLS clients consuming server certificates
1235 - TLS servers consuming client certificates
1236 - Hosting providers taking certificates or private keys from customers
1237 - Certificate authorities parsing certification requests from subscribers
1238 - Anything else which parses ASN.1 elliptic curve parameters
1240 Also any other applications that use the BN_mod_sqrt() where the attacker
1241 can control the parameter values are vulnerable to this DoS issue.
1246 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
1247 to the list of ciphersuites providing Perfect Forward Secrecy as
1248 required by SECLEVEL >= 3.
1250 *Dmitry Belyavskiy, Nicola Tuveri*
1252 * Made the AES constant time code for no-asm configurations
1253 optional due to the resulting 95% performance degradation.
1254 The AES constant time code can be enabled, for no assembly
1255 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1259 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
1264 * The negative return value handling of the certificate verification callback
1265 was reverted. The replacement is to set the verification retry state with
1266 the SSL_set_retry_verify() function.
1270 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
1272 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
1273 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
1274 verify a certificate supplied by a server. That function may return a
1275 negative return value to indicate an internal error (for example out of
1276 memory). Such a negative return value is mishandled by OpenSSL and will cause
1277 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
1278 success and a subsequent call to SSL_get_error() to return the value
1279 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
1280 returned by OpenSSL if the application has previously called
1281 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
1282 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
1283 totally unexpected and applications may not behave correctly as a result. The
1284 exact behaviour will depend on the application but it could result in
1285 crashes, infinite loops or other similar incorrect responses.
1287 This issue is made more serious in combination with a separate bug in OpenSSL
1288 3.0 that will cause X509_verify_cert() to indicate an internal error when
1289 processing a certificate chain. This will occur where a certificate does not
1290 include the Subject Alternative Name extension but where a Certificate
1291 Authority has enforced name constraints. This issue can occur even with valid
1297 * Corrected a few file name and file reference bugs in the build,
1298 installation and setup scripts, which lead to installation verification
1299 failures. Slightly enhanced the installation verification script.
1303 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
1308 * Fixed PVK encoder to properly query for the passphrase.
1312 * Multiple fixes in the OSSL_HTTP API functions.
1316 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
1317 OSSL_PARAM_INTEGER data type and return error on negative numbers
1318 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
1319 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
1323 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
1327 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
1331 * Multiple threading fixes.
1335 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
1339 * Allow fetching an operation from the provider that owns an unexportable key
1340 as a fallback if that is still allowed by the property query.
1344 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
1346 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
1351 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
1352 S390X capability vector to zero. This simplifies testing of different code
1353 paths on S390X architecture.
1357 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
1358 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
1359 SP 800-38D". The communication will fail at this point.
1363 * The EC_GROUP_clear_free() function is deprecated as there is nothing
1364 confidential in EC_GROUP data.
1368 * The byte order mark (BOM) character is ignored if encountered at the
1369 beginning of a PEM-formatted file.
1373 * Added CMS support for the Russian GOST algorithms.
1377 * Due to move of the implementation of cryptographic operations
1378 to the providers, validation of various operation parameters can
1379 be postponed until the actual operation is executed where previously
1380 it happened immediately when an operation parameter was set.
1382 For example when setting an unsupported curve with
1383 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
1384 fail but later keygen operations with the EVP_PKEY_CTX will fail.
1386 *OpenSSL team members and many third party contributors*
1388 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
1389 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
1390 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1391 instead to retrieve these algorithms from a provider.
1395 * On build targets where the multilib postfix is set in the build
1396 configuration the libdir directory was changing based on whether
1397 the lib directory with the multilib postfix exists on the system
1398 or not. This unpredictable behavior was removed and eventual
1399 multilib postfix is now always added to the default libdir. Use
1400 `--libdir=lib` to override the libdir if adding the postfix is
1405 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1406 no longer interoperable with OpenSSL 1.1.1.
1410 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1411 function codes, this function can only cause problems for calling
1416 * Add a configurable flag to output date formats as ISO 8601. Does not
1417 change the default date format.
1421 * Version of MSVC earlier than 1300 could get link warnings, which could
1422 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1423 Support for this flag has been removed.
1427 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1428 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1429 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1430 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1431 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1435 * The signatures of the functions to get and set options on SSL and
1436 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1437 Some source code changes may be required.
1441 * The public definitions of conf_method_st and conf_st have been
1442 deprecated. They will be made opaque in a future release.
1444 *Rich Salz and Tomáš Mráz*
1446 * Client-initiated renegotiation is disabled by default. To allow it, use
1447 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1448 flag, or the "ClientRenegotiation" config parameter as appropriate.
1452 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1453 or modify relative pathname inclusion.
1457 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1458 validated. Please consult the README-FIPS and
1459 README-PROVIDERS files, as well as the migration guide.
1461 *OpenSSL team members and many third party contributors*
1463 * For the key types DH and DHX the allowed settable parameters are now different.
1467 * The openssl commands that read keys, certificates, and CRLs now
1468 automatically detect the PEM or DER format of the input files.
1470 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1472 * Added enhanced PKCS#12 APIs which accept a library context.
1476 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1480 * Added support for Kernel TLS (KTLS).
1482 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1484 * Support for RFC 5746 secure renegotiation is now required by default for
1485 SSL or TLS connections to succeed.
1489 * The signature of the `copy` functional parameter of the
1490 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1491 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1492 the signature of the `pub_decode` functional parameter of the
1493 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1494 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1498 * The error return values from some control calls (ctrl) have changed.
1502 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1506 * Many functions in the EVP_ namespace that are getters of values from
1507 implementations or contexts were renamed to include get or get0 in their
1508 names. Old names are provided as macro aliases for compatibility and
1513 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1514 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1515 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1520 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1523 * The output from the command line applications may have minor
1528 * The output from numerous "printing" may have minor changes.
1532 * Windows thread synchronization uses read/write primitives (SRWLock) when
1533 supported by the OS, otherwise CriticalSection continues to be used.
1537 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1538 work on read only BIO source/sinks that do not support these functions.
1539 This allows piping or redirection of a file BIO using stdin to be buffered
1540 into memory. This is used internally in OSSL_DECODER_from_bio().
1544 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1545 this function would return one of the values OSSL_STORE_INFO_NAME,
1546 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1547 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1548 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1549 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1550 using this function should be amended to handle the changed return value.
1554 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1555 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1556 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1557 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1558 contains more than one certificate identifier: This means that all
1559 certificates referenced there MUST be part of the validation chain.
1563 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1564 RC5, DESX and DES have been moved to the legacy provider.
1568 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1569 RIPEMD-160 have been moved to the legacy provider.
1573 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1578 * The deprecated functions EVP_PKEY_get0_RSA(),
1579 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1580 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1581 well as the similarly named "get1" functions behave differently in
1586 * A number of functions handling low-level keys or engines were deprecated
1587 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1588 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1589 EVP_PKEY_get0_siphash().
1593 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1594 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1595 will need to load the legacy crypto provider. This includes these PBE
1596 algorithms which use this KDF:
1597 - NID_pbeWithMD2AndDES_CBC
1598 - NID_pbeWithMD5AndDES_CBC
1599 - NID_pbeWithSHA1AndRC2_CBC
1600 - NID_pbeWithMD2AndRC2_CBC
1601 - NID_pbeWithMD5AndRC2_CBC
1602 - NID_pbeWithSHA1AndDES_CBC
1606 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1607 BIO_debug_callback() functions.
1611 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1612 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1616 * The RAND_METHOD APIs have been deprecated.
1620 * The SRP APIs have been deprecated.
1624 * Add a compile time option to prevent the caching of provider fetched
1625 algorithms. This is enabled by including the no-cached-fetch option
1626 at configuration time.
1630 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1631 count of PKCS12_DEFAULT_ITER.
1633 *Tomáš Mráz and Sahana Prasad*
1635 * The openssl speed command does not use low-level API calls anymore.
1639 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1642 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1644 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1648 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1649 providers may supply their own group implementations (using either the "key
1650 exchange" or the "key encapsulation" methods) which will automatically be
1651 detected and used by libssl.
1653 *Matt Caswell, Nicola Tuveri*
1655 * The undocumented function X509_certificate_type() has been deprecated;
1659 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1663 * Removed RSA padding mode for SSLv23 (which was only used for
1664 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1665 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1670 * Deprecated the obsolete X9.31 RSA key generation related functions.
1672 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1673 is not allowed to return a value > 1, this is no more taken as failure.
1675 *Viktor Dukhovni and David von Oheimb*
1677 * Deprecated the obsolete X9.31 RSA key generation related functions
1678 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1679 BN_X931_generate_prime_ex().
1683 * The default key generation method for the regular 2-prime RSA keys was
1684 changed to the FIPS 186-4 B.3.6 method.
1688 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1692 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1696 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1697 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1699 *Rich Salz, Richard Levitte, and David von Oheimb*
1701 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1705 * Deprecated `OCSP_parse_url()`.
1709 * Validation of SM2 keys has been separated from the validation of regular EC
1714 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1715 switches: a validation failure triggers an early exit, returning a failure
1716 exit status to the parent process.
1720 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1721 to ignore unknown ciphers.
1725 * The `-cipher-commands` and `-digest-commands` options
1726 of the command line utility `list` have been deprecated.
1727 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1731 * Added convenience functions for generating asymmetric key pairs:
1732 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1733 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1737 * All of the low-level EC_KEY functions have been deprecated.
1739 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1741 * Deprecated all the libcrypto and libssl error string loading
1746 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1747 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1752 * The `-crypt` option to the `passwd` command line tool has been removed.
1756 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1761 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1765 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1766 EVP_PKEY_get1_tls_encodedpoint().
1770 * The security callback, which can be customised by application code, supports
1771 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1772 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1776 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1777 interface. Their functionality remains unchanged.
1781 * Added new option for 'openssl list', '-providers', which will display the
1782 list of loaded providers, their names, version and status. It optionally
1783 displays their gettable parameters.
1787 * Removed EVP_PKEY_set_alias_type().
1791 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1792 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1796 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1797 parameters are still checked for type safety at compile time via helper
1802 * Remove the RAND_DRBG API
1804 *Paul Dale and Matthias St. Pierre*
1806 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1807 as well as actual hostnames.
1811 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1812 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1813 conversely, silently ignore DTLS protocol version bounds when configuring
1814 TLS-based contexts. The commands can be repeated to set bounds of both
1815 types. The same applies with the corresponding "min_protocol" and
1816 "max_protocol" command-line switches, in case some application uses both TLS
1819 SSL_CTX instances that are created for a fixed protocol version (e.g.
1820 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1821 attempts to apply bounds to these protocol versions would result in an
1822 error. Now only the "version-flexible" SSL_CTX instances are subject to
1823 limits in configuration files in command-line options.
1827 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1832 * Reworked the recorded ERR codes to make better space for system errors.
1833 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1834 given code is a system error (true) or an OpenSSL error (false).
1838 * Reworked the test perl framework to better allow parallel testing.
1840 *Nicola Tuveri and David von Oheimb*
1842 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1843 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1847 * 'Configure' has been changed to figure out the configuration target if
1848 none is given on the command line. Consequently, the 'config' script is
1849 now only a mere wrapper. All documentation is changed to only mention
1852 *Rich Salz and Richard Levitte*
1854 * Added a library context `OSSL_LIB_CTX` that applications as well as
1855 other libraries can use to form a separate context within which
1856 libcrypto operations are performed.
1860 * Added various `_ex` functions to the OpenSSL API that support using
1861 a non-default `OSSL_LIB_CTX`.
1865 * Handshake now fails if Extended Master Secret extension is dropped
1870 * Dropped interactive mode from the `openssl` program.
1874 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1876 *David von Oheimb and Shane Lontis*
1878 * Deprecated `EC_METHOD_get_field_type()`.
1882 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1883 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1884 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1888 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1892 * Add CAdES-BES signature verification support, mostly derived
1893 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1895 *Filipe Raimundo da Silva*
1897 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1901 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1902 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1906 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1910 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1911 EC_KEY_precompute_mult().
1915 * Deprecated EC_POINTs_mul().
1919 * Removed FIPS_mode() and FIPS_mode_set().
1923 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1927 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1928 EC_POINT_get_Jprojective_coordinates_GFp().
1932 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1933 arrays to be more easily constructed via a series of utility functions.
1934 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1935 the various push functions and finally convert to a passable OSSL_PARAM
1936 array using OSSL_PARAM_BLD_to_param().
1940 * The security strength of SHA1 and MD5 based signatures in TLS has been
1945 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1946 contain a provider side internal key.
1950 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1954 * Project text documents not yet having a proper file name extension
1955 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1956 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1960 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1961 have been converted to Markdown with the goal to produce documents
1962 which not only look pretty when viewed online in the browser, but
1963 remain well readable inside a plain text editor.
1965 To achieve this goal, a 'minimalistic' Markdown style has been applied
1966 which avoids formatting elements that interfere too much with the
1967 reading flow in the text file. For example, it
1969 * avoids [ATX headings][] and uses [setext headings][] instead
1970 (which works for `<h1>` and `<h2>` headings only).
1971 * avoids [inline links][] and uses [reference links][] instead.
1972 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1974 [ATX headings]: https://github.github.com/gfm/#atx-headings
1975 [setext headings]: https://github.github.com/gfm/#setext-headings
1976 [inline links]: https://github.github.com/gfm/#inline-link
1977 [reference links]: https://github.github.com/gfm/#reference-link
1978 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1979 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1981 *Matthias St. Pierre*
1983 * The test suite is changed to preserve results of each test recipe.
1984 A new directory test-runs/ with subdirectories named like the
1985 test recipes are created in the build tree for this purpose.
1989 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1990 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1991 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1993 *David von Oheimb, Martin Peylo*
1995 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1996 It supports arbitrary request and response content types, GET redirection,
1997 TLS, connections via HTTP(S) proxies, connections and exchange via
1998 user-defined BIOs (allowing implicit connections), persistent connections,
1999 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
2000 The legacy OCSP-focused (and only partly documented) API
2001 is retained for backward compatibility, while most of it is deprecated.
2005 * Added `util/check-format.pl`, a tool for checking adherence to the
2006 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
2007 The checks performed are incomplete and yield some false positives.
2008 Still the tool should be useful for detecting most typical glitches.
2012 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
2013 If domain name resolution yields multiple IP addresses all of them are tried
2014 after `connect()` failures.
2018 * All of the low-level RSA functions have been deprecated.
2022 * X509 certificates signed using SHA1 are no longer allowed at security
2027 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
2028 modified to use PKEY APIs. These commands are now in maintenance mode
2029 and no new features will be added to them.
2033 * The command line utility rsautl has been deprecated.
2037 * The command line utilities genrsa and rsa have been modified to use PKEY
2038 APIs. They now write PKCS#8 keys by default. These commands are now in
2039 maintenance mode and no new features will be added to them.
2043 * All of the low-level DH functions have been deprecated.
2045 *Paul Dale and Matt Caswell*
2047 * All of the low-level DSA functions have been deprecated.
2051 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
2052 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
2056 * Deprecated low-level ECDH and ECDSA functions.
2060 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
2064 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
2065 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
2066 a new formulation to include all the things it can be used for,
2067 as well as words of caution.
2071 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
2075 * All of the low-level HMAC functions have been deprecated.
2077 *Paul Dale and David von Oheimb*
2079 * Over two thousand fixes were made to the documentation, including:
2080 - Common options (such as -rand/-writerand, TLS version control, etc)
2081 were refactored and point to newly-enhanced descriptions in openssl.pod.
2082 - Added style conformance for all options (with help from Richard Levitte),
2083 documented all reported missing options, added a CI build to check
2084 that all options are documented and that no unimplemented options
2086 - Documented some internals, such as all use of environment variables.
2087 - Addressed all internal broken L<> references.
2091 * All of the low-level CMAC functions have been deprecated.
2095 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
2096 functions have been deprecated.
2098 *Paul Dale and David von Oheimb*
2100 * Corrected the documentation of the return values from the `EVP_DigestSign*`
2101 set of functions. The documentation mentioned negative values for some
2102 errors, but this was never the case, so the mention of negative values
2105 Code that followed the documentation and thereby check with something
2106 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
2110 * All of the low-level cipher functions have been deprecated.
2112 *Matt Caswell and Paul Dale*
2114 * Removed include/openssl/opensslconf.h.in and replaced it with
2115 include/openssl/configuration.h.in, which differs in not including
2116 <openssl/macros.h>. A short header include/openssl/opensslconf.h
2117 was added to include both.
2119 This allows internal hacks where one might need to modify the set
2120 of configured macros, for example this if deprecated symbols are
2121 still supposed to be available internally:
2123 #include <openssl/configuration.h>
2125 #undef OPENSSL_NO_DEPRECATED
2126 #define OPENSSL_SUPPRESS_DEPRECATED
2128 #include <openssl/macros.h>
2130 This should not be used by applications that use the exported
2131 symbols, as that will lead to linking errors.
2135 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
2136 used in exponentiation with 512-bit moduli. No EC algorithms are
2137 affected. Analysis suggests that attacks against 2-prime RSA1024,
2138 3-prime RSA1536, and DSA1024 as a result of this defect would be very
2139 difficult to perform and are not believed likely. Attacks against DH512
2140 are considered just feasible. However, for an attack the target would
2141 have to reuse the DH512 private key, which is not recommended anyway.
2142 Also applications directly using the low-level API BN_mod_exp may be
2143 affected if they use BN_FLG_CONSTTIME.
2148 * Most memory-debug features have been deprecated, and the functionality
2149 replaced with no-ops.
2153 * Added documentation for the STACK API.
2157 * Introduced a new method type and API, OSSL_ENCODER, to represent
2158 generic encoders. These do the same sort of job that PEM writers
2159 and d2i functions do, but with support for methods supplied by
2160 providers, and the possibility for providers to support other
2165 * Introduced a new method type and API, OSSL_DECODER, to represent
2166 generic decoders. These do the same sort of job that PEM readers
2167 and i2d functions do, but with support for methods supplied by
2168 providers, and the possibility for providers to support other
2173 * Added a .pragma directive to the syntax of configuration files, to
2174 allow varying behavior in a supported and predictable manner.
2175 Currently added pragma:
2179 This allows dollar signs to be a keyword character unless it's
2180 followed by a opening brace or parenthesis. This is useful for
2181 platforms where dollar signs are commonly used in names, such as
2182 volume names and system directory names on VMS.
2186 * Added functionality to create an EVP_PKEY from user data.
2190 * Change the interpretation of the '--api' configuration option to
2191 mean that this is a desired API compatibility level with no
2192 further meaning. The previous interpretation, that this would
2193 also mean to remove all deprecated symbols up to and including
2194 the given version, no requires that 'no-deprecated' is also used
2195 in the configuration.
2197 When building applications, the desired API compatibility level
2198 can be set with the OPENSSL_API_COMPAT macro like before. For
2199 API compatibility version below 3.0, the old style numerical
2200 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
2201 For version 3.0 and on, the value is expected to be the decimal
2202 value calculated from the major and minor version like this:
2204 MAJOR * 10000 + MINOR * 100
2208 -DOPENSSL_API_COMPAT=30000 For 3.0
2209 -DOPENSSL_API_COMPAT=30200 For 3.2
2211 To hide declarations that are deprecated up to and including the
2212 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
2213 given when building the application as well.
2217 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
2218 access to certificate and CRL stores via URIs and OSSL_STORE
2221 This adds the following functions:
2223 - X509_LOOKUP_store()
2224 - X509_STORE_load_file()
2225 - X509_STORE_load_path()
2226 - X509_STORE_load_store()
2227 - SSL_add_store_cert_subjects_to_stack()
2228 - SSL_CTX_set_default_verify_store()
2229 - SSL_CTX_load_verify_file()
2230 - SSL_CTX_load_verify_dir()
2231 - SSL_CTX_load_verify_store()
2235 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2236 The presence of this system service is determined at run-time.
2240 * Added functionality to create an EVP_PKEY context based on data
2241 for methods from providers. This takes an algorithm name and a
2242 property query string and simply stores them, with the intent
2243 that any operation that uses this context will use those strings
2244 to fetch the needed methods implicitly, thereby making the port
2245 of application written for pre-3.0 OpenSSL easier.
2249 * The undocumented function NCONF_WIN32() has been deprecated; for
2250 conversion details see the HISTORY section of doc/man5/config.pod
2254 * Introduced the new functions EVP_DigestSignInit_ex() and
2255 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
2256 EVP_DigestVerifyUpdate() have been converted to functions. See the man
2257 pages for further details.
2261 * Over two thousand fixes were made to the documentation, including:
2262 adding missing command flags, better style conformance, documentation
2265 *Rich Salz, Richard Levitte*
2267 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
2268 X25519, X448, Ed25519 and Ed448.
2272 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2277 * Deprecated the public definition of `ERR_STATE` as well as the function
2278 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
2283 * Added ERR functionality to give callers access to the stored function
2284 names that have replaced the older function code based functions.
2286 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
2287 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
2288 ERR_peek_error_all() and ERR_peek_last_error_all().
2290 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
2291 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
2292 ERR_func_error_string().
2296 * Extended testing to be verbose for failing tests only. The make variables
2297 VERBOSE_FAILURE or VF can be used to enable this:
2299 $ make VF=1 test # Unix
2300 $ mms /macro=(VF=1) test ! OpenVMS
2301 $ nmake VF=1 test # Windows
2305 * Added the `-copy_extensions` option to the `x509` command for use with
2306 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
2307 all extensions in the request are copied to the certificate or vice versa.
2309 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
2311 * Added the `-copy_extensions` option to the `req` command for use with
2312 `-x509`. When given with the `copy` or `copyall` argument,
2313 all extensions in the certification request are copied to the certificate.
2317 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
2318 they generate are by default RFC 5280 compliant in the following sense:
2319 There is a subjectKeyIdentifier extension with a hash value of the public key
2320 and for not self-signed certs there is an authorityKeyIdentifier extension
2321 with a keyIdentifier field or issuer information identifying the signing key.
2322 This is done unless some configuration overrides the new default behavior,
2323 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
2327 * Added several checks to `X509_verify_cert()` according to requirements in
2328 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
2329 (which may be done by using the CLI option `-x509_strict`):
2330 * The basicConstraints of CA certificates must be marked critical.
2331 * CA certificates must explicitly include the keyUsage extension.
2332 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
2333 * The issuer name of any certificate must not be empty.
2334 * The subject name of CA certs, certs with keyUsage crlSign,
2335 and certs without subjectAlternativeName must not be empty.
2336 * If a subjectAlternativeName extension is given it must not be empty.
2337 * The signatureAlgorithm field and the cert signature must be consistent.
2338 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
2339 must not be marked critical.
2340 * The authorityKeyIdentifier must be given for X.509v3 certs
2341 unless they are self-signed.
2342 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
2346 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
2347 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
2351 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2352 used even when parsing explicit parameters, when loading a encoded key
2353 or calling `EC_GROUP_new_from_ecpkparameters()`/
2354 `EC_GROUP_new_from_ecparameters()`.
2355 This prevents bypass of security hardening and performance gains,
2356 especially for curves with specialized EC_METHODs.
2357 By default, if a key encoded with explicit parameters is loaded and later
2358 encoded, the output is still encoded with explicit parameters, even if
2359 internally a "named" EC_GROUP is used for computation.
2363 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2364 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2365 NULL. After this change, only the cofactor parameter can be NULL. It also
2366 does some minimal sanity checks on the passed order.
2371 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2372 An attack is simple, if the first CMS_recipientInfo is valid but the
2373 second CMS_recipientInfo is chosen ciphertext. If the second
2374 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2375 encryption key will be replaced by garbage, and the message cannot be
2376 decoded, but if the RSA decryption fails, the correct encryption key is
2377 used and the recipient will not notice the attack.
2378 As a work around for this potential attack the length of the decrypted
2379 key must be equal to the cipher default key length, in case the
2380 certificate is not given and all recipientInfo are tried out.
2381 The old behaviour can be re-enabled in the CMS code by setting the
2382 CMS_DEBUG_DECRYPT flag.
2386 * Early start up entropy quality from the DEVRANDOM seed source has been
2387 improved for older Linux systems. The RAND subsystem will wait for
2388 /dev/random to be producing output before seeding from /dev/urandom.
2389 The seeded state is stored for future library initialisations using
2390 a system global shared memory segment. The shared memory identifier
2391 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2392 the desired value. The default identifier is 114.
2396 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2397 when primes for RSA keys are computed.
2398 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2399 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2400 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2401 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2402 This avoids possible fingerprinting of newly generated RSA modules.
2406 * Correct the extended master secret constant on EBCDIC systems. Without this
2407 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2408 negotiate EMS will fail. Unfortunately this also means that TLS connections
2409 between EBCDIC systems with this fix, and EBCDIC systems without this
2410 fix will fail if they negotiate EMS.
2414 * Changed the library initialisation so that the config file is now loaded
2415 by default. This was already the case for libssl. It now occurs for both
2416 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2417 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2421 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2422 where the former acts as a replacement for `ERR_put_error()`, and the
2423 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2424 `ERR_raise_data()` adds more flexibility by taking a format string and
2425 an arbitrary number of arguments following it, to be processed with
2430 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2431 to check if a named provider is loaded and available. When called, it
2432 will also activate all fallback providers if such are still present.
2436 * Enforce a minimum DH modulus size of 512 bits.
2440 * Changed DH parameters to generate the order q subgroup instead of 2q.
2441 Previously generated DH parameters are still accepted by DH_check
2442 but DH_generate_key works around that by clearing bit 0 of the
2443 private key for those. This avoids leaking bit 0 of the private key.
2447 * Significantly reduce secure memory usage by the randomness pools.
2451 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2456 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2457 algorithms. An implementation of a key exchange algorithm can be obtained
2458 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2459 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2460 the older EVP_PKEY_derive_init() function. See the man pages for the new
2461 functions for further details.
2465 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2469 * Removed the function names from error messages and deprecated the
2474 * Removed NextStep support and the macro OPENSSL_UNISTD
2478 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2479 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2480 Also removed "export var as function" capability; we do not export
2481 variables, only functions.
2485 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2486 an error and 1 indicating success. In previous versions of OpenSSL this
2487 was a void type. If a key was set longer than the maximum possible this
2492 * Support SM2 signing and verification schemes with X509 certificate.
2496 * Use SHA256 as the default digest for TS query in the `ts` app.
2500 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2504 * Default cipher lists/suites are now available via a function, the
2505 #defines are deprecated.
2509 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2510 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2511 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2515 * Join the directories crypto/x509 and crypto/x509v3
2519 * Added command 'openssl kdf' that uses the EVP_KDF API.
2523 * Added command 'openssl mac' that uses the EVP_MAC API.
2527 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2528 as default directories. Also added the command 'openssl info'
2529 for scripting purposes.
2533 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2538 * Add prediction resistance to the DRBG reseeding process.
2542 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2543 mandated by IEEE Std 1619-2018.
2547 * Added newline escaping functionality to a filename when using openssl dgst.
2548 This output format is to replicate the output format found in the `*sum`
2549 checksum programs. This aims to preserve backward compatibility.
2551 *Matt Eaton, Richard Levitte, and Paul Dale*
2553 * Removed the heartbeat message in DTLS feature, as it has very
2554 little usage and doesn't seem to fulfill a valuable purpose.
2555 The configuration option is now deprecated.
2559 * Changed the output of 'openssl {digestname} < file' to display the
2560 digest name in its output.
2564 * Added a new generic trace API which provides support for enabling
2565 instrumentation through trace output.
2567 *Richard Levitte & Matthias St. Pierre*
2569 * Added build tests for C++. These are generated files that only do one
2570 thing, to include one public OpenSSL head file each. This tests that
2571 the public header files can be usefully included in a C++ application.
2573 This test isn't enabled by default. It can be enabled with the option
2574 'enable-buildtest-c++'.
2578 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2582 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2586 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2590 * Added KMAC to EVP_MAC.
2594 * Added property based algorithm implementation selection framework to
2599 * Added SCA hardening for modular field inversion in EC_GROUP through
2600 a new dedicated field_inv() pointer in EC_METHOD.
2601 This also addresses a leakage affecting conversions from projective
2602 to affine coordinates.
2604 *Billy Bob Brumley, Nicola Tuveri*
2606 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2607 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2608 those algorithms that were already supported through the EVP_PKEY API
2609 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2610 and scrypt are now wrappers that call EVP_KDF.
2614 * Build devcrypto engine as a dynamic engine.
2616 *Eneas U de Queiroz*
2618 * Add keyed BLAKE2 to EVP_MAC.
2622 * Fix a bug in the computation of the endpoint-pair shared secret used
2623 by DTLS over SCTP. This breaks interoperability with older versions
2624 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2625 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2626 interoperability with such broken implementations. However, enabling
2627 this switch breaks interoperability with correct implementations.
2629 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2630 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2634 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2638 * Changed the license to the Apache License v2.0.
2642 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2644 - Major releases (indicated by incrementing the MAJOR release number)
2645 may introduce incompatible API/ABI changes.
2646 - Minor releases (indicated by incrementing the MINOR release number)
2647 may introduce new features but retain API/ABI compatibility.
2648 - Patch releases (indicated by incrementing the PATCH number)
2649 are intended for bug fixes and other improvements of existing
2650 features only (like improving performance or adding documentation)
2651 and retain API/ABI compatibility.
2655 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2659 * Remove the 'dist' target and add a tarball building script. The
2660 'dist' target has fallen out of use, and it shouldn't be
2661 necessary to configure just to create a source distribution.
2665 * Recreate the OS390-Unix config target. It no longer relies on a
2666 special script like it did for OpenSSL pre-1.1.0.
2670 * Instead of having the source directories listed in Configure, add
2671 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2676 * Add GMAC to EVP_MAC.
2680 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2684 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2685 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2686 to facilitate the continued use of MACs through raw private keys in
2687 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2691 * Deprecate ECDH_KDF_X9_62().
2695 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2696 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2697 are retained for backwards compatibility.
2701 * AES-XTS mode now enforces that its two keys are different to mitigate
2702 the attacked described in "Efficient Instantiations of Tweakable
2703 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2704 Details of this attack can be obtained from:
2705 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2709 * Rename the object files, i.e. give them other names than in previous
2710 versions. Their names now include the name of the final product, as
2711 well as its type mnemonic (bin, lib, shlib).
2715 * Added new option for 'openssl list', '-objects', which will display the
2716 list of built in objects, i.e. OIDs with names.
2720 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2721 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2726 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2727 improves application performance by removing data copies and providing
2728 applications with zero-copy system calls such as sendfile and splice.
2732 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2736 * `PKCS12_parse` now maintains the order of the parsed certificates
2737 when outputting them via `*ca` (rather than reversing it).
2741 * Deprecated pthread fork support methods.
2745 * Added support for FFDHE key exchange in TLS 1.3.
2749 * Added a new concept for OpenSSL plugability: providers. This
2750 functionality is designed to replace the ENGINE API and ENGINE
2751 implementations, and to be much more dynamic, allowing provider
2752 authors to introduce new algorithms among other things, as long as
2753 there's an API that supports the algorithm type.
2755 With this concept comes a new core API for interaction between
2756 libcrypto and provider implementations. Public libcrypto functions
2757 that want to use providers do so through this core API.
2759 The main documentation for this core API is found in
2760 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2761 refer to other manuals describing the API specific for supported
2762 algorithm types (also called operations).
2769 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2771 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2773 * Avoid loading of a dynamic engine twice.
2777 * Prioritise DANE TLSA issuer certs over peer certs
2781 * Fixed random API for MacOS prior to 10.12
2783 These MacOS versions don't support the CommonCrypto APIs
2787 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2789 * Fixed an SM2 Decryption Buffer Overflow.
2791 In order to decrypt SM2 encrypted data an application is expected to
2792 call the API function EVP_PKEY_decrypt(). Typically an application will
2793 call this function twice. The first time, on entry, the "out" parameter
2794 can be NULL and, on exit, the "outlen" parameter is populated with the
2795 buffer size required to hold the decrypted plaintext. The application
2796 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2797 again, but this time passing a non-NULL value for the "out" parameter.
2799 A bug in the implementation of the SM2 decryption code means that the
2800 calculation of the buffer size required to hold the plaintext returned
2801 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2802 size required by the second call. This can lead to a buffer overflow
2803 when EVP_PKEY_decrypt() is called by the application a second time with
2804 a buffer that is too small.
2806 A malicious attacker who is able present SM2 content for decryption to
2807 an application could cause attacker chosen data to overflow the buffer
2808 by up to a maximum of 62 bytes altering the contents of other data held
2809 after the buffer, possibly changing application behaviour or causing
2810 the application to crash. The location of the buffer is application
2811 dependent but is typically heap allocated.
2816 * Fixed various read buffer overruns processing ASN.1 strings
2818 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2819 structure which contains a buffer holding the string data and a field
2820 holding the buffer length. This contrasts with normal C strings which
2821 are represented as a buffer for the string data which is terminated
2822 with a NUL (0) byte.
2824 Although not a strict requirement, ASN.1 strings that are parsed using
2825 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2826 well as any string whose value has been set with the ASN1_STRING_set()
2827 function will additionally NUL terminate the byte array in the
2828 ASN1_STRING structure.
2830 However, it is possible for applications to directly construct valid
2831 ASN1_STRING structures which do not NUL terminate the byte array by
2832 directly setting the "data" and "length" fields in the ASN1_STRING
2833 array. This can also happen by using the ASN1_STRING_set0() function.
2835 Numerous OpenSSL functions that print ASN.1 data have been found to
2836 assume that the ASN1_STRING byte array will be NUL terminated, even
2837 though this is not guaranteed for strings that have been directly
2838 constructed. Where an application requests an ASN.1 structure to be
2839 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2840 been directly constructed by the application without NUL terminating
2841 the "data" field, then a read buffer overrun can occur.
2843 The same thing can also occur during name constraints processing
2844 of certificates (for example if a certificate has been directly
2845 constructed by the application instead of loading it via the OpenSSL
2846 parsing functions, and the certificate contains non NUL terminated
2847 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2848 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2850 If a malicious actor can cause an application to directly construct an
2851 ASN1_STRING and then process it through one of the affected OpenSSL
2852 functions then this issue could be hit. This might result in a crash
2853 (causing a Denial of Service attack). It could also result in the
2854 disclosure of private memory contents (such as private keys, or
2855 sensitive plaintext).
2860 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2862 * Fixed a problem with verifying a certificate chain when using the
2863 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2864 the certificates present in a certificate chain. It is not set by default.
2866 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2867 the chain that have explicitly encoded elliptic curve parameters was added
2868 as an additional strict check.
2870 An error in the implementation of this check meant that the result of a
2871 previous check to confirm that certificates in the chain are valid CA
2872 certificates was overwritten. This effectively bypasses the check
2873 that non-CA certificates must not be able to issue other certificates.
2875 If a "purpose" has been configured then there is a subsequent opportunity
2876 for checks that the certificate is a valid CA. All of the named "purpose"
2877 values implemented in libcrypto perform this check. Therefore, where
2878 a purpose is set the certificate chain will still be rejected even when the
2879 strict flag has been used. A purpose is set by default in libssl client and
2880 server certificate verification routines, but it can be overridden or
2881 removed by an application.
2883 In order to be affected, an application must explicitly set the
2884 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2885 for the certificate verification or, in the case of TLS client or server
2886 applications, override the default purpose.
2891 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2892 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2893 renegotiation ClientHello omits the signature_algorithms extension (where it
2894 was present in the initial ClientHello), but includes a
2895 signature_algorithms_cert extension then a NULL pointer dereference will
2896 result, leading to a crash and a denial of service attack.
2898 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2899 (which is the default configuration). OpenSSL TLS clients are not impacted by
2903 *Peter Kästle and Samuel Sapalski*
2905 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2907 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2908 create a unique hash value based on the issuer and serial number data
2909 contained within an X509 certificate. However, it was failing to correctly
2910 handle any errors that may occur while parsing the issuer field (which might
2911 occur if the issuer field is maliciously constructed). This may subsequently
2912 result in a NULL pointer deref and a crash leading to a potential denial of
2918 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2919 padding mode to correctly check for rollback attacks. This is considered a
2920 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2925 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2926 functions. Previously they could overflow the output length argument in some
2927 cases where the input length is close to the maximum permissible length for
2928 an integer on the platform. In such cases the return value from the function
2929 call would be 1 (indicating success), but the output length value would be
2930 negative. This could cause applications to behave incorrectly or crash.
2935 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2936 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2937 could be exploited in a side channel attack to recover the password. Since
2938 the attack is local host only this is outside of the current OpenSSL
2939 threat model and therefore no CVE is assigned.
2941 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2946 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2948 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2949 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2950 If an attacker can control both items being compared then this could lead
2951 to a possible denial of service attack. OpenSSL itself uses the
2952 GENERAL_NAME_cmp function for two purposes:
2953 1) Comparing CRL distribution point names between an available CRL and a
2954 CRL distribution point embedded in an X509 certificate
2955 2) When verifying that a timestamp response token signer matches the
2956 timestamp authority name (exposed via the API functions
2957 TS_RESP_verify_response and TS_RESP_verify_token)
2962 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2964 * Certificates with explicit curve parameters are now disallowed in
2965 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2969 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2970 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2971 conversely, silently ignore DTLS protocol version bounds when configuring
2972 TLS-based contexts. The commands can be repeated to set bounds of both
2973 types. The same applies with the corresponding "min_protocol" and
2974 "max_protocol" command-line switches, in case some application uses both TLS
2977 SSL_CTX instances that are created for a fixed protocol version (e.g.
2978 TLSv1_server_method()) also silently ignore version bounds. Previously
2979 attempts to apply bounds to these protocol versions would result in an
2980 error. Now only the "version-flexible" SSL_CTX instances are subject to
2981 limits in configuration files in command-line options.
2985 * Handshake now fails if Extended Master Secret extension is dropped
2990 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2992 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2994 * Fixed segmentation fault in SSL_check_chain()
2995 Server or client applications that call the SSL_check_chain() function
2996 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2997 dereference as a result of incorrect handling of the
2998 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2999 or unrecognised signature algorithm is received from the peer. This could
3000 be exploited by a malicious peer in a Denial of Service attack.
3005 * Added AES consttime code for no-asm configurations
3006 an optional constant time support for AES was added
3007 when building openssl for no-asm.
3008 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
3009 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
3010 At this time this feature is by default disabled.
3011 It will be enabled by default in 3.0.
3015 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
3017 * Revert the change of EOF detection while reading in libssl to avoid
3018 regressions in applications depending on the current way of reporting
3019 the EOF. As the existing method is not fully accurate the change to
3020 reporting the EOF via SSL_ERROR_SSL is kept on the current development
3021 branch and will be present in the 3.0 release.
3025 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
3026 when primes for RSA keys are computed.
3027 Since we previously always generated primes == 2 (mod 3) for RSA keys,
3028 the 2-prime and 3-prime RSA modules were easy to distinguish, since
3029 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
3030 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
3031 This avoids possible fingerprinting of newly generated RSA modules.
3035 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
3037 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
3038 while reading in libssl then we would report an error back to the
3039 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
3040 an error to the stack (which means we instead return SSL_ERROR_SSL) and
3041 therefore give a hint as to what went wrong.
3045 * Check that ed25519 and ed448 are allowed by the security level. Previously
3046 signature algorithms not using an MD were not being checked that they were
3047 allowed by the security level.
3051 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
3052 was not quite right. The behaviour was not consistent between resumption
3053 and normal handshakes, and also not quite consistent with historical
3054 behaviour. The behaviour in various scenarios has been clarified and
3055 it has been updated to make it match historical behaviour as closely as
3060 * *[VMS only]* The header files that the VMS compilers include automatically,
3061 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
3062 that the C++ compiler doesn't understand. This is a shortcoming in the
3063 compiler, but can be worked around with `__cplusplus` guards.
3065 C++ applications that use OpenSSL libraries must be compiled using the
3066 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
3067 functions. Otherwise, only functions with symbols of less than 31
3068 characters can be used, as the linker will not be able to successfully
3069 resolve symbols with longer names.
3073 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
3074 The presence of this system service is determined at run-time.
3078 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
3083 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
3085 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
3086 number generator (RNG). This was intended to include protection in the
3087 event of a fork() system call in order to ensure that the parent and child
3088 processes did not share the same RNG state. However, this protection was not
3089 being used in the default case.
3091 A partial mitigation for this issue is that the output from a high
3092 precision timer is mixed into the RNG state so the likelihood of a parent
3093 and child process sharing state is significantly reduced.
3095 If an application already calls OPENSSL_init_crypto() explicitly using
3096 OPENSSL_INIT_ATFORK then this problem does not occur at all.
3099 *Matthias St. Pierre*
3101 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3102 used even when parsing explicit parameters, when loading a encoded key
3103 or calling `EC_GROUP_new_from_ecpkparameters()`/
3104 `EC_GROUP_new_from_ecparameters()`.
3105 This prevents bypass of security hardening and performance gains,
3106 especially for curves with specialized EC_METHODs.
3107 By default, if a key encoded with explicit parameters is loaded and later
3108 encoded, the output is still encoded with explicit parameters, even if
3109 internally a "named" EC_GROUP is used for computation.
3113 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3114 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3115 NULL. After this change, only the cofactor parameter can be NULL. It also
3116 does some minimal sanity checks on the passed order.
3121 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3122 An attack is simple, if the first CMS_recipientInfo is valid but the
3123 second CMS_recipientInfo is chosen ciphertext. If the second
3124 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3125 encryption key will be replaced by garbage, and the message cannot be
3126 decoded, but if the RSA decryption fails, the correct encryption key is
3127 used and the recipient will not notice the attack.
3128 As a work around for this potential attack the length of the decrypted
3129 key must be equal to the cipher default key length, in case the
3130 certificate is not given and all recipientInfo are tried out.
3131 The old behaviour can be re-enabled in the CMS code by setting the
3132 CMS_DEBUG_DECRYPT flag.
3137 * Early start up entropy quality from the DEVRANDOM seed source has been
3138 improved for older Linux systems. The RAND subsystem will wait for
3139 /dev/random to be producing output before seeding from /dev/urandom.
3140 The seeded state is stored for future library initialisations using
3141 a system global shared memory segment. The shared memory identifier
3142 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
3143 the desired value. The default identifier is 114.
3147 * Correct the extended master secret constant on EBCDIC systems. Without this
3148 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
3149 negotiate EMS will fail. Unfortunately this also means that TLS connections
3150 between EBCDIC systems with this fix, and EBCDIC systems without this
3151 fix will fail if they negotiate EMS.
3155 * Use Windows installation paths in the mingw builds
3157 Mingw isn't a POSIX environment per se, which means that Windows
3158 paths should be used for installation.
3163 * Changed DH_check to accept parameters with order q and 2q subgroups.
3164 With order 2q subgroups the bit 0 of the private key is not secret
3165 but DH_generate_key works around that by clearing bit 0 of the
3166 private key for those. This avoids leaking bit 0 of the private key.
3170 * Significantly reduce secure memory usage by the randomness pools.
3174 * Revert the DEVRANDOM_WAIT feature for Linux systems
3176 The DEVRANDOM_WAIT feature added a select() call to wait for the
3177 /dev/random device to become readable before reading from the
3178 /dev/urandom device.
3180 It turned out that this change had negative side effects on
3181 performance which were not acceptable. After some discussion it
3182 was decided to revert this feature and leave it up to the OS
3183 resp. the platform maintainer to ensure a proper initialization
3184 during early boot time.
3186 *Matthias St. Pierre*
3188 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
3190 * Add build tests for C++. These are generated files that only do one
3191 thing, to include one public OpenSSL head file each. This tests that
3192 the public header files can be usefully included in a C++ application.
3194 This test isn't enabled by default. It can be enabled with the option
3195 'enable-buildtest-c++'.
3199 * Enable SHA3 pre-hashing for ECDSA and DSA.
3203 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3204 This changes the size when using the `genpkey` command when no size is given.
3205 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3206 generation commands to use 2048 bits by default.
3210 * Reorganize the manual pages to consistently have RETURN VALUES,
3211 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
3212 util/fix-doc-nits accordingly.
3214 *Paul Yang, Joshua Lock*
3216 * Add the missing accessor EVP_PKEY_get0_engine()
3220 * Have commands like `s_client` and `s_server` output the signature scheme
3221 along with other cipher suite parameters when debugging.
3225 * Make OPENSSL_config() error agnostic again.
3229 * Do the error handling in RSA decryption constant time.
3233 * Prevent over long nonces in ChaCha20-Poly1305.
3235 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3236 for every encryption operation. RFC 7539 specifies that the nonce value
3237 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3238 and front pads the nonce with 0 bytes if it is less than 12
3239 bytes. However it also incorrectly allows a nonce to be set of up to 16
3240 bytes. In this case only the last 12 bytes are significant and any
3241 additional leading bytes are ignored.
3243 It is a requirement of using this cipher that nonce values are
3244 unique. Messages encrypted using a reused nonce value are susceptible to
3245 serious confidentiality and integrity attacks. If an application changes
3246 the default nonce length to be longer than 12 bytes and then makes a
3247 change to the leading bytes of the nonce expecting the new value to be a
3248 new unique nonce then such an application could inadvertently encrypt
3249 messages with a reused nonce.
3251 Additionally the ignored bytes in a long nonce are not covered by the
3252 integrity guarantee of this cipher. Any application that relies on the
3253 integrity of these ignored leading bytes of a long nonce may be further
3254 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3255 is safe because no such use sets such a long nonce value. However user
3256 applications that use this cipher directly and set a non-default nonce
3257 length to be longer than 12 bytes may be vulnerable.
3259 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3265 * Add DEVRANDOM_WAIT feature for Linux systems
3267 On older Linux systems where the getrandom() system call is not available,
3268 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
3269 Contrary to getrandom(), the /dev/urandom device will not block during
3270 early boot when the kernel CSPRNG has not been seeded yet.
3272 To mitigate this known weakness, use select() to wait for /dev/random to
3273 become readable before reading from /dev/urandom.
3275 * Ensure that SM2 only uses SM3 as digest algorithm
3279 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
3281 * Change the info callback signals for the start and end of a post-handshake
3282 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
3283 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
3284 confused by this and assume that a TLSv1.2 renegotiation has started. This
3285 can break KeyUpdate handling. Instead we no longer signal the start and end
3286 of a post handshake message exchange (although the messages themselves are
3287 still signalled). This could break some applications that were expecting
3288 the old signals. However without this KeyUpdate is not usable for many
3293 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
3295 * Timing vulnerability in DSA signature generation
3297 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3298 timing side channel attack. An attacker could use variations in the signing
3299 algorithm to recover the private key.
3301 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3306 * Timing vulnerability in ECDSA signature generation
3308 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3309 timing side channel attack. An attacker could use variations in the signing
3310 algorithm to recover the private key.
3312 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3317 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
3318 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
3319 of two gigabytes and the error handling improved.
3321 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
3322 categorized as a normal bug, not a security issue, because the DRBG reseeds
3323 automatically and is fully functional even without additional randomness
3324 provided by the application.
3326 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
3328 * Add a new ClientHello callback. Provides a callback interface that gives
3329 the application the ability to adjust the nascent SSL object at the
3330 earliest stage of ClientHello processing, immediately after extensions have
3331 been collected but before they have been processed. In particular, this
3332 callback can adjust the supported TLS versions in response to the contents
3337 * Add SM2 base algorithm support.
3341 * s390x assembly pack: add (improved) hardware-support for the following
3342 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
3343 aes-cfb/cfb8, aes-ecb.
3347 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3348 parameter is no longer accepted, as it leads to a corrupt table. NULL
3349 pem_str is reserved for alias entries only.
3353 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3354 step for prime curves. The new implementation is based on formulae from
3355 differential addition-and-doubling in homogeneous projective coordinates
3356 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
3357 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
3358 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
3359 to work in projective coordinates.
3361 *Billy Bob Brumley, Nicola Tuveri*
3363 * Change generating and checking of primes so that the error rate of not
3364 being prime depends on the intended use based on the size of the input.
3365 For larger primes this will result in more rounds of Miller-Rabin.
3366 The maximal error rate for primes with more than 1080 bits is lowered
3369 *Kurt Roeckx, Annie Yousar*
3371 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3375 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
3376 moving between systems, and to avoid confusion when a Windows build is
3377 done with mingw vs with MSVC. For POSIX installs, there's still a
3378 symlink or copy named 'tsget' to avoid that confusion as well.
3382 * Revert blinding in ECDSA sign and instead make problematic addition
3383 length-invariant. Switch even to fixed-length Montgomery multiplication.
3387 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3388 step for binary curves. The new implementation is based on formulae from
3389 differential addition-and-doubling in mixed Lopez-Dahab projective
3390 coordinates, modified to independently blind the operands.
3392 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3394 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3395 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3396 EC_METHODs to implement their own specialized "ladder step", to take
3397 advantage of more favorable coordinate systems or more efficient
3398 differential addition-and-doubling algorithms.
3400 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3402 * Modified the random device based seed sources to keep the relevant
3403 file descriptors open rather than reopening them on each access.
3404 This allows such sources to operate in a chroot() jail without
3405 the associated device nodes being available. This behaviour can be
3406 controlled using RAND_keep_random_devices_open().
3410 * Numerous side-channel attack mitigations have been applied. This may have
3411 performance impacts for some algorithms for the benefit of improved
3412 security. Specific changes are noted in this change log by their respective
3417 * AIX shared library support overhaul. Switch to AIX "natural" way of
3418 handling shared libraries, which means collecting shared objects of
3419 different versions and bitnesses in one common archive. This allows to
3420 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3421 doesn't affect the way 3rd party applications are linked, only how
3422 multi-version installation is managed.
3426 * Make ec_group_do_inverse_ord() more robust and available to other
3427 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3428 mitigations are applied to the fallback BN_mod_inverse().
3429 When using this function rather than BN_mod_inverse() directly, new
3430 EC cryptosystem implementations are then safer-by-default.
3434 * Add coordinate blinding for EC_POINT and implement projective
3435 coordinate blinding for generic prime curves as a countermeasure to
3436 chosen point SCA attacks.
3438 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3440 * Add blinding to ECDSA and DSA signatures to protect against side channel
3441 attacks discovered by Keegan Ryan (NCC Group).
3445 * Enforce checking in the `pkeyutl` command to ensure that the input
3446 length does not exceed the maximum supported digest length when performing
3447 a sign, verify or verifyrecover operation.
3451 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3452 I/O in combination with something like select() or poll() will hang. This
3453 can be turned off again using SSL_CTX_clear_mode().
3454 Many applications do not properly handle non-application data records, and
3455 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3456 around the problems in those applications, but can also break some.
3457 It's recommended to read the manpages about SSL_read(), SSL_write(),
3458 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3459 SSL_CTX_set_read_ahead() again.
3463 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3464 now allow empty (zero character) pass phrases.
3468 * Apply blinding to binary field modular inversion and remove patent
3469 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3473 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3474 binary and prime elliptic curves.
3478 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3479 constant time fixed point multiplication.
3483 * Revise elliptic curve scalar multiplication with timing attack
3484 defenses: ec_wNAF_mul redirects to a constant time implementation
3485 when computing fixed point and variable point multiplication (which
3486 in OpenSSL are mostly used with secret scalars in keygen, sign,
3487 ECDH derive operations).
3488 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3491 * Updated CONTRIBUTING
3495 * Updated DRBG / RAND to request nonce and additional low entropy
3496 randomness from the system.
3498 *Matthias St. Pierre*
3500 * Updated 'openssl rehash' to use OpenSSL consistent default.
3504 * Moved the load of the ssl_conf module to libcrypto, which helps
3505 loading engines that libssl uses before libssl is initialised.
3509 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3513 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3515 *Ingo Schwarze, Rich Salz*
3517 * Added output of accepting IP address and port for 'openssl s_server'
3521 * Added a new API for TLSv1.3 ciphersuites:
3522 SSL_CTX_set_ciphersuites()
3523 SSL_set_ciphersuites()
3527 * Memory allocation failures consistently add an error to the error
3532 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3533 in libcrypto when run as setuid/setgid.
3537 * Load any config file by default when libssl is used.
3541 * Added new public header file <openssl/rand_drbg.h> and documentation
3542 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3544 *Matthias St. Pierre*
3546 * QNX support removed (cannot find contributors to get their approval
3547 for the license change).
3551 * TLSv1.3 replay protection for early data has been implemented. See the
3552 SSL_read_early_data() man page for further details.
3556 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3557 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3558 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3559 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3560 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3561 configuration has been separated out. See the ciphers man page or the
3562 SSL_CTX_set_ciphersuites() man page for more information.
3566 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3567 in responder mode now supports the new "-multi" option, which
3568 spawns the specified number of child processes to handle OCSP
3569 requests. The "-timeout" option now also limits the OCSP
3570 responder's patience to wait to receive the full client request
3571 on a newly accepted connection. Child processes are respawned
3572 as needed, and the CA index file is automatically reloaded
3573 when changed. This makes it possible to run the "ocsp" responder
3574 as a long-running service, making the OpenSSL CA somewhat more
3575 feature-complete. In this mode, most diagnostic messages logged
3576 after entering the event loop are logged via syslog(3) rather than
3581 * Added support for X448 and Ed448. Heavily based on original work by
3586 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3587 objects loaded. This adds the functions OSSL_STORE_expect() and
3588 OSSL_STORE_find() as well as needed tools to construct searches and
3589 get the search data out of them.
3593 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3594 version of OpenSSL should review their configuration settings to ensure
3595 that they are still appropriate for TLSv1.3. For further information see:
3596 <https://wiki.openssl.org/index.php/TLS1.3>
3600 * Grand redesign of the OpenSSL random generator
3602 The default RAND method now utilizes an AES-CTR DRBG according to
3603 NIST standard SP 800-90Ar1. The new random generator is essentially
3604 a port of the default random generator from the OpenSSL FIPS 2.0
3605 object module. It is a hybrid deterministic random bit generator
3606 using an AES-CTR bit stream and which seeds and reseeds itself
3607 automatically using trusted system entropy sources.
3609 Some of its new features are:
3610 - Support for multiple DRBG instances with seed chaining.
3611 - The default RAND method makes use of a DRBG.
3612 - There is a public and private DRBG instance.
3613 - The DRBG instances are fork-safe.
3614 - Keep all global DRBG instances on the secure heap if it is enabled.
3615 - The public and private DRBG instance are per thread for lock free
3618 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3620 * Changed Configure so it only says what it does and doesn't dump
3621 so much data. Instead, ./configdata.pm should be used as a script
3622 to display all sorts of configuration data.
3626 * Added processing of "make variables" to Configure.
3630 * Added SHA512/224 and SHA512/256 algorithm support.
3634 * The last traces of Netware support, first removed in 1.1.0, have
3639 * Get rid of Makefile.shared, and in the process, make the processing
3640 of certain files (rc.obj, or the .def/.map/.opt files produced from
3641 the ordinal files) more visible and hopefully easier to trace and
3642 debug (or make silent).
3646 * Make it possible to have environment variable assignments as
3647 arguments to config / Configure.
3651 * Add multi-prime RSA (RFC 8017) support.
3655 * Add SM3 implemented according to GB/T 32905-2016
3656 *Jack Lloyd <jack.lloyd@ribose.com>,*
3657 *Ronald Tse <ronald.tse@ribose.com>,*
3658 *Erick Borsboom <erick.borsboom@ribose.com>*
3660 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3661 as documented in RFC6066.
3662 Based on a patch from Tomasz Moń
3664 *Filipe Raimundo da Silva*
3666 * Add SM4 implemented according to GB/T 32907-2016.
3667 *Jack Lloyd <jack.lloyd@ribose.com>,*
3668 *Ronald Tse <ronald.tse@ribose.com>,*
3669 *Erick Borsboom <erick.borsboom@ribose.com>*
3671 * Reimplement -newreq-nodes and ERR_error_string_n; the
3672 original author does not agree with the license change.
3676 * Add ARIA AEAD TLS support.
3680 * Some macro definitions to support VS6 have been removed. Visual
3681 Studio 6 has not worked since 1.1.0
3685 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3686 without clearing the errors.
3690 * Add "atfork" functions. If building on a system that without
3691 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3692 requirements. The RAND facility now uses/requires this.
3700 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3701 not possible to disable entirely. However, it's still possible to
3702 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3705 To disable, configure with 'no-ui-console'. 'no-ui' is still
3706 possible to use as an alias. Check at compile time with the
3707 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3708 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3712 * Add a STORE module, which implements a uniform and URI based reader of
3713 stores that can contain keys, certificates, CRLs and numerous other
3714 objects. The main API is loosely based on a few stdio functions,
3715 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3716 OSSL_STORE_error and OSSL_STORE_close.
3717 The implementation uses backends called "loaders" to implement arbitrary
3718 URI schemes. There is one built in "loader" for the 'file' scheme.
3722 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3723 then adjusted to work on FreeBSD 8.4 as well.
3724 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3725 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3729 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3730 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3731 error code calls like this:
3733 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3735 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3736 that can be encoded in C. For the foreseeable future, this will only
3739 *Richard Levitte and Tim Hudson*
3741 * Removed BSD cryptodev engine.
3745 * Add a build target 'build_all_generated', to build all generated files
3746 and only that. This can be used to prepare everything that requires
3747 things like perl for a system that lacks perl and then move everything
3748 to that system and do the rest of the build there.
3752 * In the UI interface, make it possible to duplicate the user data. This
3753 can be used by engines that need to retain the data for a longer time
3754 than just the call where this user data is passed.
3758 * Ignore the '-named_curve auto' value for compatibility of applications
3761 *Tomáš Mráz <tmraz@fedoraproject.org>*
3763 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3764 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3765 alerts across multiple records (some of which could be empty). In practice
3766 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3767 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3768 support this at all. Supporting it adds significant complexity to the
3769 record layer, and its removal is unlikely to cause interoperability
3774 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3775 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3776 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3781 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3782 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3784 *Richard Levitte, Andy Polyakov*
3786 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3791 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3792 platform rather than 'mingw'.
3796 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3797 success if they are asked to add an object which already exists
3798 in the store. This change cascades to other functions which load
3799 certificates and CRLs.
3803 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3804 facilitate stack unwinding even from assembly subroutines.
3808 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3809 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3813 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3814 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3815 which is the minimum version we support.
3819 * Certificate time validation (X509_cmp_time) enforces stricter
3820 compliance with RFC 5280. Fractional seconds and timezone offsets
3821 are no longer allowed.
3825 * Add support for ARIA
3829 * s_client will now send the Server Name Indication (SNI) extension by
3830 default unless the new "-noservername" option is used. The server name is
3831 based on the host provided to the "-connect" option unless overridden by
3832 using "-servername".
3836 * Add support for SipHash
3840 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3841 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3842 prevent issues where no progress is being made and the peer continually
3843 sends unrecognised record types, using up resources processing them.
3847 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3848 using the algorithm defined in
3849 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3853 * Heartbeat support has been removed; the ABI is changed for now.
3855 *Richard Levitte, Rich Salz*
3857 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3861 * The RSA "null" method, which was partially supported to avoid patent
3862 issues, has been replaced to always returns NULL.
3869 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3871 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3872 used even when parsing explicit parameters, when loading a encoded key
3873 or calling `EC_GROUP_new_from_ecpkparameters()`/
3874 `EC_GROUP_new_from_ecparameters()`.
3875 This prevents bypass of security hardening and performance gains,
3876 especially for curves with specialized EC_METHODs.
3877 By default, if a key encoded with explicit parameters is loaded and later
3878 encoded, the output is still encoded with explicit parameters, even if
3879 internally a "named" EC_GROUP is used for computation.
3883 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3884 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3885 NULL. After this change, only the cofactor parameter can be NULL. It also
3886 does some minimal sanity checks on the passed order.
3891 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3892 An attack is simple, if the first CMS_recipientInfo is valid but the
3893 second CMS_recipientInfo is chosen ciphertext. If the second
3894 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3895 encryption key will be replaced by garbage, and the message cannot be
3896 decoded, but if the RSA decryption fails, the correct encryption key is
3897 used and the recipient will not notice the attack.
3898 As a work around for this potential attack the length of the decrypted
3899 key must be equal to the cipher default key length, in case the
3900 certificate is not given and all recipientInfo are tried out.
3901 The old behaviour can be re-enabled in the CMS code by setting the
3902 CMS_DEBUG_DECRYPT flag.
3907 * Use Windows installation paths in the mingw builds
3909 Mingw isn't a POSIX environment per se, which means that Windows
3910 paths should be used for installation.
3915 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3917 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3918 This changes the size when using the `genpkey` command when no size is given.
3919 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3920 generation commands to use 2048 bits by default.
3924 * Prevent over long nonces in ChaCha20-Poly1305.
3926 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3927 for every encryption operation. RFC 7539 specifies that the nonce value
3928 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3929 and front pads the nonce with 0 bytes if it is less than 12
3930 bytes. However it also incorrectly allows a nonce to be set of up to 16
3931 bytes. In this case only the last 12 bytes are significant and any
3932 additional leading bytes are ignored.
3934 It is a requirement of using this cipher that nonce values are
3935 unique. Messages encrypted using a reused nonce value are susceptible to
3936 serious confidentiality and integrity attacks. If an application changes
3937 the default nonce length to be longer than 12 bytes and then makes a
3938 change to the leading bytes of the nonce expecting the new value to be a
3939 new unique nonce then such an application could inadvertently encrypt
3940 messages with a reused nonce.
3942 Additionally the ignored bytes in a long nonce are not covered by the
3943 integrity guarantee of this cipher. Any application that relies on the
3944 integrity of these ignored leading bytes of a long nonce may be further
3945 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3946 is safe because no such use sets such a long nonce value. However user
3947 applications that use this cipher directly and set a non-default nonce
3948 length to be longer than 12 bytes may be vulnerable.
3950 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3956 * Added SCA hardening for modular field inversion in EC_GROUP through
3957 a new dedicated field_inv() pointer in EC_METHOD.
3958 This also addresses a leakage affecting conversions from projective
3959 to affine coordinates.
3961 *Billy Bob Brumley, Nicola Tuveri*
3963 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3964 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3968 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3972 * Remove the 'dist' target and add a tarball building script. The
3973 'dist' target has fallen out of use, and it shouldn't be
3974 necessary to configure just to create a source distribution.
3978 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3980 * Timing vulnerability in DSA signature generation
3982 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3983 timing side channel attack. An attacker could use variations in the signing
3984 algorithm to recover the private key.
3986 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3991 * Timing vulnerability in ECDSA signature generation
3993 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3994 timing side channel attack. An attacker could use variations in the signing
3995 algorithm to recover the private key.
3997 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
4002 * Add coordinate blinding for EC_POINT and implement projective
4003 coordinate blinding for generic prime curves as a countermeasure to
4004 chosen point SCA attacks.
4006 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
4008 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
4010 * Client DoS due to large DH parameter
4012 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4013 malicious server can send a very large prime value to the client. This will
4014 cause the client to spend an unreasonably long period of time generating a
4015 key for this prime resulting in a hang until the client has finished. This
4016 could be exploited in a Denial Of Service attack.
4018 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4023 * Cache timing vulnerability in RSA Key Generation
4025 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4026 a cache timing side channel attack. An attacker with sufficient access to
4027 mount cache timing attacks during the RSA key generation process could
4028 recover the private key.
4030 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4031 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4036 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4037 parameter is no longer accepted, as it leads to a corrupt table. NULL
4038 pem_str is reserved for alias entries only.
4042 * Revert blinding in ECDSA sign and instead make problematic addition
4043 length-invariant. Switch even to fixed-length Montgomery multiplication.
4047 * Change generating and checking of primes so that the error rate of not
4048 being prime depends on the intended use based on the size of the input.
4049 For larger primes this will result in more rounds of Miller-Rabin.
4050 The maximal error rate for primes with more than 1080 bits is lowered
4053 *Kurt Roeckx, Annie Yousar*
4055 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4059 * Add blinding to ECDSA and DSA signatures to protect against side channel
4060 attacks discovered by Keegan Ryan (NCC Group).
4064 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4065 now allow empty (zero character) pass phrases.
4069 * Certificate time validation (X509_cmp_time) enforces stricter
4070 compliance with RFC 5280. Fractional seconds and timezone offsets
4071 are no longer allowed.
4075 * Fixed a text canonicalisation bug in CMS
4077 Where a CMS detached signature is used with text content the text goes
4078 through a canonicalisation process first prior to signing or verifying a
4079 signature. This process strips trailing space at the end of lines, converts
4080 line terminators to CRLF and removes additional trailing line terminators
4081 at the end of a file. A bug in the canonicalisation process meant that
4082 some characters, such as form-feed, were incorrectly treated as whitespace
4083 and removed. This is contrary to the specification (RFC5485). This fix
4084 could mean that detached text data signed with an earlier version of
4085 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
4086 signed with a fixed OpenSSL may fail to verify with an earlier version of
4087 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
4088 and use the "-binary" flag (for the "cms" command line application) or set
4089 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
4093 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
4095 * Constructed ASN.1 types with a recursive definition could exceed the stack
4097 Constructed ASN.1 types with a recursive definition (such as can be found
4098 in PKCS7) could eventually exceed the stack given malicious input with
4099 excessive recursion. This could result in a Denial Of Service attack. There
4100 are no such structures used within SSL/TLS that come from untrusted sources
4101 so this is considered safe.
4103 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4109 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
4111 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
4112 effectively reduced to only comparing the least significant bit of each
4113 byte. This allows an attacker to forge messages that would be considered as
4114 authenticated in an amount of tries lower than that guaranteed by the
4115 security claims of the scheme. The module can only be compiled by the
4116 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
4118 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
4124 * Add a build target 'build_all_generated', to build all generated files
4125 and only that. This can be used to prepare everything that requires
4126 things like perl for a system that lacks perl and then move everything
4127 to that system and do the rest of the build there.
4131 * Backport SSL_OP_NO_RENGOTIATION
4133 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
4134 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
4135 changes this is no longer possible in 1.1.0. Therefore, the new
4136 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
4137 1.1.0 to provide equivalent functionality.
4139 Note that if an application built against 1.1.0h headers (or above) is run
4140 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
4141 accepted but nothing will happen, i.e. renegotiation will not be prevented.
4145 * Removed the OS390-Unix config target. It relied on a script that doesn't
4150 * rsaz_1024_mul_avx2 overflow bug on x86_64
4152 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4153 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4154 Analysis suggests that attacks against RSA and DSA as a result of this
4155 defect would be very difficult to perform and are not believed likely.
4156 Attacks against DH1024 are considered just feasible, because most of the
4157 work necessary to deduce information about a private key may be performed
4158 offline. The amount of resources required for such an attack would be
4159 significant. However, for an attack on TLS to be meaningful, the server
4160 would have to share the DH1024 private key among multiple clients, which is
4161 no longer an option since CVE-2016-0701.
4163 This only affects processors that support the AVX2 but not ADX extensions
4164 like Intel Haswell (4th generation).
4166 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4167 was originally found via the OSS-Fuzz project.
4172 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
4174 * bn_sqrx8x_internal carry bug on x86_64
4176 There is a carry propagating bug in the x86_64 Montgomery squaring
4177 procedure. No EC algorithms are affected. Analysis suggests that attacks
4178 against RSA and DSA as a result of this defect would be very difficult to
4179 perform and are not believed likely. Attacks against DH are considered just
4180 feasible (although very difficult) because most of the work necessary to
4181 deduce information about a private key may be performed offline. The amount
4182 of resources required for such an attack would be very significant and
4183 likely only accessible to a limited number of attackers. An attacker would
4184 additionally need online access to an unpatched system using the target
4185 private key in a scenario with persistent DH parameters and a private
4186 key that is shared between multiple clients.
4188 This only affects processors that support the BMI1, BMI2 and ADX extensions
4189 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4191 This issue was reported to OpenSSL by the OSS-Fuzz project.
4196 * Malformed X.509 IPAddressFamily could cause OOB read
4198 If an X.509 certificate has a malformed IPAddressFamily extension,
4199 OpenSSL could do a one-byte buffer overread. The most likely result
4200 would be an erroneous display of the certificate in text format.
4202 This issue was reported to OpenSSL by the OSS-Fuzz project.
4207 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
4209 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4210 platform rather than 'mingw'.
4214 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
4215 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
4216 which is the minimum version we support.
4220 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
4222 * Encrypt-Then-Mac renegotiation crash
4224 During a renegotiation handshake if the Encrypt-Then-Mac extension is
4225 negotiated where it was not in the original handshake (or vice-versa) then
4226 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
4227 and servers are affected.
4229 This issue was reported to OpenSSL by Joe Orton (Red Hat).
4234 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
4236 * Truncated packet could crash via OOB read
4238 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4239 cipher is being used, then a truncated packet can cause that host to
4240 perform an out-of-bounds read, usually resulting in a crash.
4242 This issue was reported to OpenSSL by Robert Święcki of Google.
4247 * Bad (EC)DHE parameters cause a client crash
4249 If a malicious server supplies bad parameters for a DHE or ECDHE key
4250 exchange then this can result in the client attempting to dereference a
4251 NULL pointer leading to a client crash. This could be exploited in a Denial
4254 This issue was reported to OpenSSL by Guido Vranken.
4259 * BN_mod_exp may produce incorrect results on x86_64
4261 There is a carry propagating bug in the x86_64 Montgomery squaring
4262 procedure. No EC algorithms are affected. Analysis suggests that attacks
4263 against RSA and DSA as a result of this defect would be very difficult to
4264 perform and are not believed likely. Attacks against DH are considered just
4265 feasible (although very difficult) because most of the work necessary to
4266 deduce information about a private key may be performed offline. The amount
4267 of resources required for such an attack would be very significant and
4268 likely only accessible to a limited number of attackers. An attacker would
4269 additionally need online access to an unpatched system using the target
4270 private key in a scenario with persistent DH parameters and a private
4271 key that is shared between multiple clients. For example this can occur by
4272 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4273 similar to CVE-2015-3193 but must be treated as a separate problem.
4275 This issue was reported to OpenSSL by the OSS-Fuzz project.
4280 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
4282 * ChaCha20/Poly1305 heap-buffer-overflow
4284 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
4285 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
4286 crash. This issue is not considered to be exploitable beyond a DoS.
4288 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
4293 * CMS Null dereference
4295 Applications parsing invalid CMS structures can crash with a NULL pointer
4296 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
4297 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
4298 structure callback if an attempt is made to free certain invalid encodings.
4299 Only CHOICE structures using a callback which do not handle NULL value are
4302 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
4307 * Montgomery multiplication may produce incorrect results
4309 There is a carry propagating bug in the Broadwell-specific Montgomery
4310 multiplication procedure that handles input lengths divisible by, but
4311 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4312 and DH private keys are impossible. This is because the subroutine in
4313 question is not used in operations with the private key itself and an input
4314 of the attacker's direct choice. Otherwise the bug can manifest itself as
4315 transient authentication and key negotiation failures or reproducible
4316 erroneous outcome of public-key operations with specially crafted input.
4317 Among EC algorithms only Brainpool P-512 curves are affected and one
4318 presumably can attack ECDH key negotiation. Impact was not analyzed in
4319 detail, because pre-requisites for attack are considered unlikely. Namely
4320 multiple clients have to choose the curve in question and the server has to
4321 share the private key among them, neither of which is default behaviour.
4322 Even then only clients that chose the curve will be affected.
4324 This issue was publicly reported as transient failures and was not
4325 initially recognized as a security issue. Thanks to Richard Morgan for
4326 providing reproducible case.
4331 * Removed automatic addition of RPATH in shared libraries and executables,
4332 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
4336 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
4338 * Fix Use After Free for large message sizes
4340 The patch applied to address CVE-2016-6307 resulted in an issue where if a
4341 message larger than approx 16k is received then the underlying buffer to
4342 store the incoming message is reallocated and moved. Unfortunately a
4343 dangling pointer to the old location is left which results in an attempt to
4344 write to the previously freed location. This is likely to result in a
4345 crash, however it could potentially lead to execution of arbitrary code.
4347 This issue only affects OpenSSL 1.1.0a.
4349 This issue was reported to OpenSSL by Robert Święcki.
4354 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
4356 * OCSP Status Request extension unbounded memory growth
4358 A malicious client can send an excessively large OCSP Status Request
4359 extension. If that client continually requests renegotiation, sending a
4360 large OCSP Status Request extension each time, then there will be unbounded
4361 memory growth on the server. This will eventually lead to a Denial Of
4362 Service attack through memory exhaustion. Servers with a default
4363 configuration are vulnerable even if they do not support OCSP. Builds using
4364 the "no-ocsp" build time option are not affected.
4366 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4371 * SSL_peek() hang on empty record
4373 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
4374 sends an empty record. This could be exploited by a malicious peer in a
4375 Denial Of Service attack.
4377 This issue was reported to OpenSSL by Alex Gaynor.
4382 * Excessive allocation of memory in tls_get_message_header() and
4383 dtls1_preprocess_fragment()
4385 A (D)TLS message includes 3 bytes for its length in the header for the
4386 message. This would allow for messages up to 16Mb in length. Messages of
4387 this length are excessive and OpenSSL includes a check to ensure that a
4388 peer is sending reasonably sized messages in order to avoid too much memory
4389 being consumed to service a connection. A flaw in the logic of version
4390 1.1.0 means that memory for the message is allocated too early, prior to
4391 the excessive message length check. Due to way memory is allocated in
4392 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4393 to service a connection. This could lead to a Denial of Service through
4394 memory exhaustion. However, the excessive message length check still takes
4395 place, and this would cause the connection to immediately fail. Assuming
4396 that the application calls SSL_free() on the failed connection in a timely
4397 manner then the 21Mb of allocated memory will then be immediately freed
4398 again. Therefore, the excessive memory allocation will be transitory in
4399 nature. This then means that there is only a security impact if:
4401 1) The application does not call SSL_free() in a timely manner in the event
4402 that the connection fails
4404 2) The application is working in a constrained environment where there is
4405 very little free memory
4407 3) The attacker initiates multiple connection attempts such that there are
4408 multiple connections in a state where memory has been allocated for the
4409 connection; SSL_free() has not yet been called; and there is insufficient
4410 memory to service the multiple requests.
4412 Except in the instance of (1) above any Denial Of Service is likely to be
4413 transitory because as soon as the connection fails the memory is
4414 subsequently freed again in the SSL_free() call. However there is an
4415 increased risk during this period of application crashes due to the lack of
4416 memory - which would then mean a more serious Denial of Service.
4418 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4419 (CVE-2016-6307 and CVE-2016-6308)
4423 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4424 had to be removed. Primary reason is that vendor assembler can't
4425 assemble our modules with -KPIC flag. As result it, assembly
4426 support, was not even available as option. But its lack means
4427 lack of side-channel resistant code, which is incompatible with
4428 security by todays standards. Fortunately gcc is readily available
4429 prepackaged option, which we firmly point at...
4433 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4435 * Windows command-line tool supports UTF-8 opt-in option for arguments
4436 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4437 (to any value) allows Windows user to access PKCS#12 file generated
4438 with Windows CryptoAPI and protected with non-ASCII password, as well
4439 as files generated under UTF-8 locale on Linux also protected with
4444 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4445 have been disabled by default and removed from DEFAULT, just like RC4.
4446 See the RC4 item below to re-enable both.
4450 * The method for finding the storage location for the Windows RAND seed file
4451 has changed. First we check %RANDFILE%. If that is not set then we check
4452 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4453 all else fails we fall back to C:\.
4457 * The EVP_EncryptUpdate() function has had its return type changed from void
4458 to int. A return of 0 indicates and error while a return of 1 indicates
4463 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4464 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4465 off the constant time implementation for RSA, DSA and DH have been made
4466 no-ops and deprecated.
4470 * Windows RAND implementation was simplified to only get entropy by
4471 calling CryptGenRandom(). Various other RAND-related tickets
4474 *Joseph Wylie Yandle, Rich Salz*
4476 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4477 and `OPENSSL_LH_`, respectively. The old names are available
4478 with API compatibility. They new names are now completely documented.
4482 * Unify TYPE_up_ref(obj) methods signature.
4483 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4484 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4485 int (instead of void) like all others TYPE_up_ref() methods.
4486 So now these methods also check the return value of CRYPTO_atomic_add(),
4487 and the validity of object reference counter.
4489 *fdasilvayy@gmail.com*
4491 * With Windows Visual Studio builds, the .pdb files are installed
4492 alongside the installed libraries and executables. For a static
4493 library installation, ossl_static.pdb is the associate compiler
4494 generated .pdb file to be used when linking programs.
4498 * Remove openssl.spec. Packaging files belong with the packagers.
4502 * Automatic Darwin/OSX configuration has had a refresh, it will now
4503 recognise x86_64 architectures automatically. You can still decide
4504 to build for a different bitness with the environment variable
4505 KERNEL_BITS (can be 32 or 64), for example:
4507 KERNEL_BITS=32 ./config
4511 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4512 256 bit AES and HMAC with SHA256.
4516 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4520 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4524 * To enable users to have their own config files and build file templates,
4525 Configure looks in the directory indicated by the environment variable
4526 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4527 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4528 name and is used as is.
4532 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4533 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4534 X509_CERT_FILE_CTX was removed.
4538 * "shared" builds are now the default. To create only static libraries use
4539 the "no-shared" Configure option.
4543 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4544 All of these option have not worked for some while and are fundamental
4549 * Make various cleanup routines no-ops and mark them as deprecated. Most
4550 global cleanup functions are no longer required because they are handled
4551 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4552 Explicitly de-initing can cause problems (e.g. where a library that uses
4553 OpenSSL de-inits, but an application is still using it). The affected
4554 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4555 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4556 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4557 COMP_zlib_cleanup().
4561 * --strict-warnings no longer enables runtime debugging options
4562 such as REF_DEBUG. Instead, debug options are automatically
4563 enabled with '--debug' builds.
4565 *Andy Polyakov, Emilia Käsper*
4567 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4568 have been moved out of the public header files. New functions for managing
4569 these have been added.
4573 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4574 objects have been moved out of the public header files. New
4575 functions for managing these have been added.
4579 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4580 have been moved out of the public header files. New functions for managing
4581 these have been added.
4585 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4586 moved out of the public header files. New functions for managing these
4591 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4595 * Removed the mk1mf build scripts.
4599 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4600 it is always safe to #include a header now.
4604 * Removed the aged BC-32 config and all its supporting scripts
4608 * Removed support for Ultrix, Netware, and OS/2.
4612 * Add support for HKDF.
4614 *Alessandro Ghedini*
4616 * Add support for blake2b and blake2s
4620 * Added support for "pipelining". Ciphers that have the
4621 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4622 encryptions/decryptions simultaneously. There are currently no built-in
4623 ciphers with this property but the expectation is that engines will be able
4624 to offer it to significantly improve throughput. Support has been extended
4625 into libssl so that multiple records for a single connection can be
4626 processed in one go (for >=TLS 1.1).
4630 * Added the AFALG engine. This is an async capable engine which is able to
4631 offload work to the Linux kernel. In this initial version it only supports
4632 AES128-CBC. The kernel must be version 4.1.0 or greater.
4636 * OpenSSL now uses a new threading API. It is no longer necessary to
4637 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4638 are two supported threading models: pthreads and windows threads. It is
4639 also possible to configure OpenSSL at compile time for "no-threads". The
4640 old threading API should no longer be used. The functions have been
4641 replaced with "no-op" compatibility macros.
4643 *Alessandro Ghedini, Matt Caswell*
4645 * Modify behavior of ALPN to invoke callback after SNI/servername
4646 callback, such that updates to the SSL_CTX affect ALPN.
4650 * Add SSL_CIPHER queries for authentication and key-exchange.
4654 * Changes to the DEFAULT cipherlist:
4655 - Prefer (EC)DHE handshakes over plain RSA.
4656 - Prefer AEAD ciphers over legacy ciphers.
4657 - Prefer ECDSA over RSA when both certificates are available.
4658 - Prefer TLSv1.2 ciphers/PRF.
4659 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4664 * Change the ECC default curve list to be this, in order: x25519,
4665 secp256r1, secp521r1, secp384r1.
4669 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4670 disabled by default. They can be re-enabled using the
4671 enable-weak-ssl-ciphers option to Configure.
4675 * If the server has ALPN configured, but supports no protocols that the
4676 client advertises, send a fatal "no_application_protocol" alert.
4677 This behaviour is SHALL in RFC 7301, though it isn't universally
4678 implemented by other servers.
4682 * Add X25519 support.
4683 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4684 for public and private key encoding using the format documented in
4685 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4686 key generation and key derivation.
4688 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4693 * Deprecate SRP_VBASE_get_by_user.
4694 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4695 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4696 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4697 seed, even if the seed is configured.
4699 Users should use SRP_VBASE_get1_by_user instead. Note that in
4700 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4701 also that even though configuring the SRP seed attempts to hide
4702 invalid usernames by continuing the handshake with fake
4703 credentials, this behaviour is not constant time and no strong
4704 guarantees are made that the handshake is indistinguishable from
4705 that of a valid user.
4709 * Configuration change; it's now possible to build dynamic engines
4710 without having to build shared libraries and vice versa. This
4711 only applies to the engines in `engines/`, those in `crypto/engine/`
4712 will always be built into libcrypto (i.e. "static").
4714 Building dynamic engines is enabled by default; to disable, use
4715 the configuration option "disable-dynamic-engine".
4717 The only requirements for building dynamic engines are the
4718 presence of the DSO module and building with position independent
4719 code, so they will also automatically be disabled if configuring
4720 with "disable-dso" or "disable-pic".
4722 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4723 are also taken away from openssl/opensslconf.h, as they are
4728 * Configuration change; if there is a known flag to compile
4729 position independent code, it will always be applied on the
4730 libcrypto and libssl object files, and never on the application
4731 object files. This means other libraries that use routines from
4732 libcrypto / libssl can be made into shared libraries regardless
4733 of how OpenSSL was configured.
4735 If this isn't desirable, the configuration options "disable-pic"
4736 or "no-pic" can be used to disable the use of PIC. This will
4737 also disable building shared libraries and dynamic engines.
4741 * Removed JPAKE code. It was experimental and has no wide use.
4745 * The INSTALL_PREFIX Makefile variable has been renamed to
4746 DESTDIR. That makes for less confusion on what this variable
4747 is for. Also, the configuration option --install_prefix is
4752 * Heartbeat for TLS has been removed and is disabled by default
4753 for DTLS; configure with enable-heartbeats. Code that uses the
4754 old #define's might need to be updated.
4756 *Emilia Käsper, Rich Salz*
4758 * Rename REF_CHECK to REF_DEBUG.
4762 * New "unified" build system
4764 The "unified" build system is aimed to be a common system for all
4765 platforms we support. With it comes new support for VMS.
4767 This system builds supports building in a different directory tree
4768 than the source tree. It produces one Makefile (for unix family
4769 or lookalikes), or one descrip.mms (for VMS).
4771 The source of information to make the Makefile / descrip.mms is
4772 small files called 'build.info', holding the necessary
4773 information for each directory with source to compile, and a
4774 template in Configurations, like unix-Makefile.tmpl or
4777 With this change, the library names were also renamed on Windows
4778 and on VMS. They now have names that are closer to the standard
4779 on Unix, and include the major version number, and in certain
4780 cases, the architecture they are built for. See "Notes on shared
4781 libraries" in INSTALL.
4783 We rely heavily on the perl module Text::Template.
4787 * Added support for auto-initialisation and de-initialisation of the library.
4788 OpenSSL no longer requires explicit init or deinit routines to be called,
4789 except in certain circumstances. See the OPENSSL_init_crypto() and
4790 OPENSSL_init_ssl() man pages for further information.
4794 * The arguments to the DTLSv1_listen function have changed. Specifically the
4795 "peer" argument is now expected to be a BIO_ADDR object.
4797 * Rewrite of BIO networking library. The BIO library lacked consistent
4798 support of IPv6, and adding it required some more extensive
4799 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4800 which hold all types of addresses and chains of address information.
4801 It also introduces a new API, with functions like BIO_socket,
4802 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4803 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4804 have been adapted accordingly.
4808 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4813 * CRIME protection: disable compression by default, even if OpenSSL is
4814 compiled with zlib enabled. Applications can still enable compression
4815 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4816 using the SSL_CONF library to configure compression.
4820 * The signature of the session callback configured with
4821 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4822 was explicitly marked as `const unsigned char*` instead of
4827 * Always DPURIFY. Remove the use of uninitialized memory in the
4828 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4832 * Removed many obsolete configuration items, including
4833 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4834 MD2_CHAR, MD2_INT, MD2_LONG
4836 IDEA_SHORT, IDEA_LONG
4837 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4839 *Rich Salz, with advice from Andy Polyakov*
4841 * Many BN internals have been moved to an internal header file.
4843 *Rich Salz with help from Andy Polyakov*
4845 * Configuration and writing out the results from it has changed.
4846 Files such as Makefile include/openssl/opensslconf.h and are now
4847 produced through general templates, such as Makefile.in and
4848 crypto/opensslconf.h.in and some help from the perl module
4851 Also, the center of configuration information is no longer
4852 Makefile. Instead, Configure produces a perl module in
4853 configdata.pm which holds most of the config data (in the hash
4854 table %config), the target data that comes from the target
4855 configuration in one of the `Configurations/*.conf` files (in
4860 * To clarify their intended purposes, the Configure options
4861 --prefix and --openssldir change their semantics, and become more
4862 straightforward and less interdependent.
4864 --prefix shall be used exclusively to give the location INSTALLTOP
4865 where programs, scripts, libraries, include files and manuals are
4866 going to be installed. The default is now /usr/local.
4868 --openssldir shall be used exclusively to give the default
4869 location OPENSSLDIR where certificates, private keys, CRLs are
4870 managed. This is also where the default openssl.cnf gets
4872 If the directory given with this option is a relative path, the
4873 values of both the --prefix value and the --openssldir value will
4874 be combined to become OPENSSLDIR.
4875 The default for --openssldir is INSTALLTOP/ssl.
4877 Anyone who uses --openssldir to specify where OpenSSL is to be
4878 installed MUST change to use --prefix instead.
4882 * The GOST engine was out of date and therefore it has been removed. An up
4883 to date GOST engine is now being maintained in an external repository.
4884 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4885 support for GOST ciphersuites (these are only activated if a GOST engine
4890 * EGD is no longer supported by default; use enable-egd when
4893 *Ben Kaduk and Rich Salz*
4895 * The distribution now has Makefile.in files, which are used to
4896 create Makefile's when Configure is run. *Configure must be run
4897 before trying to build now.*
4901 * The return value for SSL_CIPHER_description() for error conditions
4906 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4908 Obtaining and performing DNSSEC validation of TLSA records is
4909 the application's responsibility. The application provides
4910 the TLSA records of its choice to OpenSSL, and these are then
4911 used to authenticate the peer.
4913 The TLSA records need not even come from DNS. They can, for
4914 example, be used to implement local end-entity certificate or
4915 trust-anchor "pinning", where the "pin" data takes the form
4916 of TLSA records, which can augment or replace verification
4917 based on the usual WebPKI public certification authorities.
4921 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4922 continues to support deprecated interfaces in default builds.
4923 However, applications are strongly advised to compile their
4924 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4925 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4926 or the 1.1.0 releases.
4928 In environments in which all applications have been ported to
4929 not use any deprecated interfaces OpenSSL's Configure script
4930 should be used with the --api=1.1.0 option to entirely remove
4931 support for the deprecated features from the library and
4932 unconditionally disable them in the installed headers.
4933 Essentially the same effect can be achieved with the "no-deprecated"
4934 argument to Configure, except that this will always restrict
4935 the build to just the latest API, rather than a fixed API
4938 As applications are ported to future revisions of the API,
4939 they should update their compile-time OPENSSL_API_COMPAT define
4940 accordingly, but in most cases should be able to continue to
4941 compile with later releases.
4943 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4944 0x10000000L and 0x00908000L, respectively. However those
4945 versions did not support the OPENSSL_API_COMPAT feature, and
4946 so applications are not typically tested for explicit support
4947 of just the undeprecated features of either release.
4951 * Add support for setting the minimum and maximum supported protocol.
4952 It can bet set via the SSL_set_min_proto_version() and
4953 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4954 MaxProtocol. It's recommended to use the new APIs to disable
4955 protocols instead of disabling individual protocols using
4956 SSL_set_options() or SSL_CONF's Protocol. This change also
4957 removes support for disabling TLS 1.2 in the OpenSSL TLS
4958 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4962 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4966 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4967 and integrates ECDSA and ECDH functionality into EC. Implementations can
4968 now redirect key generation and no longer need to convert to or from
4971 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4972 include the ec.h header file instead.
4976 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4977 ciphers who are no longer supported and drops support the ephemeral RSA key
4978 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4982 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4983 opaque. For HMAC_CTX, the following constructors and destructors
4986 HMAC_CTX *HMAC_CTX_new(void);
4987 void HMAC_CTX_free(HMAC_CTX *ctx);
4989 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4990 destroy such methods has been added. See EVP_MD_meth_new(3) and
4991 EVP_CIPHER_meth_new(3) for documentation.
4994 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4995 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4996 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4997 an already created structure.
4998 2) For consistency with the majority of our object creators and
4999 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
5000 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
5001 for deprecated builds.
5005 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
5006 cryptographic operations to be performed asynchronously as long as an
5007 asynchronous capable engine is used. See the ASYNC_start_job() man page for
5008 further details. Libssl has also had this capability integrated with the
5009 introduction of the new mode SSL_MODE_ASYNC and associated error
5010 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
5011 pages. This work was developed in partnership with Intel Corp.
5015 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
5016 always enabled now. If you want to disable the support you should
5017 exclude it using the list of supported ciphers. This also means that the
5018 "-no_ecdhe" option has been removed from s_server.
5022 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
5023 SSL_{CTX_}set1_curves() which can set a list.
5027 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
5028 curve you want to support using SSL_{CTX_}set1_curves().
5032 * State machine rewrite. The state machine code has been significantly
5033 refactored in order to remove much duplication of code and solve issues
5034 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
5035 further details). This change does have some associated API changes.
5036 Notably the SSL_state() function has been removed and replaced by
5037 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
5038 SSL_set_state() has been removed altogether. The previous handshake states
5039 defined in ssl.h and ssl3.h have also been removed.
5043 * All instances of the string "ssleay" in the public API were replaced
5044 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
5045 Some error codes related to internal RSA_eay API's were renamed.
5049 * The demo files in crypto/threads were moved to demo/threads.
5053 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
5056 *Matt Caswell, Rich Salz*
5058 * New ASN.1 embed macro.
5060 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
5061 structure is not allocated: it is part of the parent. That is instead of
5069 This reduces memory fragmentation and make it impossible to accidentally
5070 set a mandatory field to NULL.
5072 This currently only works for some fields specifically a SEQUENCE, CHOICE,
5073 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
5074 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
5079 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
5083 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
5084 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
5085 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
5086 DES and RC4 ciphersuites.
5090 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5091 This changes the decoding behaviour for some invalid messages,
5092 though the change is mostly in the more lenient direction, and
5093 legacy behaviour is preserved as much as possible.
5097 * Fix no-stdio build.
5098 *David Woodhouse <David.Woodhouse@intel.com> and also*
5099 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
5101 * New testing framework
5102 The testing framework has been largely rewritten and is now using
5103 perl and the perl modules Test::Harness and an extended variant of
5104 Test::More called OpenSSL::Test to do its work. All test scripts in
5105 test/ have been rewritten into test recipes, and all direct calls to
5106 executables in test/Makefile have become individual recipes using the
5107 simplified testing OpenSSL::Test::Simple.
5109 For documentation on our testing modules, do:
5111 perldoc test/testlib/OpenSSL/Test/Simple.pm
5112 perldoc test/testlib/OpenSSL/Test.pm
5116 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
5117 are used; the latter aborts on memory leaks (usually checked on exit).
5118 Some undocumented "set malloc, etc., hooks" functions were removed
5119 and others were changed. All are now documented.
5123 * In DSA_generate_parameters_ex, if the provided seed is too short,
5126 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5128 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
5129 from RFC4279, RFC4785, RFC5487, RFC5489.
5131 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
5132 original RSA_PSK patch.
5136 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
5137 era flag was never set throughout the codebase (only read). Also removed
5138 SSL3_FLAGS_POP_BUFFER which was only used if
5139 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
5143 * Changed the default name options in the "ca", "crl", "req" and "x509"
5144 to be "oneline" instead of "compat".
5148 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
5149 not aware of clients that still exhibit this bug, and the workaround
5150 hasn't been working properly for a while.
5154 * The return type of BIO_number_read() and BIO_number_written() as well as
5155 the corresponding num_read and num_write members in the BIO structure has
5156 changed from unsigned long to uint64_t. On platforms where an unsigned
5157 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
5162 * Given the pervasive nature of TLS extensions it is inadvisable to run
5163 OpenSSL without support for them. It also means that maintaining
5164 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
5165 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
5169 * Removed support for the two export grade static DH ciphersuites
5170 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
5171 were newly added (along with a number of other static DH ciphersuites) to
5172 1.0.2. However the two export ones have *never* worked since they were
5173 introduced. It seems strange in any case to be adding new export
5174 ciphersuites, and given "logjam" it also does not seem correct to fix them.
5178 * Version negotiation has been rewritten. In particular SSLv23_method(),
5179 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
5180 and turned into macros which simply call the new preferred function names
5181 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
5182 should use the new names instead. Also as part of this change the ssl23.h
5183 header file has been removed.
5187 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
5188 code and the associated standard is no longer considered fit-for-purpose.
5192 * RT2547 was closed. When generating a private key, try to make the
5193 output file readable only by the owner. This behavior change might
5194 be noticeable when interacting with other software.
5196 * Documented all exdata functions. Added CRYPTO_free_ex_index.
5201 * Added HTTP GET support to the ocsp command.
5205 * Changed default digest for the dgst and enc commands from MD5 to
5210 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
5214 * Added support for TLS extended master secret from
5215 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
5216 initial patch which was a great help during development.
5220 * All libssl internal structures have been removed from the public header
5221 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
5222 now redundant). Users should not attempt to access internal structures
5223 directly. Instead they should use the provided API functions.
5227 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
5228 Access to deprecated functions can be re-enabled by running config with
5229 "enable-deprecated". In addition applications wishing to use deprecated
5230 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
5231 will, by default, disable some transitive includes that previously existed
5232 in the header files (e.g. ec.h will no longer, by default, include bn.h)
5236 * Added support for OCB mode. OpenSSL has been granted a patent license
5237 compatible with the OpenSSL license for use of OCB. Details are available
5238 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
5239 for OCB can be removed by calling config with no-ocb.
5243 * SSLv2 support has been removed. It still supports receiving an SSLv2
5244 compatible client hello.
5248 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
5249 done while fixing the error code for the key-too-small case.
5251 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
5253 * CA.sh has been removed; use CA.pl instead.
5257 * Removed old DES API.
5261 * Remove various unsupported platforms:
5267 Sinix/ReliantUNIX RM400
5272 16-bit platforms such as WIN16
5276 * Clean up OPENSSL_NO_xxx #define's
5277 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
5278 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
5279 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
5280 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
5281 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
5282 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
5283 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
5284 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
5285 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
5286 - Remove MS_STATIC; it's a relic from platforms <32 bits.
5290 * Cleaned up dead code
5291 Remove all but one '#ifdef undef' which is to be looked at.
5295 * Clean up calling of xxx_free routines.
5296 Just like free(), fix most of the xxx_free routines to accept
5297 NULL. Remove the non-null checks from callers. Save much code.
5301 * Add secure heap for storage of private keys (when possible).
5302 Add BIO_s_secmem(), CBIGNUM, etc.
5303 Contributed by Akamai Technologies under our Corporate CLA.
5307 * Experimental support for a new, fast, unbiased prime candidate generator,
5308 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
5310 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
5312 * New output format NSS in the sess_id command line tool. This allows
5313 exporting the session id and the master key in NSS keylog format.
5315 *Martin Kaiser <martin@kaiser.cx>*
5317 * Harmonize version and its documentation. -f flag is used to display
5320 *mancha <mancha1@zoho.com>*
5322 * Fix eckey_priv_encode so it immediately returns an error upon a failure
5323 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
5325 *mancha <mancha1@zoho.com>*
5327 * Fix some double frees. These are not thought to be exploitable.
5329 *mancha <mancha1@zoho.com>*
5331 * A missing bounds check in the handling of the TLS heartbeat extension
5332 can be used to reveal up to 64k of memory to a connected client or
5335 Thanks for Neel Mehta of Google Security for discovering this bug and to
5336 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
5337 preparing the fix ([CVE-2014-0160])
5339 *Adam Langley, Bodo Moeller*
5341 * Fix for the attack described in the paper "Recovering OpenSSL
5342 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
5343 by Yuval Yarom and Naomi Benger. Details can be obtained from:
5344 <http://eprint.iacr.org/2014/140>
5346 Thanks to Yuval Yarom and Naomi Benger for discovering this
5347 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5349 *Yuval Yarom and Naomi Benger*
5351 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5352 this fixes a limitation in previous versions of OpenSSL.
5356 * Experimental encrypt-then-mac support.
5358 Experimental support for encrypt then mac from
5359 draft-gutmann-tls-encrypt-then-mac-02.txt
5361 To enable it set the appropriate extension number (0x42 for the test
5362 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
5364 For non-compliant peers (i.e. just about everything) this should have no
5367 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
5371 * Add EVP support for key wrapping algorithms, to avoid problems with
5372 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5373 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5374 algorithms and include tests cases.
5378 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
5383 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5384 MGF1 digest and OAEP label.
5388 * Make openssl verify return errors.
5390 *Chris Palmer <palmer@google.com> and Ben Laurie*
5392 * New function ASN1_TIME_diff to calculate the difference between two
5393 ASN1_TIME structures or one structure and the current time.
5397 * Update fips_test_suite to support multiple command line options. New
5398 test to induce all self test errors in sequence and check expected
5403 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5404 sign or verify all in one operation.
5408 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5409 test programs and fips_test_suite. Includes functionality to parse
5410 the minimal script output of fipsalgest.pl directly.
5414 * Add authorisation parameter to FIPS_module_mode_set().
5418 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5422 * Use separate DRBG fields for internal and external flags. New function
5423 FIPS_drbg_health_check() to perform on demand health checking. Add
5424 generation tests to fips_test_suite with reduced health check interval to
5425 demonstrate periodic health checking. Add "nodh" option to
5426 fips_test_suite to skip very slow DH test.
5430 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5435 * More extensive health check for DRBG checking many more failure modes.
5436 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5437 combination: call this in fips_test_suite.
5441 * Add support for canonical generation of DSA parameter 'g'. See
5444 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5445 POST to handle HMAC cases.
5449 * Add functions FIPS_module_version() and FIPS_module_version_text()
5450 to return numerical and string versions of the FIPS module number.
5454 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5455 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5456 outside the validated module in the FIPS capable OpenSSL.
5460 * Minor change to DRBG entropy callback semantics. In some cases
5461 there is no multiple of the block length between min_len and
5462 max_len. Allow the callback to return more than max_len bytes
5463 of entropy but discard any extra: it is the callback's responsibility
5464 to ensure that the extra data discarded does not impact the
5465 requested amount of entropy.
5469 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5470 information in FIPS186-3, SP800-57 and SP800-131A.
5474 * CCM support via EVP. Interface is very similar to GCM case except we
5475 must supply all data in one chunk (i.e. no update, final) and the
5476 message length must be supplied if AAD is used. Add algorithm test
5481 * Initial version of POST overhaul. Add POST callback to allow the status
5482 of POST to be monitored and/or failures induced. Modify fips_test_suite
5483 to use callback. Always run all selftests even if one fails.
5487 * XTS support including algorithm test driver in the fips_gcmtest program.
5488 Note: this does increase the maximum key length from 32 to 64 bytes but
5489 there should be no binary compatibility issues as existing applications
5490 will never use XTS mode.
5494 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5495 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5496 performs algorithm blocking for unapproved PRNG types. Also do not
5497 set PRNG type in FIPS_mode_set(): leave this to the application.
5498 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5499 the standard OpenSSL PRNG: set additional data to a date time vector.
5503 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5504 This shouldn't present any incompatibility problems because applications
5505 shouldn't be using these directly and any that are will need to rethink
5506 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5510 * Extensive self tests and health checking required by SP800-90 DRBG.
5511 Remove strength parameter from FIPS_drbg_instantiate and always
5512 instantiate at maximum supported strength.
5516 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5520 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5524 * New function DH_compute_key_padded() to compute a DH key and pad with
5525 leading zeroes if needed: this complies with SP800-56A et al.
5529 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5530 anything, incomplete, subject to change and largely untested at present.
5534 * Modify fipscanisteronly build option to only build the necessary object
5535 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5539 * Add experimental option FIPSSYMS to give all symbols in
5540 fipscanister.o and FIPS or fips prefix. This will avoid
5541 conflicts with future versions of OpenSSL. Add perl script
5542 util/fipsas.pl to preprocess assembly language source files
5543 and rename any affected symbols.
5547 * Add selftest checks and algorithm block of non-fips algorithms in
5548 FIPS mode. Remove DES2 from selftests.
5552 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5553 return internal method without any ENGINE dependencies. Add new
5554 tiny fips sign and verify functions.
5558 * New build option no-ec2m to disable characteristic 2 code.
5562 * New build option "fipscanisteronly". This only builds fipscanister.o
5563 and (currently) associated fips utilities. Uses the file Makefile.fips
5564 instead of Makefile.org as the prototype.
5568 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5569 Update fips_gcmtest to use IV generator.
5573 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5574 setting output buffer to NULL. The `*Final` function must be
5575 called although it will not retrieve any additional data. The tag
5576 can be set or retrieved with a ctrl. The IV length is by default 12
5577 bytes (96 bits) but can be set to an alternative value. If the IV
5578 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5583 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5584 underlying do_cipher function handles all cipher semantics itself
5585 including padding and finalisation. This is useful if (for example)
5586 an ENGINE cipher handles block padding itself. The behaviour of
5587 do_cipher is subtly changed if this flag is set: the return value
5588 is the number of characters written to the output buffer (zero is
5589 no longer an error code) or a negative error code. Also if the
5590 input buffer is NULL and length 0 finalisation should be performed.
5594 * If a candidate issuer certificate is already part of the constructed
5595 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5599 * Improve forward-security support: add functions
5601 void SSL_CTX_set_not_resumable_session_callback(
5602 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5603 void SSL_set_not_resumable_session_callback(
5604 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5606 for use by SSL/TLS servers; the callback function will be called whenever a
5607 new session is created, and gets to decide whether the session may be
5608 cached to make it resumable (return 0) or not (return 1). (As by the
5609 SSL/TLS protocol specifications, the session_id sent by the server will be
5610 empty to indicate that the session is not resumable; also, the server will
5611 not generate RFC 4507 (RFC 5077) session tickets.)
5613 A simple reasonable callback implementation is to return is_forward_secure.
5614 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5615 by the SSL/TLS server library, indicating whether it can provide forward
5618 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5620 * New -verify_name option in command line utilities to set verification
5625 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5626 Add CMAC pkey methods.
5630 * Experimental renegotiation in s_server -www mode. If the client
5631 browses /reneg connection is renegotiated. If /renegcert it is
5632 renegotiated requesting a certificate.
5636 * Add an "external" session cache for debugging purposes to s_server. This
5637 should help trace issues which normally are only apparent in deployed
5638 multi-process servers.
5642 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5643 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5644 BIO_set_cipher() and some obscure PEM functions were changed so they
5645 can now return an error. The RAND changes required a change to the
5646 RAND_METHOD structure.
5650 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5651 a gcc attribute to warn if the result of a function is ignored. This
5652 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5653 whose return value is often ignored.
5657 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5658 These allow SCTs (signed certificate timestamps) to be requested and
5659 validated when establishing a connection.
5661 *Rob Percival <robpercival@google.com>*
5666 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5668 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5669 used even when parsing explicit parameters, when loading a encoded key
5670 or calling `EC_GROUP_new_from_ecpkparameters()`/
5671 `EC_GROUP_new_from_ecparameters()`.
5672 This prevents bypass of security hardening and performance gains,
5673 especially for curves with specialized EC_METHODs.
5674 By default, if a key encoded with explicit parameters is loaded and later
5675 encoded, the output is still encoded with explicit parameters, even if
5676 internally a "named" EC_GROUP is used for computation.
5680 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5681 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5682 NULL. After this change, only the cofactor parameter can be NULL. It also
5683 does some minimal sanity checks on the passed order.
5688 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5689 An attack is simple, if the first CMS_recipientInfo is valid but the
5690 second CMS_recipientInfo is chosen ciphertext. If the second
5691 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5692 encryption key will be replaced by garbage, and the message cannot be
5693 decoded, but if the RSA decryption fails, the correct encryption key is
5694 used and the recipient will not notice the attack.
5695 As a work around for this potential attack the length of the decrypted
5696 key must be equal to the cipher default key length, in case the
5697 certificate is not given and all recipientInfo are tried out.
5698 The old behaviour can be re-enabled in the CMS code by setting the
5699 CMS_DEBUG_DECRYPT flag.
5704 * Document issue with installation paths in diverse Windows builds
5706 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5707 binaries and run-time config file.
5712 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5714 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5715 This changes the size when using the `genpkey` command when no size is given.
5716 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5717 generation commands to use 2048 bits by default.
5721 * Add FIPS support for Android Arm 64-bit
5723 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5724 Module in Version 2.0.10. For some reason, the corresponding target
5725 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5726 built with FIPS support on Android Arm 64-bit. This omission has been
5729 *Matthias St. Pierre*
5731 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5733 * 0-byte record padding oracle
5735 If an application encounters a fatal protocol error and then calls
5736 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5737 then OpenSSL can respond differently to the calling application if a 0 byte
5738 record is received with invalid padding compared to if a 0 byte record is
5739 received with an invalid MAC. If the application then behaves differently
5740 based on that in a way that is detectable to the remote peer, then this
5741 amounts to a padding oracle that could be used to decrypt data.
5743 In order for this to be exploitable "non-stitched" ciphersuites must be in
5744 use. Stitched ciphersuites are optimised implementations of certain
5745 commonly used ciphersuites. Also the application must call SSL_shutdown()
5746 twice even if a protocol error has occurred (applications should not do
5747 this but some do anyway).
5749 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5750 Aviram, with additional investigation by Steven Collison and Andrew
5751 Hourselt. It was reported to OpenSSL on 10th December 2018.
5756 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5760 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5762 * Microarchitecture timing vulnerability in ECC scalar multiplication
5764 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5765 shown to be vulnerable to a microarchitecture timing side channel attack.
5766 An attacker with sufficient access to mount local timing attacks during
5767 ECDSA signature generation could recover the private key.
5769 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5770 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5776 * Timing vulnerability in DSA signature generation
5778 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5779 timing side channel attack. An attacker could use variations in the signing
5780 algorithm to recover the private key.
5782 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5787 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5788 Module, accidentally introduced while backporting security fixes from the
5789 development branch and hindering the use of ECC in FIPS mode.
5793 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5795 * Client DoS due to large DH parameter
5797 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5798 malicious server can send a very large prime value to the client. This will
5799 cause the client to spend an unreasonably long period of time generating a
5800 key for this prime resulting in a hang until the client has finished. This
5801 could be exploited in a Denial Of Service attack.
5803 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5808 * Cache timing vulnerability in RSA Key Generation
5810 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5811 a cache timing side channel attack. An attacker with sufficient access to
5812 mount cache timing attacks during the RSA key generation process could
5813 recover the private key.
5815 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5816 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5821 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5822 parameter is no longer accepted, as it leads to a corrupt table. NULL
5823 pem_str is reserved for alias entries only.
5827 * Revert blinding in ECDSA sign and instead make problematic addition
5828 length-invariant. Switch even to fixed-length Montgomery multiplication.
5832 * Change generating and checking of primes so that the error rate of not
5833 being prime depends on the intended use based on the size of the input.
5834 For larger primes this will result in more rounds of Miller-Rabin.
5835 The maximal error rate for primes with more than 1080 bits is lowered
5838 *Kurt Roeckx, Annie Yousar*
5840 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5844 * Add blinding to ECDSA and DSA signatures to protect against side channel
5845 attacks discovered by Keegan Ryan (NCC Group).
5849 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5850 now allow empty (zero character) pass phrases.
5854 * Certificate time validation (X509_cmp_time) enforces stricter
5855 compliance with RFC 5280. Fractional seconds and timezone offsets
5856 are no longer allowed.
5860 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5862 * Constructed ASN.1 types with a recursive definition could exceed the stack
5864 Constructed ASN.1 types with a recursive definition (such as can be found
5865 in PKCS7) could eventually exceed the stack given malicious input with
5866 excessive recursion. This could result in a Denial Of Service attack. There
5867 are no such structures used within SSL/TLS that come from untrusted sources
5868 so this is considered safe.
5870 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5876 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5878 * Read/write after SSL object in error state
5880 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5881 mechanism. The intent was that if a fatal error occurred during a handshake
5882 then OpenSSL would move into the error state and would immediately fail if
5883 you attempted to continue the handshake. This works as designed for the
5884 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5885 SSL_connect()), however due to a bug it does not work correctly if
5886 SSL_read() or SSL_write() is called directly. In that scenario, if the
5887 handshake fails then a fatal error will be returned in the initial function
5888 call. If SSL_read()/SSL_write() is subsequently called by the application
5889 for the same SSL object then it will succeed and the data is passed without
5890 being decrypted/encrypted directly from the SSL/TLS record layer.
5892 In order to exploit this issue an application bug would have to be present
5893 that resulted in a call to SSL_read()/SSL_write() being issued after having
5894 already received a fatal error.
5896 This issue was reported to OpenSSL by David Benjamin (Google).
5901 * rsaz_1024_mul_avx2 overflow bug on x86_64
5903 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5904 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5905 Analysis suggests that attacks against RSA and DSA as a result of this
5906 defect would be very difficult to perform and are not believed likely.
5907 Attacks against DH1024 are considered just feasible, because most of the
5908 work necessary to deduce information about a private key may be performed
5909 offline. The amount of resources required for such an attack would be
5910 significant. However, for an attack on TLS to be meaningful, the server
5911 would have to share the DH1024 private key among multiple clients, which is
5912 no longer an option since CVE-2016-0701.
5914 This only affects processors that support the AVX2 but not ADX extensions
5915 like Intel Haswell (4th generation).
5917 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5918 was originally found via the OSS-Fuzz project.
5923 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5925 * bn_sqrx8x_internal carry bug on x86_64
5927 There is a carry propagating bug in the x86_64 Montgomery squaring
5928 procedure. No EC algorithms are affected. Analysis suggests that attacks
5929 against RSA and DSA as a result of this defect would be very difficult to
5930 perform and are not believed likely. Attacks against DH are considered just
5931 feasible (although very difficult) because most of the work necessary to
5932 deduce information about a private key may be performed offline. The amount
5933 of resources required for such an attack would be very significant and
5934 likely only accessible to a limited number of attackers. An attacker would
5935 additionally need online access to an unpatched system using the target
5936 private key in a scenario with persistent DH parameters and a private
5937 key that is shared between multiple clients.
5939 This only affects processors that support the BMI1, BMI2 and ADX extensions
5940 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5942 This issue was reported to OpenSSL by the OSS-Fuzz project.
5947 * Malformed X.509 IPAddressFamily could cause OOB read
5949 If an X.509 certificate has a malformed IPAddressFamily extension,
5950 OpenSSL could do a one-byte buffer overread. The most likely result
5951 would be an erroneous display of the certificate in text format.
5953 This issue was reported to OpenSSL by the OSS-Fuzz project.
5957 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5959 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5960 platform rather than 'mingw'.
5964 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5966 * Truncated packet could crash via OOB read
5968 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5969 cipher is being used, then a truncated packet can cause that host to
5970 perform an out-of-bounds read, usually resulting in a crash.
5972 This issue was reported to OpenSSL by Robert Święcki of Google.
5977 * BN_mod_exp may produce incorrect results on x86_64
5979 There is a carry propagating bug in the x86_64 Montgomery squaring
5980 procedure. No EC algorithms are affected. Analysis suggests that attacks
5981 against RSA and DSA as a result of this defect would be very difficult to
5982 perform and are not believed likely. Attacks against DH are considered just
5983 feasible (although very difficult) because most of the work necessary to
5984 deduce information about a private key may be performed offline. The amount
5985 of resources required for such an attack would be very significant and
5986 likely only accessible to a limited number of attackers. An attacker would
5987 additionally need online access to an unpatched system using the target
5988 private key in a scenario with persistent DH parameters and a private
5989 key that is shared between multiple clients. For example this can occur by
5990 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5991 similar to CVE-2015-3193 but must be treated as a separate problem.
5993 This issue was reported to OpenSSL by the OSS-Fuzz project.
5998 * Montgomery multiplication may produce incorrect results
6000 There is a carry propagating bug in the Broadwell-specific Montgomery
6001 multiplication procedure that handles input lengths divisible by, but
6002 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
6003 and DH private keys are impossible. This is because the subroutine in
6004 question is not used in operations with the private key itself and an input
6005 of the attacker's direct choice. Otherwise the bug can manifest itself as
6006 transient authentication and key negotiation failures or reproducible
6007 erroneous outcome of public-key operations with specially crafted input.
6008 Among EC algorithms only Brainpool P-512 curves are affected and one
6009 presumably can attack ECDH key negotiation. Impact was not analyzed in
6010 detail, because pre-requisites for attack are considered unlikely. Namely
6011 multiple clients have to choose the curve in question and the server has to
6012 share the private key among them, neither of which is default behaviour.
6013 Even then only clients that chose the curve will be affected.
6015 This issue was publicly reported as transient failures and was not
6016 initially recognized as a security issue. Thanks to Richard Morgan for
6017 providing reproducible case.
6022 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
6023 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
6024 prevent issues where no progress is being made and the peer continually
6025 sends unrecognised record types, using up resources processing them.
6029 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
6031 * Missing CRL sanity check
6033 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
6034 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
6035 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
6037 This issue only affects the OpenSSL 1.0.2i
6042 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
6044 * OCSP Status Request extension unbounded memory growth
6046 A malicious client can send an excessively large OCSP Status Request
6047 extension. If that client continually requests renegotiation, sending a
6048 large OCSP Status Request extension each time, then there will be unbounded
6049 memory growth on the server. This will eventually lead to a Denial Of
6050 Service attack through memory exhaustion. Servers with a default
6051 configuration are vulnerable even if they do not support OCSP. Builds using
6052 the "no-ocsp" build time option are not affected.
6054 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6059 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6062 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6068 * OOB write in MDC2_Update()
6070 An overflow can occur in MDC2_Update() either if called directly or
6071 through the EVP_DigestUpdate() function using MDC2. If an attacker
6072 is able to supply very large amounts of input data after a previous
6073 call to EVP_EncryptUpdate() with a partial block then a length check
6074 can overflow resulting in a heap corruption.
6076 The amount of data needed is comparable to SIZE_MAX which is impractical
6079 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6084 * Malformed SHA512 ticket DoS
6086 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6087 DoS attack where a malformed ticket will result in an OOB read which will
6090 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6091 a custom server callback and ticket lookup mechanism.
6093 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6098 * OOB write in BN_bn2dec()
6100 The function BN_bn2dec() does not check the return value of BN_div_word().
6101 This can cause an OOB write if an application uses this function with an
6102 overly large BIGNUM. This could be a problem if an overly large certificate
6103 or CRL is printed out from an untrusted source. TLS is not affected because
6104 record limits will reject an oversized certificate before it is parsed.
6106 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6111 * OOB read in TS_OBJ_print_bio()
6113 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6114 the total length the OID text representation would use and not the amount
6115 of data written. This will result in OOB reads when large OIDs are
6118 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6123 * Pointer arithmetic undefined behaviour
6125 Avoid some undefined pointer arithmetic
6127 A common idiom in the codebase is to check limits in the following manner:
6130 Where "p" points to some malloc'd data of SIZE bytes and
6133 "len" here could be from some externally supplied data (e.g. from a TLS
6136 The rules of C pointer arithmetic are such that "p + len" is only well
6137 defined where len <= SIZE. Therefore the above idiom is actually
6138 undefined behaviour.
6140 For example this could cause problems if some malloc implementation
6141 provides an address for "p" such that "p + len" actually overflows for
6142 values of len that are too big and therefore p + len < limit.
6144 This issue was reported to OpenSSL by Guido Vranken
6149 * Constant time flag not preserved in DSA signing
6151 Operations in the DSA signing algorithm should run in constant time in
6152 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6153 implementation means that a non-constant time codepath is followed for
6154 certain operations. This has been demonstrated through a cache-timing
6155 attack to be sufficient for an attacker to recover the private DSA key.
6157 This issue was reported by César Pereida (Aalto University), Billy Brumley
6158 (Tampere University of Technology), and Yuval Yarom (The University of
6159 Adelaide and NICTA).
6164 * DTLS buffered message DoS
6166 In a DTLS connection where handshake messages are delivered out-of-order
6167 those messages that OpenSSL is not yet ready to process will be buffered
6168 for later use. Under certain circumstances, a flaw in the logic means that
6169 those messages do not get removed from the buffer even though the handshake
6170 has been completed. An attacker could force up to approx. 15 messages to
6171 remain in the buffer when they are no longer required. These messages will
6172 be cleared when the DTLS connection is closed. The default maximum size for
6173 a message is 100k. Therefore, the attacker could force an additional 1500k
6174 to be consumed per connection. By opening many simultaneous connections an
6175 attacker could cause a DoS attack through memory exhaustion.
6177 This issue was reported to OpenSSL by Quan Luo.
6182 * DTLS replay protection DoS
6184 A flaw in the DTLS replay attack protection mechanism means that records
6185 that arrive for future epochs update the replay protection "window" before
6186 the MAC for the record has been validated. This could be exploited by an
6187 attacker by sending a record for the next epoch (which does not have to
6188 decrypt or have a valid MAC), with a very large sequence number. This means
6189 that all subsequent legitimate packets are dropped causing a denial of
6190 service for a specific DTLS connection.
6192 This issue was reported to OpenSSL by the OCAP audit team.
6197 * Certificate message OOB reads
6199 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6200 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6201 theoretical DoS risk but this has not been observed in practice on common
6204 The messages affected are client certificate, client certificate request
6205 and server certificate. As a result the attack can only be performed
6206 against a client or a server which enables client authentication.
6208 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6213 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
6215 * Prevent padding oracle in AES-NI CBC MAC check
6217 A MITM attacker can use a padding oracle attack to decrypt traffic
6218 when the connection uses an AES CBC cipher and the server support
6221 This issue was introduced as part of the fix for Lucky 13 padding
6222 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6223 constant time by making sure that always the same bytes are read and
6224 compared against either the MAC or padding bytes. But it no longer
6225 checked that there was enough data to have both the MAC and padding
6228 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6232 * Fix EVP_EncodeUpdate overflow
6234 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6235 Base64 encoding of binary data. If an attacker is able to supply very large
6236 amounts of input data then a length check can overflow resulting in a heap
6239 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6240 the `PEM_write_bio*` family of functions. These are mainly used within the
6241 OpenSSL command line applications, so any application which processes data
6242 from an untrusted source and outputs it as a PEM file should be considered
6243 vulnerable to this issue. User applications that call these APIs directly
6244 with large amounts of untrusted data may also be vulnerable.
6246 This issue was reported by Guido Vranken.
6251 * Fix EVP_EncryptUpdate overflow
6253 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6254 is able to supply very large amounts of input data after a previous call to
6255 EVP_EncryptUpdate() with a partial block then a length check can overflow
6256 resulting in a heap corruption. Following an analysis of all OpenSSL
6257 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6258 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6259 the first called function after an EVP_EncryptInit(), and therefore that
6260 specific call must be safe. The second form is where the length passed to
6261 EVP_EncryptUpdate() can be seen from the code to be some small value and
6262 therefore there is no possibility of an overflow. Since all instances are
6263 one of these two forms, it is believed that there can be no overflows in
6264 internal code due to this problem. It should be noted that
6265 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6266 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6267 of these calls have also been analysed too and it is believed there are no
6268 instances in internal usage where an overflow could occur.
6270 This issue was reported by Guido Vranken.
6275 * Prevent ASN.1 BIO excessive memory allocation
6277 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6278 a short invalid encoding can cause allocation of large amounts of memory
6279 potentially consuming excessive resources or exhausting memory.
6281 Any application parsing untrusted data through d2i BIO functions is
6282 affected. The memory based functions such as d2i_X509() are *not* affected.
6283 Since the memory based functions are used by the TLS library, TLS
6284 applications are not affected.
6286 This issue was reported by Brian Carpenter.
6293 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6294 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6295 in arbitrary stack data being returned in the buffer.
6297 This issue was reported by Guido Vranken.
6302 * Modify behavior of ALPN to invoke callback after SNI/servername
6303 callback, such that updates to the SSL_CTX affect ALPN.
6307 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6312 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6313 methods are enabled and ssl2 is disabled the methods return NULL.
6317 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
6319 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6320 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6321 provide any "EXPORT" or "LOW" strength ciphers.
6325 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6326 is by default disabled at build-time. Builds that are not configured with
6327 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6328 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6329 will need to explicitly call either of:
6331 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6333 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6335 as appropriate. Even if either of those is used, or the application
6336 explicitly uses the version-specific SSLv2_method() or its client and
6337 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6338 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6339 ciphers, and SSLv2 56-bit DES are no longer available.
6344 * Fix a double-free in DSA code
6346 A double free bug was discovered when OpenSSL parses malformed DSA private
6347 keys and could lead to a DoS attack or memory corruption for applications
6348 that receive DSA private keys from untrusted sources. This scenario is
6351 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6357 * Disable SRP fake user seed to address a server memory leak.
6359 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6361 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6362 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6363 was changed to ignore the "fake user" SRP seed, even if the seed
6366 Users should use SRP_VBASE_get1_by_user instead. Note that in
6367 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6368 also that even though configuring the SRP seed attempts to hide
6369 invalid usernames by continuing the handshake with fake
6370 credentials, this behaviour is not constant time and no strong
6371 guarantees are made that the handshake is indistinguishable from
6372 that of a valid user.
6377 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6379 In the BN_hex2bn function the number of hex digits is calculated using an
6380 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6381 large values of `i` this can result in `bn_expand` not allocating any
6382 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6383 field as NULL leading to a subsequent NULL ptr deref. For very large values
6384 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6385 In this case memory is allocated to the internal BIGNUM data field, but it
6386 is insufficiently sized leading to heap corruption. A similar issue exists
6387 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6388 is ever called by user applications with very large untrusted hex/dec data.
6389 This is anticipated to be a rare occurrence.
6391 All OpenSSL internal usage of these functions use data that is not expected
6392 to be untrusted, e.g. config file data or application command line
6393 arguments. If user developed applications generate config file data based
6394 on untrusted data then it is possible that this could also lead to security
6395 consequences. This is also anticipated to be rare.
6397 This issue was reported to OpenSSL by Guido Vranken.
6402 * Fix memory issues in `BIO_*printf` functions
6404 The internal `fmtstr` function used in processing a "%s" format string in
6405 the `BIO_*printf` functions could overflow while calculating the length of a
6406 string and cause an OOB read when printing very long strings.
6408 Additionally the internal `doapr_outch` function can attempt to write to an
6409 OOB memory location (at an offset from the NULL pointer) in the event of a
6410 memory allocation failure. In 1.0.2 and below this could be caused where
6411 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6412 could be in processing a very long "%s" format string. Memory leaks can
6415 The first issue may mask the second issue dependent on compiler behaviour.
6416 These problems could enable attacks where large amounts of untrusted data
6417 is passed to the `BIO_*printf` functions. If applications use these functions
6418 in this way then they could be vulnerable. OpenSSL itself uses these
6419 functions when printing out human-readable dumps of ASN.1 data. Therefore
6420 applications that print this data could be vulnerable if the data is from
6421 untrusted sources. OpenSSL command line applications could also be
6422 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6423 as command line arguments.
6425 Libssl is not considered directly vulnerable. Additionally certificates etc
6426 received via remote connections via libssl are also unlikely to be able to
6427 trigger these issues because of message size limits enforced within libssl.
6429 This issue was reported to OpenSSL Guido Vranken.
6434 * Side channel attack on modular exponentiation
6436 A side-channel attack was found which makes use of cache-bank conflicts on
6437 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6438 of RSA keys. The ability to exploit this issue is limited as it relies on
6439 an attacker who has control of code in a thread running on the same
6440 hyper-threaded core as the victim thread which is performing decryptions.
6442 This issue was reported to OpenSSL by Yuval Yarom, The University of
6443 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6444 Nadia Heninger, University of Pennsylvania with more information at
6445 <http://cachebleed.info>.
6450 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6451 if no keysize is specified with default_bits. This fixes an
6452 omission in an earlier change that changed all RSA/DSA key generation
6453 commands to use 2048 bits by default.
6457 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6459 * DH small subgroups
6461 Historically OpenSSL only ever generated DH parameters based on "safe"
6462 primes. More recently (in version 1.0.2) support was provided for
6463 generating X9.42 style parameter files such as those required for RFC 5114
6464 support. The primes used in such files may not be "safe". Where an
6465 application is using DH configured with parameters based on primes that are
6466 not "safe" then an attacker could use this fact to find a peer's private
6467 DH exponent. This attack requires that the attacker complete multiple
6468 handshakes in which the peer uses the same private DH exponent. For example
6469 this could be used to discover a TLS server's private DH exponent if it's
6470 reusing the private DH exponent or it's using a static DH ciphersuite.
6472 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6473 TLS. It is not on by default. If the option is not set then the server
6474 reuses the same private DH exponent for the life of the server process and
6475 would be vulnerable to this attack. It is believed that many popular
6476 applications do set this option and would therefore not be at risk.
6478 The fix for this issue adds an additional check where a "q" parameter is
6479 available (as is the case in X9.42 based parameters). This detects the
6480 only known attack, and is the only possible defense for static DH
6481 ciphersuites. This could have some performance impact.
6483 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6484 default and cannot be disabled. This could have some performance impact.
6486 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6491 * SSLv2 doesn't block disabled ciphers
6493 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6494 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6495 been disabled, provided that the SSLv2 protocol was not also disabled via
6498 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6499 and Sebastian Schinzel.
6504 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6506 * BN_mod_exp may produce incorrect results on x86_64
6508 There is a carry propagating bug in the x86_64 Montgomery squaring
6509 procedure. No EC algorithms are affected. Analysis suggests that attacks
6510 against RSA and DSA as a result of this defect would be very difficult to
6511 perform and are not believed likely. Attacks against DH are considered just
6512 feasible (although very difficult) because most of the work necessary to
6513 deduce information about a private key may be performed offline. The amount
6514 of resources required for such an attack would be very significant and
6515 likely only accessible to a limited number of attackers. An attacker would
6516 additionally need online access to an unpatched system using the target
6517 private key in a scenario with persistent DH parameters and a private
6518 key that is shared between multiple clients. For example this can occur by
6519 default in OpenSSL DHE based SSL/TLS ciphersuites.
6521 This issue was reported to OpenSSL by Hanno Böck.
6526 * Certificate verify crash with missing PSS parameter
6528 The signature verification routines will crash with a NULL pointer
6529 dereference if presented with an ASN.1 signature using the RSA PSS
6530 algorithm and absent mask generation function parameter. Since these
6531 routines are used to verify certificate signature algorithms this can be
6532 used to crash any certificate verification operation and exploited in a
6533 DoS attack. Any application which performs certificate verification is
6534 vulnerable including OpenSSL clients and servers which enable client
6537 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6542 * X509_ATTRIBUTE memory leak
6544 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6545 memory. This structure is used by the PKCS#7 and CMS routines so any
6546 application which reads PKCS#7 or CMS data from untrusted sources is
6547 affected. SSL/TLS is not affected.
6549 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6555 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6556 This changes the decoding behaviour for some invalid messages,
6557 though the change is mostly in the more lenient direction, and
6558 legacy behaviour is preserved as much as possible.
6562 * In DSA_generate_parameters_ex, if the provided seed is too short,
6565 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6567 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6569 * Alternate chains certificate forgery
6571 During certificate verification, OpenSSL will attempt to find an
6572 alternative certificate chain if the first attempt to build such a chain
6573 fails. An error in the implementation of this logic can mean that an
6574 attacker could cause certain checks on untrusted certificates to be
6575 bypassed, such as the CA flag, enabling them to use a valid leaf
6576 certificate to act as a CA and "issue" an invalid certificate.
6578 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6583 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6585 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6586 incompatibility in the handling of HMAC. The previous ABI has now been
6591 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6593 * Malformed ECParameters causes infinite loop
6595 When processing an ECParameters structure OpenSSL enters an infinite loop
6596 if the curve specified is over a specially malformed binary polynomial
6599 This can be used to perform denial of service against any
6600 system which processes public keys, certificate requests or
6601 certificates. This includes TLS clients and TLS servers with
6602 client authentication enabled.
6604 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6609 * Exploitable out-of-bounds read in X509_cmp_time
6611 X509_cmp_time does not properly check the length of the ASN1_TIME
6612 string and can read a few bytes out of bounds. In addition,
6613 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6616 An attacker can use this to craft malformed certificates and CRLs of
6617 various sizes and potentially cause a segmentation fault, resulting in
6618 a DoS on applications that verify certificates or CRLs. TLS clients
6619 that verify CRLs are affected. TLS clients and servers with client
6620 authentication enabled may be affected if they use custom verification
6623 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6624 independently by Hanno Böck.
6629 * PKCS7 crash with missing EnvelopedContent
6631 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6632 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6633 with missing content and trigger a NULL pointer dereference on parsing.
6635 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6636 structures from untrusted sources are affected. OpenSSL clients and
6637 servers are not affected.
6639 This issue was reported to OpenSSL by Michal Zalewski (Google).
6644 * CMS verify infinite loop with unknown hash function
6646 When verifying a signedData message the CMS code can enter an infinite loop
6647 if presented with an unknown hash function OID. This can be used to perform
6648 denial of service against any system which verifies signedData messages using
6650 This issue was reported to OpenSSL by Johannes Bauer.
6655 * Race condition handling NewSessionTicket
6657 If a NewSessionTicket is received by a multi-threaded client when attempting to
6658 reuse a previous ticket then a race condition can occur potentially leading to
6659 a double free of the ticket data.
6664 * Only support 256-bit or stronger elliptic curves with the
6665 'ecdh_auto' setting (server) or by default (client). Of supported
6666 curves, prefer P-256 (both).
6670 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6672 * ClientHello sigalgs DoS fix
6674 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6675 invalid signature algorithms extension a NULL pointer dereference will
6676 occur. This can be exploited in a DoS attack against the server.
6678 This issue was was reported to OpenSSL by David Ramos of Stanford
6682 *Stephen Henson and Matt Caswell*
6684 * Multiblock corrupted pointer fix
6686 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6687 feature only applies on 64 bit x86 architecture platforms that support AES
6688 NI instructions. A defect in the implementation of "multiblock" can cause
6689 OpenSSL's internal write buffer to become incorrectly set to NULL when
6690 using non-blocking IO. Typically, when the user application is using a
6691 socket BIO for writing, this will only result in a failed connection.
6692 However if some other BIO is used then it is likely that a segmentation
6693 fault will be triggered, thus enabling a potential DoS attack.
6695 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6700 * Segmentation fault in DTLSv1_listen fix
6702 The DTLSv1_listen function is intended to be stateless and processes the
6703 initial ClientHello from many peers. It is common for user code to loop
6704 over the call to DTLSv1_listen until a valid ClientHello is received with
6705 an associated cookie. A defect in the implementation of DTLSv1_listen means
6706 that state is preserved in the SSL object from one invocation to the next
6707 that can lead to a segmentation fault. Errors processing the initial
6708 ClientHello can trigger this scenario. An example of such an error could be
6709 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6712 This issue was reported to OpenSSL by Per Allansson.
6717 * Segmentation fault in ASN1_TYPE_cmp fix
6719 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6720 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6721 certificate signature algorithm consistency this can be used to crash any
6722 certificate verification operation and exploited in a DoS attack. Any
6723 application which performs certificate verification is vulnerable including
6724 OpenSSL clients and servers which enable client authentication.
6729 * Segmentation fault for invalid PSS parameters fix
6731 The signature verification routines will crash with a NULL pointer
6732 dereference if presented with an ASN.1 signature using the RSA PSS
6733 algorithm and invalid parameters. Since these routines are used to verify
6734 certificate signature algorithms this can be used to crash any
6735 certificate verification operation and exploited in a DoS attack. Any
6736 application which performs certificate verification is vulnerable including
6737 OpenSSL clients and servers which enable client authentication.
6739 This issue was was reported to OpenSSL by Brian Carpenter.
6744 * ASN.1 structure reuse memory corruption fix
6746 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6747 memory corruption via an invalid write. Such reuse is and has been
6748 strongly discouraged and is believed to be rare.
6750 Applications that parse structures containing CHOICE or ANY DEFINED BY
6751 components may be affected. Certificate parsing (d2i_X509 and related
6752 functions) are however not affected. OpenSSL clients and servers are
6758 * PKCS7 NULL pointer dereferences fix
6760 The PKCS#7 parsing code does not handle missing outer ContentInfo
6761 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6762 missing content and trigger a NULL pointer dereference on parsing.
6764 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6765 otherwise parse PKCS#7 structures from untrusted sources are
6766 affected. OpenSSL clients and servers are not affected.
6768 This issue was reported to OpenSSL by Michal Zalewski (Google).
6773 * DoS via reachable assert in SSLv2 servers fix
6775 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6776 servers that both support SSLv2 and enable export cipher suites by sending
6777 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6779 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6780 (OpenSSL development team).
6785 * Empty CKE with client auth and DHE fix
6787 If client auth is used then a server can seg fault in the event of a DHE
6788 ciphersuite being selected and a zero length ClientKeyExchange message
6789 being sent by the client. This could be exploited in a DoS attack.
6794 * Handshake with unseeded PRNG fix
6796 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6797 with an unseeded PRNG. The conditions are:
6798 - The client is on a platform where the PRNG has not been seeded
6799 automatically, and the user has not seeded manually
6800 - A protocol specific client method version has been used (i.e. not
6801 SSL_client_methodv23)
6802 - A ciphersuite is used that does not require additional random data from
6803 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6805 If the handshake succeeds then the client random that has been used will
6806 have been generated from a PRNG with insufficient entropy and therefore the
6807 output may be predictable.
6809 For example using the following command with an unseeded openssl will
6810 succeed on an unpatched platform:
6812 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6817 * Use After Free following d2i_ECPrivatekey error fix
6819 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6820 could cause a use after free condition. This, in turn, could cause a double
6821 free in several private key parsing functions (such as d2i_PrivateKey
6822 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6823 for applications that receive EC private keys from untrusted
6824 sources. This scenario is considered rare.
6826 This issue was discovered by the BoringSSL project and fixed in their
6832 * X509_to_X509_REQ NULL pointer deref fix
6834 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6835 the certificate key is invalid. This function is rarely used in practice.
6837 This issue was discovered by Brian Carpenter.
6842 * Removed the export ciphers from the DEFAULT ciphers
6846 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6848 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6849 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6850 So far those who have to target multiple platforms would compromise
6851 and argue that binary targeting say ARMv5 would still execute on
6852 ARMv8. "Universal" build resolves this compromise by providing
6853 near-optimal performance even on newer platforms.
6857 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6858 (other platforms pending).
6860 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6862 * Add support for the SignedCertificateTimestampList certificate and
6863 OCSP response extensions from RFC6962.
6867 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6868 for corner cases. (Certain input points at infinity could lead to
6869 bogus results, with non-infinity inputs mapped to infinity too.)
6873 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6874 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6875 common cases are optimized and there still is room for further
6876 improvements. Vector Permutation AES for Altivec is also added.
6880 * Add support for little-endian ppc64 Linux target.
6882 *Marcelo Cerri (IBM)*
6884 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6885 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6886 are optimized and there still is room for further improvements.
6887 Both 32- and 64-bit modes are supported.
6889 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6891 * Improved ARMv7 NEON support.
6895 * Support for SPARC Architecture 2011 crypto extensions, first
6896 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6897 SHA256/512, MD5, GHASH and modular exponentiation.
6899 *Andy Polyakov, David Miller*
6901 * Accelerated modular exponentiation for Intel processors, a.k.a.
6904 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6906 * Support for new and upcoming Intel processors, including AVX2,
6907 BMI and SHA ISA extensions. This includes additional "stitched"
6908 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6911 This work was sponsored by Intel Corp.
6915 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6916 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6917 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6921 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6922 this fixes a limitation in previous versions of OpenSSL.
6926 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6927 MGF1 digest and OAEP label.
6931 * Add EVP support for key wrapping algorithms, to avoid problems with
6932 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6933 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6934 algorithms and include tests cases.
6938 * Add functions to allocate and set the fields of an ECDSA_METHOD
6941 *Douglas E. Engert, Steve Henson*
6943 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6944 difference in days and seconds between two tm or ASN1_TIME structures.
6948 * Add -rev test option to s_server to just reverse order of characters
6949 received by client and send back to server. Also prints an abbreviated
6950 summary of the connection parameters.
6954 * New option -brief for s_client and s_server to print out a brief summary
6955 of connection parameters.
6959 * Add callbacks for arbitrary TLS extensions.
6961 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6963 * New option -crl_download in several openssl utilities to download CRLs
6964 from CRLDP extension in certificates.
6968 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6972 * New function X509_CRL_diff to generate a delta CRL from the difference
6973 of two full CRLs. Add support to "crl" utility.
6977 * New functions to set lookup_crls function and to retrieve
6978 X509_STORE from X509_STORE_CTX.
6982 * Print out deprecated issuer and subject unique ID fields in
6987 * Extend OCSP I/O functions so they can be used for simple general purpose
6988 HTTP as well as OCSP. New wrapper function which can be used to download
6989 CRLs using the OCSP API.
6993 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6997 * `SSL_CONF*` functions. These provide a common framework for application
6998 configuration using configuration files or command lines.
7002 * SSL/TLS tracing code. This parses out SSL/TLS records using the
7003 message callback and prints the results. Needs compile time option
7004 "enable-ssl-trace". New options to s_client and s_server to enable
7009 * New ctrl and macro to retrieve supported points extensions.
7010 Print out extension in s_server and s_client.
7014 * New functions to retrieve certificate signature and signature
7019 * Add functions to retrieve and manipulate the raw cipherlist sent by a
7024 * New Suite B modes for TLS code. These use and enforce the requirements
7025 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
7026 only use Suite B curves. The Suite B modes can be set by using the
7027 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
7031 * New chain verification flags for Suite B levels of security. Check
7032 algorithms are acceptable when flags are set in X509_verify_cert.
7036 * Make tls1_check_chain return a set of flags indicating checks passed
7037 by a certificate chain. Add additional tests to handle client
7038 certificates: checks for matching certificate type and issuer name
7043 * If an attempt is made to use a signature algorithm not in the peer
7044 preference list abort the handshake. If client has no suitable
7045 signature algorithms in response to a certificate request do not
7046 use the certificate.
7050 * If server EC tmp key is not in client preference list abort handshake.
7054 * Add support for certificate stores in CERT structure. This makes it
7055 possible to have different stores per SSL structure or one store in
7056 the parent SSL_CTX. Include distinct stores for certificate chain
7057 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
7058 to build and store a certificate chain in CERT structure: returning
7059 an error if the chain cannot be built: this will allow applications
7060 to test if a chain is correctly configured.
7062 Note: if the CERT based stores are not set then the parent SSL_CTX
7063 store is used to retain compatibility with existing behaviour.
7067 * New function ssl_set_client_disabled to set a ciphersuite disabled
7068 mask based on the current session, check mask when sending client
7069 hello and checking the requested ciphersuite.
7073 * New ctrls to retrieve and set certificate types in a certificate
7074 request message. Print out received values in s_client. If certificate
7075 types is not set with custom values set sensible values based on
7076 supported signature algorithms.
7080 * Support for distinct client and server supported signature algorithms.
7084 * Add certificate callback. If set this is called whenever a certificate
7085 is required by client or server. An application can decide which
7086 certificate chain to present based on arbitrary criteria: for example
7087 supported signature algorithms. Add very simple example to s_server.
7088 This fixes many of the problems and restrictions of the existing client
7089 certificate callback: for example you can now clear an existing
7090 certificate and specify the whole chain.
7094 * Add new "valid_flags" field to CERT_PKEY structure which determines what
7095 the certificate can be used for (if anything). Set valid_flags field
7096 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
7097 to have similar checks in it.
7099 Add new "cert_flags" field to CERT structure and include a "strict mode".
7100 This enforces some TLS certificate requirements (such as only permitting
7101 certificate signature algorithms contained in the supported algorithms
7102 extension) which some implementations ignore: this option should be used
7103 with caution as it could cause interoperability issues.
7107 * Update and tidy signature algorithm extension processing. Work out
7108 shared signature algorithms based on preferences and peer algorithms
7109 and print them out in s_client and s_server. Abort handshake if no
7110 shared signature algorithms.
7114 * Add new functions to allow customised supported signature algorithms
7115 for SSL and SSL_CTX structures. Add options to s_client and s_server
7120 * New function SSL_certs_clear() to delete all references to certificates
7121 from an SSL structure. Before this once a certificate had been added
7122 it couldn't be removed.
7126 * Integrate hostname, email address and IP address checking with certificate
7127 verification. New verify options supporting checking in openssl utility.
7131 * Fixes and wildcard matching support to hostname and email checking
7132 functions. Add manual page.
7134 *Florian Weimer (Red Hat Product Security Team)*
7136 * New functions to check a hostname email or IP address against a
7137 certificate. Add options x509 utility to print results of checks against
7142 * Fix OCSP checking.
7144 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
7146 * Initial experimental support for explicitly trusted non-root CAs.
7147 OpenSSL still tries to build a complete chain to a root but if an
7148 intermediate CA has a trust setting included that is used. The first
7149 setting is used: whether to trust (e.g., -addtrust option to the x509
7154 * Add -trusted_first option which attempts to find certificates in the
7155 trusted store even if an untrusted chain is also supplied.
7159 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
7160 platform support for Linux and Android.
7164 * Support for linux-x32, ILP32 environment in x86_64 framework.
7168 * Experimental multi-implementation support for FIPS capable OpenSSL.
7169 When in FIPS mode the approved implementations are used as normal,
7170 when not in FIPS mode the internal unapproved versions are used instead.
7171 This means that the FIPS capable OpenSSL isn't forced to use the
7172 (often lower performance) FIPS implementations outside FIPS mode.
7176 * Transparently support X9.42 DH parameters when calling
7177 PEM_read_bio_DHparameters. This means existing applications can handle
7178 the new parameter format automatically.
7182 * Initial experimental support for X9.42 DH parameter format: mainly
7183 to support use of 'q' parameter for RFC5114 parameters.
7187 * Add DH parameters from RFC5114 including test data to dhtest.
7191 * Support for automatic EC temporary key parameter selection. If enabled
7192 the most preferred EC parameters are automatically used instead of
7193 hardcoded fixed parameters. Now a server just has to call:
7194 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
7195 support ECDH and use the most appropriate parameters.
7199 * Enhance and tidy EC curve and point format TLS extension code. Use
7200 static structures instead of allocation if default values are used.
7201 New ctrls to set curves we wish to support and to retrieve shared curves.
7202 Print out shared curves in s_server. New options to s_server and s_client
7203 to set list of supported curves.
7207 * New ctrls to retrieve supported signature algorithms and
7208 supported curve values as an array of NIDs. Extend openssl utility
7209 to print out received values.
7213 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
7214 between NIDs and the more common NIST names such as "P-256". Enhance
7215 ecparam utility and ECC method to recognise the NIST names for curves.
7219 * Enhance SSL/TLS certificate chain handling to support different
7220 chains for each certificate instead of one chain in the parent SSL_CTX.
7224 * Support for fixed DH ciphersuite client authentication: where both
7225 server and client use DH certificates with common parameters.
7229 * Support for fixed DH ciphersuites: those requiring DH server
7234 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
7236 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
7237 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
7238 X509_CINF_get_signature were reverted post internal team review.
7243 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
7245 * OCSP Status Request extension unbounded memory growth
7247 A malicious client can send an excessively large OCSP Status Request
7248 extension. If that client continually requests renegotiation, sending a
7249 large OCSP Status Request extension each time, then there will be unbounded
7250 memory growth on the server. This will eventually lead to a Denial Of
7251 Service attack through memory exhaustion. Servers with a default
7252 configuration are vulnerable even if they do not support OCSP. Builds using
7253 the "no-ocsp" build time option are not affected.
7255 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7260 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
7263 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
7269 * OOB write in MDC2_Update()
7271 An overflow can occur in MDC2_Update() either if called directly or
7272 through the EVP_DigestUpdate() function using MDC2. If an attacker
7273 is able to supply very large amounts of input data after a previous
7274 call to EVP_EncryptUpdate() with a partial block then a length check
7275 can overflow resulting in a heap corruption.
7277 The amount of data needed is comparable to SIZE_MAX which is impractical
7280 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7285 * Malformed SHA512 ticket DoS
7287 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
7288 DoS attack where a malformed ticket will result in an OOB read which will
7291 The use of SHA512 in TLS session tickets is comparatively rare as it requires
7292 a custom server callback and ticket lookup mechanism.
7294 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7299 * OOB write in BN_bn2dec()
7301 The function BN_bn2dec() does not check the return value of BN_div_word().
7302 This can cause an OOB write if an application uses this function with an
7303 overly large BIGNUM. This could be a problem if an overly large certificate
7304 or CRL is printed out from an untrusted source. TLS is not affected because
7305 record limits will reject an oversized certificate before it is parsed.
7307 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7312 * OOB read in TS_OBJ_print_bio()
7314 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
7315 the total length the OID text representation would use and not the amount
7316 of data written. This will result in OOB reads when large OIDs are
7319 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7324 * Pointer arithmetic undefined behaviour
7326 Avoid some undefined pointer arithmetic
7328 A common idiom in the codebase is to check limits in the following manner:
7331 Where "p" points to some malloc'd data of SIZE bytes and
7334 "len" here could be from some externally supplied data (e.g. from a TLS
7337 The rules of C pointer arithmetic are such that "p + len" is only well
7338 defined where len <= SIZE. Therefore, the above idiom is actually
7339 undefined behaviour.
7341 For example this could cause problems if some malloc implementation
7342 provides an address for "p" such that "p + len" actually overflows for
7343 values of len that are too big and therefore p + len < limit.
7345 This issue was reported to OpenSSL by Guido Vranken
7350 * Constant time flag not preserved in DSA signing
7352 Operations in the DSA signing algorithm should run in constant time in
7353 order to avoid side channel attacks. A flaw in the OpenSSL DSA
7354 implementation means that a non-constant time codepath is followed for
7355 certain operations. This has been demonstrated through a cache-timing
7356 attack to be sufficient for an attacker to recover the private DSA key.
7358 This issue was reported by César Pereida (Aalto University), Billy Brumley
7359 (Tampere University of Technology), and Yuval Yarom (The University of
7360 Adelaide and NICTA).
7365 * DTLS buffered message DoS
7367 In a DTLS connection where handshake messages are delivered out-of-order
7368 those messages that OpenSSL is not yet ready to process will be buffered
7369 for later use. Under certain circumstances, a flaw in the logic means that
7370 those messages do not get removed from the buffer even though the handshake
7371 has been completed. An attacker could force up to approx. 15 messages to
7372 remain in the buffer when they are no longer required. These messages will
7373 be cleared when the DTLS connection is closed. The default maximum size for
7374 a message is 100k. Therefore, the attacker could force an additional 1500k
7375 to be consumed per connection. By opening many simultaneous connections an
7376 attacker could cause a DoS attack through memory exhaustion.
7378 This issue was reported to OpenSSL by Quan Luo.
7383 * DTLS replay protection DoS
7385 A flaw in the DTLS replay attack protection mechanism means that records
7386 that arrive for future epochs update the replay protection "window" before
7387 the MAC for the record has been validated. This could be exploited by an
7388 attacker by sending a record for the next epoch (which does not have to
7389 decrypt or have a valid MAC), with a very large sequence number. This means
7390 that all subsequent legitimate packets are dropped causing a denial of
7391 service for a specific DTLS connection.
7393 This issue was reported to OpenSSL by the OCAP audit team.
7398 * Certificate message OOB reads
7400 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7401 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7402 theoretical DoS risk but this has not been observed in practice on common
7405 The messages affected are client certificate, client certificate request
7406 and server certificate. As a result the attack can only be performed
7407 against a client or a server which enables client authentication.
7409 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7414 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7416 * Prevent padding oracle in AES-NI CBC MAC check
7418 A MITM attacker can use a padding oracle attack to decrypt traffic
7419 when the connection uses an AES CBC cipher and the server support
7422 This issue was introduced as part of the fix for Lucky 13 padding
7423 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7424 constant time by making sure that always the same bytes are read and
7425 compared against either the MAC or padding bytes. But it no longer
7426 checked that there was enough data to have both the MAC and padding
7429 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7434 * Fix EVP_EncodeUpdate overflow
7436 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7437 Base64 encoding of binary data. If an attacker is able to supply very large
7438 amounts of input data then a length check can overflow resulting in a heap
7441 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7442 the `PEM_write_bio*` family of functions. These are mainly used within the
7443 OpenSSL command line applications, so any application which processes data
7444 from an untrusted source and outputs it as a PEM file should be considered
7445 vulnerable to this issue. User applications that call these APIs directly
7446 with large amounts of untrusted data may also be vulnerable.
7448 This issue was reported by Guido Vranken.
7453 * Fix EVP_EncryptUpdate overflow
7455 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7456 is able to supply very large amounts of input data after a previous call to
7457 EVP_EncryptUpdate() with a partial block then a length check can overflow
7458 resulting in a heap corruption. Following an analysis of all OpenSSL
7459 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7460 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7461 the first called function after an EVP_EncryptInit(), and therefore that
7462 specific call must be safe. The second form is where the length passed to
7463 EVP_EncryptUpdate() can be seen from the code to be some small value and
7464 therefore there is no possibility of an overflow. Since all instances are
7465 one of these two forms, it is believed that there can be no overflows in
7466 internal code due to this problem. It should be noted that
7467 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7468 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7469 of these calls have also been analysed too and it is believed there are no
7470 instances in internal usage where an overflow could occur.
7472 This issue was reported by Guido Vranken.
7477 * Prevent ASN.1 BIO excessive memory allocation
7479 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7480 a short invalid encoding can casuse allocation of large amounts of memory
7481 potentially consuming excessive resources or exhausting memory.
7483 Any application parsing untrusted data through d2i BIO functions is
7484 affected. The memory based functions such as d2i_X509() are *not* affected.
7485 Since the memory based functions are used by the TLS library, TLS
7486 applications are not affected.
7488 This issue was reported by Brian Carpenter.
7495 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7496 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7497 in arbitrary stack data being returned in the buffer.
7499 This issue was reported by Guido Vranken.
7504 * Modify behavior of ALPN to invoke callback after SNI/servername
7505 callback, such that updates to the SSL_CTX affect ALPN.
7509 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7514 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7515 methods are enabled and ssl2 is disabled the methods return NULL.
7519 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7521 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7522 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7523 provide any "EXPORT" or "LOW" strength ciphers.
7527 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7528 is by default disabled at build-time. Builds that are not configured with
7529 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7530 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7531 will need to explicitly call either of:
7533 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7535 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7537 as appropriate. Even if either of those is used, or the application
7538 explicitly uses the version-specific SSLv2_method() or its client and
7539 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7540 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7541 ciphers, and SSLv2 56-bit DES are no longer available.
7546 * Fix a double-free in DSA code
7548 A double free bug was discovered when OpenSSL parses malformed DSA private
7549 keys and could lead to a DoS attack or memory corruption for applications
7550 that receive DSA private keys from untrusted sources. This scenario is
7553 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7559 * Disable SRP fake user seed to address a server memory leak.
7561 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7563 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7564 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7565 was changed to ignore the "fake user" SRP seed, even if the seed
7568 Users should use SRP_VBASE_get1_by_user instead. Note that in
7569 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7570 also that even though configuring the SRP seed attempts to hide
7571 invalid usernames by continuing the handshake with fake
7572 credentials, this behaviour is not constant time and no strong
7573 guarantees are made that the handshake is indistinguishable from
7574 that of a valid user.
7579 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7581 In the BN_hex2bn function the number of hex digits is calculated using an
7582 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7583 large values of `i` this can result in `bn_expand` not allocating any
7584 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7585 field as NULL leading to a subsequent NULL ptr deref. For very large values
7586 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7587 In this case memory is allocated to the internal BIGNUM data field, but it
7588 is insufficiently sized leading to heap corruption. A similar issue exists
7589 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7590 is ever called by user applications with very large untrusted hex/dec data.
7591 This is anticipated to be a rare occurrence.
7593 All OpenSSL internal usage of these functions use data that is not expected
7594 to be untrusted, e.g. config file data or application command line
7595 arguments. If user developed applications generate config file data based
7596 on untrusted data then it is possible that this could also lead to security
7597 consequences. This is also anticipated to be rare.
7599 This issue was reported to OpenSSL by Guido Vranken.
7604 * Fix memory issues in `BIO_*printf` functions
7606 The internal `fmtstr` function used in processing a "%s" format string in
7607 the `BIO_*printf` functions could overflow while calculating the length of a
7608 string and cause an OOB read when printing very long strings.
7610 Additionally the internal `doapr_outch` function can attempt to write to an
7611 OOB memory location (at an offset from the NULL pointer) in the event of a
7612 memory allocation failure. In 1.0.2 and below this could be caused where
7613 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7614 could be in processing a very long "%s" format string. Memory leaks can
7617 The first issue may mask the second issue dependent on compiler behaviour.
7618 These problems could enable attacks where large amounts of untrusted data
7619 is passed to the `BIO_*printf` functions. If applications use these functions
7620 in this way then they could be vulnerable. OpenSSL itself uses these
7621 functions when printing out human-readable dumps of ASN.1 data. Therefore
7622 applications that print this data could be vulnerable if the data is from
7623 untrusted sources. OpenSSL command line applications could also be
7624 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7625 as command line arguments.
7627 Libssl is not considered directly vulnerable. Additionally certificates etc
7628 received via remote connections via libssl are also unlikely to be able to
7629 trigger these issues because of message size limits enforced within libssl.
7631 This issue was reported to OpenSSL Guido Vranken.
7636 * Side channel attack on modular exponentiation
7638 A side-channel attack was found which makes use of cache-bank conflicts on
7639 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7640 of RSA keys. The ability to exploit this issue is limited as it relies on
7641 an attacker who has control of code in a thread running on the same
7642 hyper-threaded core as the victim thread which is performing decryptions.
7644 This issue was reported to OpenSSL by Yuval Yarom, The University of
7645 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7646 Nadia Heninger, University of Pennsylvania with more information at
7647 <http://cachebleed.info>.
7652 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7653 if no keysize is specified with default_bits. This fixes an
7654 omission in an earlier change that changed all RSA/DSA key generation
7655 commands to use 2048 bits by default.
7659 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7661 * Protection for DH small subgroup attacks
7663 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7664 switched on by default and cannot be disabled. This could have some
7669 * SSLv2 doesn't block disabled ciphers
7671 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7672 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7673 been disabled, provided that the SSLv2 protocol was not also disabled via
7676 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7677 and Sebastian Schinzel.
7682 * Reject DH handshakes with parameters shorter than 1024 bits.
7686 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7688 * Certificate verify crash with missing PSS parameter
7690 The signature verification routines will crash with a NULL pointer
7691 dereference if presented with an ASN.1 signature using the RSA PSS
7692 algorithm and absent mask generation function parameter. Since these
7693 routines are used to verify certificate signature algorithms this can be
7694 used to crash any certificate verification operation and exploited in a
7695 DoS attack. Any application which performs certificate verification is
7696 vulnerable including OpenSSL clients and servers which enable client
7699 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7704 * X509_ATTRIBUTE memory leak
7706 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7707 memory. This structure is used by the PKCS#7 and CMS routines so any
7708 application which reads PKCS#7 or CMS data from untrusted sources is
7709 affected. SSL/TLS is not affected.
7711 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7717 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7718 This changes the decoding behaviour for some invalid messages,
7719 though the change is mostly in the more lenient direction, and
7720 legacy behaviour is preserved as much as possible.
7724 * In DSA_generate_parameters_ex, if the provided seed is too short,
7725 use a random seed, as already documented.
7727 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7729 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7731 * Alternate chains certificate forgery
7733 During certificate verification, OpenSSL will attempt to find an
7734 alternative certificate chain if the first attempt to build such a chain
7735 fails. An error in the implementation of this logic can mean that an
7736 attacker could cause certain checks on untrusted certificates to be
7737 bypassed, such as the CA flag, enabling them to use a valid leaf
7738 certificate to act as a CA and "issue" an invalid certificate.
7740 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7746 * Race condition handling PSK identify hint
7748 If PSK identity hints are received by a multi-threaded client then
7749 the values are wrongly updated in the parent SSL_CTX structure. This can
7750 result in a race condition potentially leading to a double free of the
7756 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7758 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7759 incompatibility in the handling of HMAC. The previous ABI has now been
7762 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7764 * Malformed ECParameters causes infinite loop
7766 When processing an ECParameters structure OpenSSL enters an infinite loop
7767 if the curve specified is over a specially malformed binary polynomial
7770 This can be used to perform denial of service against any
7771 system which processes public keys, certificate requests or
7772 certificates. This includes TLS clients and TLS servers with
7773 client authentication enabled.
7775 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7780 * Exploitable out-of-bounds read in X509_cmp_time
7782 X509_cmp_time does not properly check the length of the ASN1_TIME
7783 string and can read a few bytes out of bounds. In addition,
7784 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7787 An attacker can use this to craft malformed certificates and CRLs of
7788 various sizes and potentially cause a segmentation fault, resulting in
7789 a DoS on applications that verify certificates or CRLs. TLS clients
7790 that verify CRLs are affected. TLS clients and servers with client
7791 authentication enabled may be affected if they use custom verification
7794 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7795 independently by Hanno Böck.
7800 * PKCS7 crash with missing EnvelopedContent
7802 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7803 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7804 with missing content and trigger a NULL pointer dereference on parsing.
7806 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7807 structures from untrusted sources are affected. OpenSSL clients and
7808 servers are not affected.
7810 This issue was reported to OpenSSL by Michal Zalewski (Google).
7815 * CMS verify infinite loop with unknown hash function
7817 When verifying a signedData message the CMS code can enter an infinite loop
7818 if presented with an unknown hash function OID. This can be used to perform
7819 denial of service against any system which verifies signedData messages using
7821 This issue was reported to OpenSSL by Johannes Bauer.
7826 * Race condition handling NewSessionTicket
7828 If a NewSessionTicket is received by a multi-threaded client when attempting to
7829 reuse a previous ticket then a race condition can occur potentially leading to
7830 a double free of the ticket data.
7835 * Reject DH handshakes with parameters shorter than 768 bits.
7837 *Kurt Roeckx and Emilia Kasper*
7839 * dhparam: generate 2048-bit parameters by default.
7841 *Kurt Roeckx and Emilia Kasper*
7843 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7845 * Segmentation fault in ASN1_TYPE_cmp fix
7847 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7848 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7849 certificate signature algorithm consistency this can be used to crash any
7850 certificate verification operation and exploited in a DoS attack. Any
7851 application which performs certificate verification is vulnerable including
7852 OpenSSL clients and servers which enable client authentication.
7857 * ASN.1 structure reuse memory corruption fix
7859 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7860 memory corruption via an invalid write. Such reuse is and has been
7861 strongly discouraged and is believed to be rare.
7863 Applications that parse structures containing CHOICE or ANY DEFINED BY
7864 components may be affected. Certificate parsing (d2i_X509 and related
7865 functions) are however not affected. OpenSSL clients and servers are
7871 * PKCS7 NULL pointer dereferences fix
7873 The PKCS#7 parsing code does not handle missing outer ContentInfo
7874 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7875 missing content and trigger a NULL pointer dereference on parsing.
7877 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7878 otherwise parse PKCS#7 structures from untrusted sources are
7879 affected. OpenSSL clients and servers are not affected.
7881 This issue was reported to OpenSSL by Michal Zalewski (Google).
7886 * DoS via reachable assert in SSLv2 servers fix
7888 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7889 servers that both support SSLv2 and enable export cipher suites by sending
7890 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7892 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7893 (OpenSSL development team).
7898 * Use After Free following d2i_ECPrivatekey error fix
7900 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7901 could cause a use after free condition. This, in turn, could cause a double
7902 free in several private key parsing functions (such as d2i_PrivateKey
7903 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7904 for applications that receive EC private keys from untrusted
7905 sources. This scenario is considered rare.
7907 This issue was discovered by the BoringSSL project and fixed in their
7913 * X509_to_X509_REQ NULL pointer deref fix
7915 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7916 the certificate key is invalid. This function is rarely used in practice.
7918 This issue was discovered by Brian Carpenter.
7923 * Removed the export ciphers from the DEFAULT ciphers
7927 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7929 * Build fixes for the Windows and OpenVMS platforms
7931 *Matt Caswell and Richard Levitte*
7933 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7935 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7936 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7937 dereference. This could lead to a Denial Of Service attack. Thanks to
7938 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7943 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7944 dtls1_buffer_record function under certain conditions. In particular this
7945 could occur if an attacker sent repeated DTLS records with the same
7946 sequence number but for the next epoch. The memory leak could be exploited
7947 by an attacker in a Denial of Service attack through memory exhaustion.
7948 Thanks to Chris Mueller for reporting this issue.
7953 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7954 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7955 method would be set to NULL which could later result in a NULL pointer
7956 dereference. Thanks to Frank Schmirler for reporting this issue.
7961 * Abort handshake if server key exchange message is omitted for ephemeral
7964 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7965 reporting this issue.
7970 * Remove non-export ephemeral RSA code on client and server. This code
7971 violated the TLS standard by allowing the use of temporary RSA keys in
7972 non-export ciphersuites and could be used by a server to effectively
7973 downgrade the RSA key length used to a value smaller than the server
7974 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7975 INRIA or reporting this issue.
7980 * Fixed issue where DH client certificates are accepted without verification.
7981 An OpenSSL server will accept a DH certificate for client authentication
7982 without the certificate verify message. This effectively allows a client to
7983 authenticate without the use of a private key. This only affects servers
7984 which trust a client certificate authority which issues certificates
7985 containing DH keys: these are extremely rare and hardly ever encountered.
7986 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7992 * Ensure that the session ID context of an SSL is updated when its
7993 SSL_CTX is updated via SSL_set_SSL_CTX.
7995 The session ID context is typically set from the parent SSL_CTX,
7996 and can vary with the CTX.
8000 * Fix various certificate fingerprint issues.
8002 By using non-DER or invalid encodings outside the signed portion of a
8003 certificate the fingerprint can be changed without breaking the signature.
8004 Although no details of the signed portion of the certificate can be changed
8005 this can cause problems with some applications: e.g. those using the
8006 certificate fingerprint for blacklists.
8008 1. Reject signatures with non zero unused bits.
8010 If the BIT STRING containing the signature has non zero unused bits reject
8011 the signature. All current signature algorithms require zero unused bits.
8013 2. Check certificate algorithm consistency.
8015 Check the AlgorithmIdentifier inside TBS matches the one in the
8016 certificate signature. NB: this will result in signature failure
8017 errors for some broken certificates.
8019 Thanks to Konrad Kraszewski from Google for reporting this issue.
8021 3. Check DSA/ECDSA signatures use DER.
8023 Re-encode DSA/ECDSA signatures and compare with the original received
8024 signature. Return an error if there is a mismatch.
8026 This will reject various cases including garbage after signature
8027 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8028 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8029 (negative or with leading zeroes).
8031 Further analysis was conducted and fixes were developed by Stephen Henson
8032 of the OpenSSL core team.
8038 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8039 results on some platforms, including x86_64. This bug occurs at random
8040 with a very low probability, and is not known to be exploitable in any
8041 way, though its exact impact is difficult to determine. Thanks to Pieter
8042 Wuille (Blockstream) who reported this issue and also suggested an initial
8043 fix. Further analysis was conducted by the OpenSSL development team and
8044 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8045 the OpenSSL core team.
8050 * Do not resume sessions on the server if the negotiated protocol
8051 version does not match the session's version. Resuming with a different
8052 version, while not strictly forbidden by the RFC, is of questionable
8053 sanity and breaks all known clients.
8055 *David Benjamin, Emilia Käsper*
8057 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
8058 early CCS messages during renegotiation. (Note that because
8059 renegotiation is encrypted, this early CCS was not exploitable.)
8063 * Tighten client-side session ticket handling during renegotiation:
8064 ensure that the client only accepts a session ticket if the server sends
8065 the extension anew in the ServerHello. Previously, a TLS client would
8066 reuse the old extension state and thus accept a session ticket if one was
8067 announced in the initial ServerHello.
8069 Similarly, ensure that the client requires a session ticket if one
8070 was advertised in the ServerHello. Previously, a TLS client would
8071 ignore a missing NewSessionTicket message.
8075 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
8079 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
8080 sends a carefully crafted handshake message, to cause OpenSSL to fail
8081 to free up to 64k of memory causing a memory leak. This could be
8082 exploited in a Denial Of Service attack. This issue affects OpenSSL
8083 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
8084 whether SRTP is used or configured. Implementations of OpenSSL that
8085 have been compiled with OPENSSL_NO_SRTP defined are not affected.
8087 The fix was developed by the OpenSSL team.
8092 * Session Ticket Memory Leak.
8094 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8095 integrity of that ticket is first verified. In the event of a session
8096 ticket integrity check failing, OpenSSL will fail to free memory
8097 causing a memory leak. By sending a large number of invalid session
8098 tickets an attacker could exploit this issue in a Denial Of Service
8104 * Build option no-ssl3 is incomplete.
8106 When OpenSSL is configured with "no-ssl3" as a build option, servers
8107 could accept and complete an SSL 3.0 handshake, and clients could be
8108 configured to send them.
8111 *Akamai and the OpenSSL team*
8113 * Add support for TLS_FALLBACK_SCSV.
8114 Client applications doing fallback retries should call
8115 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8118 *Adam Langley, Bodo Moeller*
8120 * Add additional DigestInfo checks.
8122 Re-encode DigestInto in DER and check against the original when
8123 verifying RSA signature: this will reject any improperly encoded
8124 DigestInfo structures.
8126 Note: this is a precautionary measure and no attacks are currently known.
8130 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
8132 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
8133 SRP code can be overrun an internal buffer. Add sanity check that
8134 g, A, B < N to SRP code.
8136 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
8137 Group for discovering this issue.
8142 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
8143 TLS 1.0 instead of higher protocol versions when the ClientHello message
8144 is badly fragmented. This allows a man-in-the-middle attacker to force a
8145 downgrade to TLS 1.0 even if both the server and the client support a
8146 higher protocol version, by modifying the client's TLS records.
8148 Thanks to David Benjamin and Adam Langley (Google) for discovering and
8149 researching this issue.
8154 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8155 to a denial of service attack. A malicious server can crash the client
8156 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8157 ciphersuite and sending carefully crafted handshake messages.
8159 Thanks to Felix Gröbert (Google) for discovering and researching this
8165 * By sending carefully crafted DTLS packets an attacker could cause openssl
8166 to leak memory. This can be exploited through a Denial of Service attack.
8167 Thanks to Adam Langley for discovering and researching this issue.
8172 * An attacker can force openssl to consume large amounts of memory whilst
8173 processing DTLS handshake messages. This can be exploited through a
8174 Denial of Service attack.
8175 Thanks to Adam Langley for discovering and researching this issue.
8180 * An attacker can force an error condition which causes openssl to crash
8181 whilst processing DTLS packets due to memory being freed twice. This
8182 can be exploited through a Denial of Service attack.
8183 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8189 * If a multithreaded client connects to a malicious server using a resumed
8190 session and the server sends an ec point format extension it could write
8191 up to 255 bytes to freed memory.
8193 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8199 * A malicious server can crash an OpenSSL client with a null pointer
8200 dereference (read) by specifying an SRP ciphersuite even though it was not
8201 properly negotiated with the client. This can be exploited through a
8202 Denial of Service attack.
8204 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
8205 discovering and researching this issue.
8210 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8211 X509_name_oneline, X509_name_print_ex et al. to leak some information
8212 from the stack. Applications may be affected if they echo pretty printing
8213 output to the attacker.
8215 Thanks to Ivan Fratric (Google) for discovering this issue.
8218 *Emilia Käsper, and Steve Henson*
8220 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8221 for corner cases. (Certain input points at infinity could lead to
8222 bogus results, with non-infinity inputs mapped to infinity too.)
8226 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
8228 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8229 handshake can force the use of weak keying material in OpenSSL
8230 SSL/TLS clients and servers.
8232 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8233 researching this issue. ([CVE-2014-0224])
8235 *KIKUCHI Masashi, Steve Henson*
8237 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8238 OpenSSL DTLS client the code can be made to recurse eventually crashing
8241 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8244 *Imre Rad, Steve Henson*
8246 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8247 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8248 client or server. This is potentially exploitable to run arbitrary
8249 code on a vulnerable client or server.
8251 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8253 *Jüri Aedla, Steve Henson*
8255 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8256 are subject to a denial of service attack.
8258 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8259 this issue. ([CVE-2014-3470])
8261 *Felix Gröbert, Ivan Fratric, Steve Henson*
8263 * Harmonize version and its documentation. -f flag is used to display
8266 *mancha <mancha1@zoho.com>*
8268 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8269 in i2d_ECPrivateKey.
8271 *mancha <mancha1@zoho.com>*
8273 * Fix some double frees. These are not thought to be exploitable.
8275 *mancha <mancha1@zoho.com>*
8277 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
8279 * A missing bounds check in the handling of the TLS heartbeat extension
8280 can be used to reveal up to 64k of memory to a connected client or
8283 Thanks for Neel Mehta of Google Security for discovering this bug and to
8284 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
8285 preparing the fix ([CVE-2014-0160])
8287 *Adam Langley, Bodo Moeller*
8289 * Fix for the attack described in the paper "Recovering OpenSSL
8290 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8291 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8292 <http://eprint.iacr.org/2014/140>
8294 Thanks to Yuval Yarom and Naomi Benger for discovering this
8295 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8297 *Yuval Yarom and Naomi Benger*
8299 * TLS pad extension: draft-agl-tls-padding-03
8301 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
8302 TLS client Hello record length value would otherwise be > 255 and
8303 less that 512 pad with a dummy extension containing zeroes so it
8304 is at least 512 bytes long.
8306 *Adam Langley, Steve Henson*
8308 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
8310 * Fix for TLS record tampering bug. A carefully crafted invalid
8311 handshake could crash OpenSSL with a NULL pointer exception.
8312 Thanks to Anton Johansson for reporting this issues.
8315 * Keep original DTLS digest and encryption contexts in retransmission
8316 structures so we can use the previous session parameters if they need
8317 to be resent. ([CVE-2013-6450])
8321 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8322 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8323 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8324 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8325 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8326 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8328 *Rob Stradling, Adam Langley*
8330 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
8332 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
8333 supporting platforms or when small records were transferred.
8335 *Andy Polyakov, Steve Henson*
8337 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
8339 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8341 This addresses the flaw in CBC record processing discovered by
8342 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8343 at: <http://www.isg.rhul.ac.uk/tls/>
8345 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8346 Security Group at Royal Holloway, University of London
8347 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8348 Emilia Käsper for the initial patch.
8351 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8353 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
8354 ciphersuites which can be exploited in a denial of service attack.
8355 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
8356 and detecting this bug and to Wolfgang Ettlinger
8357 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
8362 * Return an error when checking OCSP signatures when key is NULL.
8363 This fixes a DoS attack. ([CVE-2013-0166])
8367 * Make openssl verify return errors.
8369 *Chris Palmer <palmer@google.com> and Ben Laurie*
8371 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8372 the right response is stapled. Also change SSL_get_certificate()
8373 so it returns the certificate actually sent.
8374 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8376 *Rob Stradling <rob.stradling@comodo.com>*
8378 * Fix possible deadlock when decoding public keys.
8382 * Don't use TLS 1.0 record version number in initial client hello
8387 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
8389 * Sanity check record length before skipping explicit IV in TLS
8390 1.2, 1.1 and DTLS to fix DoS attack.
8392 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8393 fuzzing as a service testing platform.
8398 * Initialise tkeylen properly when encrypting CMS messages.
8399 Thanks to Solar Designer of Openwall for reporting this issue.
8403 * In FIPS mode don't try to use composite ciphers as they are not
8408 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8410 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8411 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8412 mean any application compiled against OpenSSL 1.0.0 headers setting
8413 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8414 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8415 0x10000000L Any application which was previously compiled against
8416 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8417 will need to be recompiled as a result. Letting be results in
8418 inability to disable specifically TLS 1.1 and in client context,
8419 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8423 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8424 disable just protocol X, but all protocols above X *if* there are
8425 protocols *below* X still enabled. In more practical terms it means
8426 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8427 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8428 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8433 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8435 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8436 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8437 in CRYPTO_realloc_clean.
8439 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8440 issue and to Adam Langley <agl@chromium.org> for fixing it.
8443 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8445 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8449 * Workarounds for some broken servers that "hang" if a client hello
8450 record length exceeds 255 bytes.
8452 1. Do not use record version number > TLS 1.0 in initial client
8453 hello: some (but not all) hanging servers will now work.
8454 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8455 the number of ciphers sent in the client hello. This should be
8456 set to an even number, such as 50, for example by passing:
8457 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8458 Most broken servers should now work.
8459 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8460 TLS 1.2 client support entirely.
8464 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8468 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8470 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8471 STRING form instead of a DigestInfo.
8475 * The format used for MDC2 RSA signatures is inconsistent between EVP
8476 and the RSA_sign/RSA_verify functions. This was made more apparent when
8477 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8478 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8479 the correct format in RSA_verify so both forms transparently work.
8483 * Some servers which support TLS 1.0 can choke if we initially indicate
8484 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8485 encrypted premaster secret. As a workaround use the maximum permitted
8486 client version in client hello, this should keep such servers happy
8487 and still work with previous versions of OpenSSL.
8491 * Add support for TLS/DTLS heartbeats.
8493 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8495 * Add support for SCTP.
8497 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8499 * Improved PRNG seeding for VOS.
8501 *Paul Green <Paul.Green@stratus.com>*
8503 * Extensive assembler packs updates, most notably:
8505 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8506 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8507 - x86_64: bit-sliced AES implementation;
8508 - ARM: NEON support, contemporary platforms optimizations;
8509 - s390x: z196 support;
8510 - `*`: GHASH and GF(2^m) multiplication implementations;
8514 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8515 (removal of unnecessary code)
8517 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8519 * Add TLS key material exporter from RFC 5705.
8523 * Add DTLS-SRTP negotiation from RFC 5764.
8527 * Add Next Protocol Negotiation,
8528 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8529 disabled with a no-npn flag to config or Configure. Code donated
8532 *Adam Langley <agl@google.com> and Ben Laurie*
8534 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8535 NIST-P256, NIST-P521, with constant-time single point multiplication on
8536 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8537 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8538 Code made available under Apache License version 2.0.
8540 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8541 line to include this in your build of OpenSSL, and run "make depend" (or
8542 "make update"). This enables the following EC_METHODs:
8544 EC_GFp_nistp224_method()
8545 EC_GFp_nistp256_method()
8546 EC_GFp_nistp521_method()
8548 EC_GROUP_new_by_curve_name() will automatically use these (while
8549 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8552 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8554 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8555 all platforms. Move ssize_t definition from e_os.h to the public
8556 header file e_os2.h as it now appears in public header file cms.h
8560 * New -sigopt option to the ca, req and x509 utilities. Additional
8561 signature parameters can be passed using this option and in
8566 * Add RSA PSS signing function. This will generate and set the
8567 appropriate AlgorithmIdentifiers for PSS based on those in the
8568 corresponding EVP_MD_CTX structure. No application support yet.
8572 * Support for companion algorithm specific ASN1 signing routines.
8573 New function ASN1_item_sign_ctx() signs a pre-initialised
8574 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8575 the appropriate parameters.
8579 * Add new algorithm specific ASN1 verification initialisation function
8580 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8581 handling will be the same no matter what EVP_PKEY_METHOD is used.
8582 Add a PSS handler to support verification of PSS signatures: checked
8583 against a number of sample certificates.
8587 * Add signature printing for PSS. Add PSS OIDs.
8589 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8591 * Add algorithm specific signature printing. An individual ASN1 method
8592 can now print out signatures instead of the standard hex dump.
8594 More complex signatures (e.g. PSS) can print out more meaningful
8595 information. Include DSA version that prints out the signature
8600 * Password based recipient info support for CMS library: implementing
8605 * Split password based encryption into PBES2 and PBKDF2 functions. This
8606 neatly separates the code into cipher and PBE sections and is required
8607 for some algorithms that split PBES2 into separate pieces (such as
8608 password based CMS).
8612 * Session-handling fixes:
8613 - Fix handling of connections that are resuming with a session ID,
8614 but also support Session Tickets.
8615 - Fix a bug that suppressed issuing of a new ticket if the client
8616 presented a ticket with an expired session.
8617 - Try to set the ticket lifetime hint to something reasonable.
8618 - Make tickets shorter by excluding irrelevant information.
8619 - On the client side, don't ignore renewed tickets.
8621 *Adam Langley, Bodo Moeller (Google)*
8623 * Fix PSK session representation.
8627 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8629 This work was sponsored by Intel.
8633 * Add GCM support to TLS library. Some custom code is needed to split
8634 the IV between the fixed (from PRF) and explicit (from TLS record)
8635 portions. This adds all GCM ciphersuites supported by RFC5288 and
8636 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8637 add a special AESGCM string for GCM only.
8641 * Expand range of ctrls for AES GCM. Permit setting invocation
8642 field on decrypt and retrieval of invocation field only on encrypt.
8646 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8647 As required by RFC5289 these ciphersuites cannot be used if for
8648 versions of TLS earlier than 1.2.
8652 * For FIPS capable OpenSSL interpret a NULL default public key method
8653 as unset and return the appropriate default but do *not* set the default.
8654 This means we can return the appropriate method in applications that
8655 switch between FIPS and non-FIPS modes.
8659 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8660 ENGINE is used then we cannot handle that in the FIPS module so we
8661 keep original code iff non-FIPS operations are allowed.
8665 * Add -attime option to openssl utilities.
8667 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8669 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8673 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8674 FIPS EC methods unconditionally for now.
8678 * New build option no-ec2m to disable characteristic 2 code.
8682 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8683 all cases can be covered as some introduce binary incompatibilities.
8687 * Redirect RSA operations to FIPS module including keygen,
8688 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8692 * Add similar low-level API blocking to ciphers.
8696 * low-level digest APIs are not approved in FIPS mode: any attempt
8697 to use these will cause a fatal error. Applications that *really* want
8698 to use them can use the `private_*` version instead.
8702 * Redirect cipher operations to FIPS module for FIPS builds.
8706 * Redirect digest operations to FIPS module for FIPS builds.
8710 * Update build system to add "fips" flag which will link in fipscanister.o
8711 for static and shared library builds embedding a signature if needed.
8715 * Output TLS supported curves in preference order instead of numerical
8716 order. This is currently hardcoded for the highest order curves first.
8717 This should be configurable so applications can judge speed vs strength.
8721 * Add TLS v1.2 server support for client authentication.
8725 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8730 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8731 FIPS modules versions.
8735 * Add TLS v1.2 client side support for client authentication. Keep cache
8736 of handshake records longer as we don't know the hash algorithm to use
8737 until after the certificate request message is received.
8741 * Initial TLS v1.2 client support. Add a default signature algorithms
8742 extension including all the algorithms we support. Parse new signature
8743 format in client key exchange. Relax some ECC signing restrictions for
8744 TLS v1.2 as indicated in RFC5246.
8748 * Add server support for TLS v1.2 signature algorithms extension. Switch
8749 to new signature format when needed using client digest preference.
8750 All server ciphersuites should now work correctly in TLS v1.2. No client
8751 support yet and no support for client certificates.
8755 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8756 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8757 ciphersuites. At present only RSA key exchange ciphersuites work with
8758 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8759 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8760 and version checking.
8764 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8765 with this defined it will not be affected by any changes to ssl internal
8766 structures. Add several utility functions to allow openssl application
8767 to work with OPENSSL_NO_SSL_INTERN defined.
8771 * A long standing patch to add support for SRP from EdelWeb (Peter
8772 Sylvester and Christophe Renou) was integrated.
8773 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8774 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8777 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8781 * Permit abbreviated handshakes when renegotiating using the function
8782 SSL_renegotiate_abbreviated().
8784 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8786 * Add call to ENGINE_register_all_complete() to
8787 ENGINE_load_builtin_engines(), so some implementations get used
8788 automatically instead of needing explicit application support.
8792 * Add support for TLS key exporter as described in RFC5705.
8794 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8796 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8797 a few changes are required:
8799 Add SSL_OP_NO_TLSv1_1 flag.
8800 Add TLSv1_1 methods.
8801 Update version checking logic to handle version 1.1.
8802 Add explicit IV handling (ported from DTLS code).
8803 Add command line options to s_client/s_server.
8810 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8812 * X509_ATTRIBUTE memory leak
8814 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8815 memory. This structure is used by the PKCS#7 and CMS routines so any
8816 application which reads PKCS#7 or CMS data from untrusted sources is
8817 affected. SSL/TLS is not affected.
8819 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8825 * Race condition handling PSK identify hint
8827 If PSK identity hints are received by a multi-threaded client then
8828 the values are wrongly updated in the parent SSL_CTX structure. This can
8829 result in a race condition potentially leading to a double free of the
8835 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8837 * Malformed ECParameters causes infinite loop
8839 When processing an ECParameters structure OpenSSL enters an infinite loop
8840 if the curve specified is over a specially malformed binary polynomial
8843 This can be used to perform denial of service against any
8844 system which processes public keys, certificate requests or
8845 certificates. This includes TLS clients and TLS servers with
8846 client authentication enabled.
8848 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8853 * Exploitable out-of-bounds read in X509_cmp_time
8855 X509_cmp_time does not properly check the length of the ASN1_TIME
8856 string and can read a few bytes out of bounds. In addition,
8857 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8860 An attacker can use this to craft malformed certificates and CRLs of
8861 various sizes and potentially cause a segmentation fault, resulting in
8862 a DoS on applications that verify certificates or CRLs. TLS clients
8863 that verify CRLs are affected. TLS clients and servers with client
8864 authentication enabled may be affected if they use custom verification
8867 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8868 independently by Hanno Böck.
8873 * PKCS7 crash with missing EnvelopedContent
8875 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8876 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8877 with missing content and trigger a NULL pointer dereference on parsing.
8879 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8880 structures from untrusted sources are affected. OpenSSL clients and
8881 servers are not affected.
8883 This issue was reported to OpenSSL by Michal Zalewski (Google).
8888 * CMS verify infinite loop with unknown hash function
8890 When verifying a signedData message the CMS code can enter an infinite loop
8891 if presented with an unknown hash function OID. This can be used to perform
8892 denial of service against any system which verifies signedData messages using
8894 This issue was reported to OpenSSL by Johannes Bauer.
8899 * Race condition handling NewSessionTicket
8901 If a NewSessionTicket is received by a multi-threaded client when attempting to
8902 reuse a previous ticket then a race condition can occur potentially leading to
8903 a double free of the ticket data.
8908 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8910 * Segmentation fault in ASN1_TYPE_cmp fix
8912 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8913 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8914 certificate signature algorithm consistency this can be used to crash any
8915 certificate verification operation and exploited in a DoS attack. Any
8916 application which performs certificate verification is vulnerable including
8917 OpenSSL clients and servers which enable client authentication.
8922 * ASN.1 structure reuse memory corruption fix
8924 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8925 memory corruption via an invalid write. Such reuse is and has been
8926 strongly discouraged and is believed to be rare.
8928 Applications that parse structures containing CHOICE or ANY DEFINED BY
8929 components may be affected. Certificate parsing (d2i_X509 and related
8930 functions) are however not affected. OpenSSL clients and servers are
8936 * PKCS7 NULL pointer dereferences fix
8938 The PKCS#7 parsing code does not handle missing outer ContentInfo
8939 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8940 missing content and trigger a NULL pointer dereference on parsing.
8942 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8943 otherwise parse PKCS#7 structures from untrusted sources are
8944 affected. OpenSSL clients and servers are not affected.
8946 This issue was reported to OpenSSL by Michal Zalewski (Google).
8951 * DoS via reachable assert in SSLv2 servers fix
8953 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8954 servers that both support SSLv2 and enable export cipher suites by sending
8955 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8957 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8958 (OpenSSL development team).
8963 * Use After Free following d2i_ECPrivatekey error fix
8965 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8966 could cause a use after free condition. This, in turn, could cause a double
8967 free in several private key parsing functions (such as d2i_PrivateKey
8968 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8969 for applications that receive EC private keys from untrusted
8970 sources. This scenario is considered rare.
8972 This issue was discovered by the BoringSSL project and fixed in their
8978 * X509_to_X509_REQ NULL pointer deref fix
8980 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8981 the certificate key is invalid. This function is rarely used in practice.
8983 This issue was discovered by Brian Carpenter.
8988 * Removed the export ciphers from the DEFAULT ciphers
8992 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8994 * Build fixes for the Windows and OpenVMS platforms
8996 *Matt Caswell and Richard Levitte*
8998 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
9000 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
9001 message can cause a segmentation fault in OpenSSL due to a NULL pointer
9002 dereference. This could lead to a Denial Of Service attack. Thanks to
9003 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
9008 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
9009 dtls1_buffer_record function under certain conditions. In particular this
9010 could occur if an attacker sent repeated DTLS records with the same
9011 sequence number but for the next epoch. The memory leak could be exploited
9012 by an attacker in a Denial of Service attack through memory exhaustion.
9013 Thanks to Chris Mueller for reporting this issue.
9018 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
9019 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
9020 method would be set to NULL which could later result in a NULL pointer
9021 dereference. Thanks to Frank Schmirler for reporting this issue.
9026 * Abort handshake if server key exchange message is omitted for ephemeral
9029 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
9030 reporting this issue.
9035 * Remove non-export ephemeral RSA code on client and server. This code
9036 violated the TLS standard by allowing the use of temporary RSA keys in
9037 non-export ciphersuites and could be used by a server to effectively
9038 downgrade the RSA key length used to a value smaller than the server
9039 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
9040 INRIA or reporting this issue.
9045 * Fixed issue where DH client certificates are accepted without verification.
9046 An OpenSSL server will accept a DH certificate for client authentication
9047 without the certificate verify message. This effectively allows a client to
9048 authenticate without the use of a private key. This only affects servers
9049 which trust a client certificate authority which issues certificates
9050 containing DH keys: these are extremely rare and hardly ever encountered.
9051 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
9057 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
9058 results on some platforms, including x86_64. This bug occurs at random
9059 with a very low probability, and is not known to be exploitable in any
9060 way, though its exact impact is difficult to determine. Thanks to Pieter
9061 Wuille (Blockstream) who reported this issue and also suggested an initial
9062 fix. Further analysis was conducted by the OpenSSL development team and
9063 Adam Langley of Google. The final fix was developed by Andy Polyakov of
9064 the OpenSSL core team.
9069 * Fix various certificate fingerprint issues.
9071 By using non-DER or invalid encodings outside the signed portion of a
9072 certificate the fingerprint can be changed without breaking the signature.
9073 Although no details of the signed portion of the certificate can be changed
9074 this can cause problems with some applications: e.g. those using the
9075 certificate fingerprint for blacklists.
9077 1. Reject signatures with non zero unused bits.
9079 If the BIT STRING containing the signature has non zero unused bits reject
9080 the signature. All current signature algorithms require zero unused bits.
9082 2. Check certificate algorithm consistency.
9084 Check the AlgorithmIdentifier inside TBS matches the one in the
9085 certificate signature. NB: this will result in signature failure
9086 errors for some broken certificates.
9088 Thanks to Konrad Kraszewski from Google for reporting this issue.
9090 3. Check DSA/ECDSA signatures use DER.
9092 Re-encode DSA/ECDSA signatures and compare with the original received
9093 signature. Return an error if there is a mismatch.
9095 This will reject various cases including garbage after signature
9096 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
9097 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
9098 (negative or with leading zeroes).
9100 Further analysis was conducted and fixes were developed by Stephen Henson
9101 of the OpenSSL core team.
9107 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
9109 * Session Ticket Memory Leak.
9111 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
9112 integrity of that ticket is first verified. In the event of a session
9113 ticket integrity check failing, OpenSSL will fail to free memory
9114 causing a memory leak. By sending a large number of invalid session
9115 tickets an attacker could exploit this issue in a Denial Of Service
9121 * Build option no-ssl3 is incomplete.
9123 When OpenSSL is configured with "no-ssl3" as a build option, servers
9124 could accept and complete an SSL 3.0 handshake, and clients could be
9125 configured to send them.
9128 *Akamai and the OpenSSL team*
9130 * Add support for TLS_FALLBACK_SCSV.
9131 Client applications doing fallback retries should call
9132 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
9135 *Adam Langley, Bodo Moeller*
9137 * Add additional DigestInfo checks.
9139 Re-encode DigestInto in DER and check against the original when
9140 verifying RSA signature: this will reject any improperly encoded
9141 DigestInfo structures.
9143 Note: this is a precautionary measure and no attacks are currently known.
9147 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
9149 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
9150 to a denial of service attack. A malicious server can crash the client
9151 with a null pointer dereference (read) by specifying an anonymous (EC)DH
9152 ciphersuite and sending carefully crafted handshake messages.
9154 Thanks to Felix Gröbert (Google) for discovering and researching this
9160 * By sending carefully crafted DTLS packets an attacker could cause openssl
9161 to leak memory. This can be exploited through a Denial of Service attack.
9162 Thanks to Adam Langley for discovering and researching this issue.
9167 * An attacker can force openssl to consume large amounts of memory whilst
9168 processing DTLS handshake messages. This can be exploited through a
9169 Denial of Service attack.
9170 Thanks to Adam Langley for discovering and researching this issue.
9175 * An attacker can force an error condition which causes openssl to crash
9176 whilst processing DTLS packets due to memory being freed twice. This
9177 can be exploited through a Denial of Service attack.
9178 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
9184 * If a multithreaded client connects to a malicious server using a resumed
9185 session and the server sends an ec point format extension it could write
9186 up to 255 bytes to freed memory.
9188 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
9194 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
9195 X509_name_oneline, X509_name_print_ex et al. to leak some information
9196 from the stack. Applications may be affected if they echo pretty printing
9197 output to the attacker.
9199 Thanks to Ivan Fratric (Google) for discovering this issue.
9202 *Emilia Käsper, and Steve Henson*
9204 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
9205 for corner cases. (Certain input points at infinity could lead to
9206 bogus results, with non-infinity inputs mapped to infinity too.)
9210 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
9212 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
9213 handshake can force the use of weak keying material in OpenSSL
9214 SSL/TLS clients and servers.
9216 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
9217 researching this issue. ([CVE-2014-0224])
9219 *KIKUCHI Masashi, Steve Henson*
9221 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
9222 OpenSSL DTLS client the code can be made to recurse eventually crashing
9225 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
9228 *Imre Rad, Steve Henson*
9230 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
9231 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
9232 client or server. This is potentially exploitable to run arbitrary
9233 code on a vulnerable client or server.
9235 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
9237 *Jüri Aedla, Steve Henson*
9239 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
9240 are subject to a denial of service attack.
9242 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
9243 this issue. ([CVE-2014-3470])
9245 *Felix Gröbert, Ivan Fratric, Steve Henson*
9247 * Harmonize version and its documentation. -f flag is used to display
9250 *mancha <mancha1@zoho.com>*
9252 * Fix eckey_priv_encode so it immediately returns an error upon a failure
9253 in i2d_ECPrivateKey.
9255 *mancha <mancha1@zoho.com>*
9257 * Fix some double frees. These are not thought to be exploitable.
9259 *mancha <mancha1@zoho.com>*
9261 * Fix for the attack described in the paper "Recovering OpenSSL
9262 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
9263 by Yuval Yarom and Naomi Benger. Details can be obtained from:
9264 <http://eprint.iacr.org/2014/140>
9266 Thanks to Yuval Yarom and Naomi Benger for discovering this
9267 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
9269 *Yuval Yarom and Naomi Benger*
9271 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
9273 * Keep original DTLS digest and encryption contexts in retransmission
9274 structures so we can use the previous session parameters if they need
9275 to be resent. ([CVE-2013-6450])
9279 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
9280 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
9281 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
9282 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
9283 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
9284 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
9286 *Rob Stradling, Adam Langley*
9288 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
9290 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
9292 This addresses the flaw in CBC record processing discovered by
9293 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
9294 at: <http://www.isg.rhul.ac.uk/tls/>
9296 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9297 Security Group at Royal Holloway, University of London
9298 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
9299 Emilia Käsper for the initial patch.
9302 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
9304 * Return an error when checking OCSP signatures when key is NULL.
9305 This fixes a DoS attack. ([CVE-2013-0166])
9309 * Call OCSP Stapling callback after ciphersuite has been chosen, so
9310 the right response is stapled. Also change SSL_get_certificate()
9311 so it returns the certificate actually sent.
9312 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
9313 (This is a backport)
9315 *Rob Stradling <rob.stradling@comodo.com>*
9317 * Fix possible deadlock when decoding public keys.
9321 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
9323 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
9326 * Sanity check record length before skipping explicit IV in DTLS
9329 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
9330 fuzzing as a service testing platform.
9335 * Initialise tkeylen properly when encrypting CMS messages.
9336 Thanks to Solar Designer of Openwall for reporting this issue.
9340 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
9342 * Check for potentially exploitable overflows in asn1_d2i_read_bio
9343 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
9344 in CRYPTO_realloc_clean.
9346 Thanks to Tavis Ormandy, Google Security Team, for discovering this
9347 issue and to Adam Langley <agl@chromium.org> for fixing it.
9350 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
9352 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
9354 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
9355 in CMS and PKCS7 code. When RSA decryption fails use a random key for
9356 content decryption and always return the same error. Note: this attack
9357 needs on average 2^20 messages so it only affects automated senders. The
9358 old behaviour can be re-enabled in the CMS code by setting the
9359 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
9360 an MMA defence is not necessary.
9361 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
9362 this issue. ([CVE-2012-0884])
9366 * Fix CVE-2011-4619: make sure we really are receiving a
9367 client hello before rejecting multiple SGC restarts. Thanks to
9368 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
9372 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
9374 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
9375 Thanks to Antonio Martin, Enterprise Secure Access Research and
9376 Development, Cisco Systems, Inc. for discovering this bug and
9377 preparing a fix. ([CVE-2012-0050])
9381 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
9383 * Nadhem Alfardan and Kenny Paterson have discovered an extension
9384 of the Vaudenay padding oracle attack on CBC mode encryption
9385 which enables an efficient plaintext recovery attack against
9386 the OpenSSL implementation of DTLS. Their attack exploits timing
9387 differences arising during decryption processing. A research
9388 paper describing this attack can be found at:
9389 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
9390 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9391 Security Group at Royal Holloway, University of London
9392 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9393 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9394 for preparing the fix. ([CVE-2011-4108])
9396 *Robin Seggelmann, Michael Tuexen*
9398 * Clear bytes used for block padding of SSL 3.0 records.
9401 *Adam Langley (Google)*
9403 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9404 Kadianakis <desnacked@gmail.com> for discovering this issue and
9405 Adam Langley for preparing the fix. ([CVE-2011-4619])
9407 *Adam Langley (Google)*
9409 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9411 *Andrey Kulikov <amdeich@gmail.com>*
9413 * Prevent malformed RFC3779 data triggering an assertion failure.
9414 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9415 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9417 *Rob Austein <sra@hactrn.net>*
9419 * Improved PRNG seeding for VOS.
9421 *Paul Green <Paul.Green@stratus.com>*
9423 * Fix ssl_ciph.c set-up race.
9425 *Adam Langley (Google)*
9427 * Fix spurious failures in ecdsatest.c.
9429 *Emilia Käsper (Google)*
9431 * Fix the BIO_f_buffer() implementation (which was mixing different
9432 interpretations of the `..._len` fields).
9434 *Adam Langley (Google)*
9436 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9437 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9438 threads won't reuse the same blinding coefficients.
9440 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9441 lock to call BN_BLINDING_invert_ex, and avoids one use of
9442 BN_BLINDING_update for each BN_BLINDING structure (previously,
9443 the last update always remained unused).
9445 *Emilia Käsper (Google)*
9447 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9449 *Bob Buckholz (Google)*
9451 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9453 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9454 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9456 *Kaspar Brand <ossl@velox.ch>*
9458 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9459 for multi-threaded use of ECDH. ([CVE-2011-3210])
9461 *Adam Langley (Google)*
9463 * Fix x509_name_ex_d2i memory leak on bad inputs.
9467 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9468 signature public key algorithm by using OID xref utilities instead.
9469 Before this you could only use some ECC ciphersuites with SHA1 only.
9473 * Add protection against ECDSA timing attacks as mentioned in the paper
9474 by Billy Bob Brumley and Nicola Tuveri, see:
9475 <http://eprint.iacr.org/2011/232.pdf>
9477 *Billy Bob Brumley and Nicola Tuveri*
9479 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9481 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9483 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9485 * Fix bug in string printing code: if *any* escaping is enabled we must
9486 escape the escape character (backslash) or the resulting string is
9491 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9493 * Disable code workaround for ancient and obsolete Netscape browsers
9494 and servers: an attacker can use it in a ciphersuite downgrade attack.
9495 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9499 * Fixed J-PAKE implementation error, originally discovered by
9500 Sebastien Martini, further info and confirmation from Stefan
9501 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9505 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9507 * Fix extension code to avoid race conditions which can result in a buffer
9508 overrun vulnerability: resumed sessions must not be modified as they can
9509 be shared by multiple threads. CVE-2010-3864
9513 * Fix WIN32 build system to correctly link an ENGINE directory into
9518 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9520 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9523 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9525 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9527 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9528 context. The operation can be customised via the ctrl mechanism in
9529 case ENGINEs want to include additional functionality.
9533 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9537 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9538 output hashes compatible with older versions of OpenSSL.
9540 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9542 * Fix compression algorithm handling: if resuming a session use the
9543 compression algorithm of the resumed session instead of determining
9544 it from client hello again. Don't allow server to change algorithm.
9548 * Add load_crls() function to commands tidying load_certs() too. Add option
9549 to verify utility to allow additional CRLs to be included.
9553 * Update OCSP request code to permit adding custom headers to the request:
9554 some responders need this.
9558 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9561 *Julia Lawall <julia@diku.dk>*
9563 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9564 needlessly dereferenced structures, used obsolete functions and
9565 didn't handle all updated verify codes correctly.
9569 * Disable MD2 in the default configuration.
9573 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9574 indicate the initial BIO being pushed or popped. This makes it possible
9575 to determine whether the BIO is the one explicitly called or as a result
9576 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9577 it handles reference counts correctly and doesn't zero out the I/O bio
9578 when it is not being explicitly popped. WARNING: applications which
9579 included workarounds for the old buggy behaviour will need to be modified
9580 or they could free up already freed BIOs.
9584 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9585 renaming to all platforms (within the 0.9.8 branch, this was
9586 done conditionally on Netware platforms to avoid a name clash).
9588 *Guenter <lists@gknw.net>*
9590 * Add ECDHE and PSK support to DTLS.
9592 *Michael Tuexen <tuexen@fh-muenster.de>*
9594 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9599 * Add "missing" function EVP_MD_flags() (without this the only way to
9600 retrieve a digest flags is by accessing the structure directly. Update
9601 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9602 or cipher is registered as in the "from" argument. Print out all
9603 registered digests in the dgst usage message instead of manually
9604 attempting to work them out.
9608 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9609 this allows the use of compression and extensions. Change default cipher
9610 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9611 by default unless an application cipher string requests it.
9615 * Alter match criteria in PKCS12_parse(). It used to try to use local
9616 key ids to find matching certificates and keys but some PKCS#12 files
9617 don't follow the (somewhat unwritten) rules and this strategy fails.
9618 Now just gather all certificates together and the first private key
9619 then look for the first certificate that matches the key.
9623 * Support use of registered digest and cipher names for dgst and cipher
9624 commands instead of having to add each one as a special case. So now
9631 openssl dgst -sha256 foo
9633 and this works for ENGINE based algorithms too.
9637 * Update Gost ENGINE to support parameter files.
9639 *Victor B. Wagner <vitus@cryptocom.ru>*
9641 * Support GeneralizedTime in ca utility.
9643 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9645 * Enhance the hash format used for certificate directory links. The new
9646 form uses the canonical encoding (meaning equivalent names will work
9647 even if they aren't identical) and uses SHA1 instead of MD5. This form
9648 is incompatible with the older format and as a result c_rehash should
9649 be used to rebuild symbolic links.
9653 * Make PKCS#8 the default write format for private keys, replacing the
9654 traditional format. This form is standardised, more secure and doesn't
9655 include an implicit MD5 dependency.
9659 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9660 committed to OpenSSL should pass this lot as a minimum.
9664 * Add session ticket override functionality for use by EAP-FAST.
9666 *Jouni Malinen <j@w1.fi>*
9668 * Modify HMAC functions to return a value. Since these can be implemented
9669 in an ENGINE errors can occur.
9673 * Type-checked OBJ_bsearch_ex.
9677 * Type-checked OBJ_bsearch. Also some constification necessitated
9678 by type-checking. Still to come: TXT_DB, bsearch(?),
9679 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9684 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9685 seconds to a tm structure directly, instead of going through OS
9686 specific date routines. This avoids any issues with OS routines such
9687 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9688 and X509_time_adj_ex() to cover the extended range. The existing
9689 X509_time_adj() is still usable and will no longer have any date issues.
9693 * Delta CRL support. New use deltas option which will attempt to locate
9694 and search any appropriate delta CRLs available.
9696 This work was sponsored by Google.
9700 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9701 code and add additional score elements. Validate alternate CRL paths
9702 as part of the CRL checking and indicate a new error "CRL path validation
9703 error" in this case. Applications wanting additional details can use
9704 the verify callback and check the new "parent" field. If this is not
9705 NULL CRL path validation is taking place. Existing applications won't
9706 see this because it requires extended CRL support which is off by
9709 This work was sponsored by Google.
9713 * Support for freshest CRL extension.
9715 This work was sponsored by Google.
9719 * Initial indirect CRL support. Currently only supported in the CRLs
9720 passed directly and not via lookup. Process certificate issuer
9721 CRL entry extension and lookup CRL entries by bother issuer name
9722 and serial number. Check and process CRL issuer entry in IDP extension.
9724 This work was sponsored by Google.
9728 * Add support for distinct certificate and CRL paths. The CRL issuer
9729 certificate is validated separately in this case. Only enabled if
9730 an extended CRL support flag is set: this flag will enable additional
9731 CRL functionality in future.
9733 This work was sponsored by Google.
9737 * Add support for policy mappings extension.
9739 This work was sponsored by Google.
9743 * Fixes to pathlength constraint, self issued certificate handling,
9744 policy processing to align with RFC3280 and PKITS tests.
9746 This work was sponsored by Google.
9750 * Support for name constraints certificate extension. DN, email, DNS
9751 and URI types are currently supported.
9753 This work was sponsored by Google.
9757 * To cater for systems that provide a pointer-based thread ID rather
9758 than numeric, deprecate the current numeric thread ID mechanism and
9759 replace it with a structure and associated callback type. This
9760 mechanism allows a numeric "hash" to be extracted from a thread ID in
9761 either case, and on platforms where pointers are larger than 'long',
9762 mixing is done to help ensure the numeric 'hash' is usable even if it
9763 can't be guaranteed unique. The default mechanism is to use "&errno"
9764 as a pointer-based thread ID to distinguish between threads.
9766 Applications that want to provide their own thread IDs should now use
9767 CRYPTO_THREADID_set_callback() to register a callback that will call
9768 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9770 Note that ERR_remove_state() is now deprecated, because it is tied
9771 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9772 to free the current thread's error state should be replaced by
9773 ERR_remove_thread_state(NULL).
9775 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9776 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9777 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9778 application was previously providing a numeric thread callback that
9779 was inappropriate for distinguishing threads, then uniqueness might
9780 have been obtained with &errno that happened immediately in the
9781 intermediate development versions of OpenSSL; this is no longer the
9782 case, the numeric thread callback will now override the automatic use
9785 *Geoff Thorpe, with help from Bodo Moeller*
9787 * Initial support for different CRL issuing certificates. This covers a
9788 simple case where the self issued certificates in the chain exist and
9789 the real CRL issuer is higher in the existing chain.
9791 This work was sponsored by Google.
9795 * Removed effectively defunct crypto/store from the build.
9799 * Revamp of STACK to provide stronger type-checking. Still to come:
9800 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9801 ASN1_STRING, CONF_VALUE.
9805 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9806 RAM on SSL connections. This option can save about 34k per idle SSL.
9810 * Revamp of LHASH to provide stronger type-checking. Still to come:
9811 STACK, TXT_DB, bsearch, qsort.
9815 * Initial support for Cryptographic Message Syntax (aka CMS) based
9816 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9817 support for data, signedData, compressedData, digestedData and
9818 encryptedData, envelopedData types included. Scripts to check against
9819 RFC4134 examples draft and interop and consistency checks of many
9820 content types and variants.
9824 * Add options to enc utility to support use of zlib compression BIO.
9828 * Extend mk1mf to support importing of options and assembly language
9829 files from Configure script, currently only included in VC-WIN32.
9830 The assembly language rules can now optionally generate the source
9831 files from the associated perl scripts.
9835 * Implement remaining functionality needed to support GOST ciphersuites.
9836 Interop testing has been performed using CryptoPro implementations.
9838 *Victor B. Wagner <vitus@cryptocom.ru>*
9840 * s390x assembler pack.
9844 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9849 * Implement Opaque PRF Input TLS extension as specified in
9850 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9851 official specification yet and no extension type assignment by
9852 IANA exists, this extension (for now) will have to be explicitly
9853 enabled when building OpenSSL by providing the extension number
9854 to use. For example, specify an option
9856 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9858 to the "config" or "Configure" script to enable the extension,
9859 assuming extension number 0x9527 (which is a completely arbitrary
9860 and unofficial assignment based on the MD5 hash of the Internet
9861 Draft). Note that by doing so, you potentially lose
9862 interoperability with other TLS implementations since these might
9863 be using the same extension number for other purposes.
9865 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9866 opaque PRF input value to use in the handshake. This will create
9867 an internal copy of the length-'len' string at 'src', and will
9868 return non-zero for success.
9870 To get more control and flexibility, provide a callback function
9873 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9874 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9878 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9881 Callback function 'cb' will be called in handshakes, and is
9882 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9883 Argument 'arg' is for application purposes (the value as given to
9884 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9885 be provided to the callback function). The callback function
9886 has to return non-zero to report success: usually 1 to use opaque
9887 PRF input just if possible, or 2 to enforce use of the opaque PRF
9888 input. In the latter case, the library will abort the handshake
9889 if opaque PRF input is not successfully negotiated.
9891 Arguments 'peerinput' and 'len' given to the callback function
9892 will always be NULL and 0 in the case of a client. A server will
9893 see the client's opaque PRF input through these variables if
9894 available (NULL and 0 otherwise). Note that if the server
9895 provides an opaque PRF input, the length must be the same as the
9896 length of the client's opaque PRF input.
9898 Note that the callback function will only be called when creating
9899 a new session (session resumption can resume whatever was
9900 previously negotiated), and will not be called in SSL 2.0
9901 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9902 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9903 for applications that need to enforce opaque PRF input.
9907 * Update ssl code to support digests other than SHA1+MD5 for handshake
9910 *Victor B. Wagner <vitus@cryptocom.ru>*
9912 * Add RFC4507 support to OpenSSL. This includes the corrections in
9913 RFC4507bis. The encrypted ticket format is an encrypted encoded
9914 SSL_SESSION structure, that way new session features are automatically
9917 If a client application caches session in an SSL_SESSION structure
9918 support is transparent because tickets are now stored in the encoded
9921 The SSL_CTX structure automatically generates keys for ticket
9922 protection in servers so again support should be possible
9923 with no application modification.
9925 If a client or server wishes to disable RFC4507 support then the option
9926 SSL_OP_NO_TICKET can be set.
9928 Add a TLS extension debugging callback to allow the contents of any client
9929 or server extensions to be examined.
9931 This work was sponsored by Google.
9935 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9936 OpenSSL should now compile cleanly on gcc 4.2
9938 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9940 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9941 support including streaming MAC support: this is required for GOST
9942 ciphersuite support.
9944 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9946 * Add option -stream to use PKCS#7 streaming in smime utility. New
9947 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9948 to output in BER and PEM format.
9952 * Experimental support for use of HMAC via EVP_PKEY interface. This
9953 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9954 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9955 ENGINE support for HMAC keys which are unextractable. New -mac and
9956 -macopt options to dgst utility.
9960 * New option -sigopt to dgst utility. Update dgst to use
9961 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9962 alternative signing parameters such as X9.31 or PSS in the dgst
9967 * Change ssl_cipher_apply_rule(), the internal function that does
9968 the work each time a ciphersuite string requests enabling
9969 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9970 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9971 the order of disabled ciphersuites such that those ciphersuites
9972 that most recently went from enabled to disabled not only stay
9973 in order with respect to each other, but also have higher priority
9974 than other disabled ciphersuites the next time ciphersuites are
9977 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9978 the same ciphersuites as with "HIGH" alone, but in a specific
9979 order where the PSK ciphersuites come first (since they are the
9980 most recently disabled ciphersuites when "HIGH" is parsed).
9982 Also, change ssl_create_cipher_list() (using this new
9983 functionality) such that between otherwise identical
9984 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9989 * Change ssl_create_cipher_list() so that it automatically
9990 arranges the ciphersuites in reasonable order before starting
9991 to process the rule string. Thus, the definition for "DEFAULT"
9992 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9993 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9994 This makes it much easier to arrive at a reasonable default order
9995 in applications for which anonymous ciphers are OK (meaning
9996 that you can't actually use DEFAULT).
9998 *Bodo Moeller; suggested by Victor Duchovni*
10000 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
10001 processing) into multiple integers instead of setting
10002 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
10003 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
10004 (These masks as well as the individual bit definitions are hidden
10005 away into the non-exported interface ssl/ssl_locl.h, so this
10006 change to the definition of the SSL_CIPHER structure shouldn't
10007 affect applications.) This give us more bits for each of these
10008 categories, so there is no longer a need to coagulate AES128 and
10009 AES256 into a single algorithm bit, and to coagulate Camellia128
10010 and Camellia256 into a single algorithm bit, which has led to all
10013 Thus, among other things, the kludge introduced in 0.9.7m and
10014 0.9.8e for masking out AES256 independently of AES128 or masking
10015 out Camellia256 independently of AES256 is not needed here in 0.9.9.
10017 With the change, we also introduce new ciphersuite aliases that
10018 so far were missing: "AES128", "AES256", "CAMELLIA128", and
10023 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
10024 Use the leftmost N bytes of the signature input if the input is
10025 larger than the prime q (with N being the size in bytes of q).
10029 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
10030 it yet and it is largely untested.
10034 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
10038 * Initial incomplete changes to avoid need for function casts in OpenSSL
10039 some compilers (gcc 4.2 and later) reject their use. Safestack is
10040 reimplemented. Update ASN1 to avoid use of legacy functions.
10044 * Win32/64 targets are linked with Winsock2.
10048 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
10049 to external functions. This can be used to increase CRL handling
10050 efficiency especially when CRLs are very large by (for example) storing
10051 the CRL revoked certificates in a database.
10055 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
10056 new CRLs added to a directory can be used. New command line option
10057 -verify_return_error to s_client and s_server. This causes real errors
10058 to be returned by the verify callback instead of carrying on no matter
10059 what. This reflects the way a "real world" verify callback would behave.
10063 * GOST engine, supporting several GOST algorithms and public key formats.
10064 Kindly donated by Cryptocom.
10068 * Partial support for Issuing Distribution Point CRL extension. CRLs
10069 partitioned by DP are handled but no indirect CRL or reason partitioning
10070 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
10071 selected via a scoring technique which handles IDP and AKID in CRLs.
10075 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
10076 will ultimately be used for all verify operations: this will remove the
10077 X509_STORE dependency on certificate verification and allow alternative
10078 lookup methods. X509_STORE based implementations of these two callbacks.
10082 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
10083 Modify get_crl() to find a valid (unexpired) CRL if possible.
10087 * New function X509_CRL_match() to check if two CRLs are identical. Normally
10088 this would be called X509_CRL_cmp() but that name is already used by
10089 a function that just compares CRL issuer names. Cache several CRL
10090 extensions in X509_CRL structure and cache CRLDP in X509.
10094 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
10095 this maps equivalent X509_NAME structures into a consistent structure.
10096 Name comparison can then be performed rapidly using memcmp().
10100 * Non-blocking OCSP request processing. Add -timeout option to ocsp
10105 * Allow digests to supply their own micalg string for S/MIME type using
10106 the ctrl EVP_MD_CTRL_MICALG.
10110 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
10111 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
10112 ctrl. It can then customise the structure before and/or after signing
10117 * New function OBJ_add_sigid() to allow application defined signature OIDs
10118 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
10119 to free up any added signature OIDs.
10123 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
10124 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
10125 digest and cipher tables. New options added to openssl utility:
10126 list-message-digest-algorithms and list-cipher-algorithms.
10130 * Change the array representation of binary polynomials: the list
10131 of degrees of non-zero coefficients is now terminated with -1.
10132 Previously it was terminated with 0, which was also part of the
10133 value; thus, the array representation was not applicable to
10134 polynomials where t^0 has coefficient zero. This change makes
10135 the array representation useful in a more general context.
10139 * Various modifications and fixes to SSL/TLS cipher string
10140 handling. For ECC, the code now distinguishes between fixed ECDH
10141 with RSA certificates on the one hand and with ECDSA certificates
10142 on the other hand, since these are separate ciphersuites. The
10143 unused code for Fortezza ciphersuites has been removed.
10145 For consistency with EDH, ephemeral ECDH is now called "EECDH"
10146 (not "ECDHE"). For consistency with the code for DH
10147 certificates, use of ECDH certificates is now considered ECDH
10148 authentication, not RSA or ECDSA authentication (the latter is
10149 merely the CA's signing algorithm and not actively used in the
10152 The temporary ciphersuite alias "ECCdraft" is no longer
10153 available, and ECC ciphersuites are no longer excluded from "ALL"
10154 and "DEFAULT". The following aliases now exist for RFC 4492
10155 ciphersuites, most of these by analogy with the DH case:
10157 kECDHr - ECDH cert, signed with RSA
10158 kECDHe - ECDH cert, signed with ECDSA
10159 kECDH - ECDH cert (signed with either RSA or ECDSA)
10160 kEECDH - ephemeral ECDH
10161 ECDH - ECDH cert or ephemeral ECDH
10164 aECDSA - ECDSA cert
10167 AECDH - anonymous ECDH
10168 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
10172 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
10173 Use correct micalg parameters depending on digest(s) in signed message.
10177 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
10178 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
10182 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
10183 an engine to register a method. Add ENGINE lookups for methods and
10184 functional reference processing.
10188 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
10189 `EVP_{Sign,Verify}*` which allow an application to customise the signature
10194 * New -resign option to smime utility. This adds one or more signers
10195 to an existing PKCS#7 signedData structure. Also -md option to use an
10196 alternative message digest algorithm for signing.
10200 * Tidy up PKCS#7 routines and add new functions to make it easier to
10201 create PKCS7 structures containing multiple signers. Update smime
10202 application to support multiple signers.
10206 * New -macalg option to pkcs12 utility to allow setting of an alternative
10211 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
10212 Reorganize PBE internals to lookup from a static table using NIDs,
10213 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
10214 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
10215 PRF which will be automatically used with PBES2.
10219 * Replace the algorithm specific calls to generate keys in "req" with the
10224 * Update PKCS#7 enveloped data routines to use new API. This is now
10225 supported by any public key method supporting the encrypt operation. A
10226 ctrl is added to allow the public key algorithm to examine or modify
10227 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
10232 * Add a ctrl to asn1 method to allow a public key algorithm to express
10233 a default digest type to use. In most cases this will be SHA1 but some
10234 algorithms (such as GOST) need to specify an alternative digest. The
10235 return value indicates how strong the preference is 1 means optional and
10236 2 is mandatory (that is it is the only supported type). Modify
10237 ASN1_item_sign() to accept a NULL digest argument to indicate it should
10238 use the default md. Update openssl utilities to use the default digest
10239 type for signing if it is not explicitly indicated.
10243 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
10244 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
10245 signing method from the key type. This effectively removes the link
10246 between digests and public key types.
10250 * Add an OID cross reference table and utility functions. Its purpose is to
10251 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
10252 rsaEncryption. This will allow some of the algorithm specific hackery
10253 needed to use the correct OID to be removed.
10257 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
10258 structures for PKCS7_sign(). They are now set up by the relevant public
10263 * Add provisional EC pkey method with support for ECDSA and ECDH.
10267 * Add support for key derivation (agreement) in the API, DH method and
10272 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
10273 public and private key formats. As a side effect these add additional
10274 command line functionality not previously available: DSA signatures can be
10275 generated and verified using pkeyutl and DH key support and generation in
10282 *Oliver Tappe <zooey@hirschkaefer.de>*
10284 * New make target "install_html_docs" installs HTML renditions of the
10287 *Oliver Tappe <zooey@hirschkaefer.de>*
10289 * New utility "genpkey" this is analogous to "genrsa" etc except it can
10290 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
10291 support key and parameter generation and add initial key generation
10292 functionality for RSA.
10296 * Add functions for main EVP_PKEY_method operations. The undocumented
10297 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
10298 `EVP_PKEY_{encrypt,decrypt}_old`.
10302 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
10303 key API, doesn't do much yet.
10307 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
10308 public key algorithms. New option to openssl utility:
10309 "list-public-key-algorithms" to print out info.
10313 * Implement the Supported Elliptic Curves Extension for
10314 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10318 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
10319 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
10323 * New utilities pkey and pkeyparam. These are similar to algorithm specific
10324 utilities such as rsa, dsa, dsaparam etc except they process any key
10329 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
10330 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
10331 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
10336 * Initial support for pluggable public key ASN1.
10337 De-spaghettify the public key ASN1 handling. Move public and private
10338 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
10339 algorithm specific handling to a single module within the relevant
10340 algorithm directory. Add functions to allow (near) opaque processing
10341 of public and private key structures.
10345 * Implement the Supported Point Formats Extension for
10346 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10350 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
10351 for the psk identity [hint] and the psk callback functions to the
10352 SSL_SESSION, SSL and SSL_CTX structure.
10355 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
10359 SSL_CTX_use_psk_identity_hint
10360 SSL_get_psk_identity_hint
10361 SSL_get_psk_identity
10362 SSL_use_psk_identity_hint
10364 *Mika Kousa and Pasi Eronen of Nokia Corporation*
10366 * Add RFC 3161 compliant time stamp request creation, response generation
10367 and response verification functionality.
10369 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
10371 * Add initial support for TLS extensions, specifically for the server_name
10372 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10373 have new members for a hostname. The SSL data structure has an
10374 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10375 stored in that context to allow for session resumption, even after the
10376 SSL has been switched to a new SSL_CTX in reaction to a client's
10377 server_name extension.
10379 New functions (subject to change):
10381 SSL_get_servername()
10382 SSL_get_servername_type()
10385 New CTRL codes and macros (subject to change):
10387 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10388 - SSL_CTX_set_tlsext_servername_callback()
10389 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10390 - SSL_CTX_set_tlsext_servername_arg()
10391 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10393 openssl s_client has a new '-servername ...' option.
10395 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10396 '-key2 ...', '-servername_fatal' (subject to change). This allows
10397 testing the HostName extension for a specific single hostname ('-cert'
10398 and '-key' remain fallbacks for handshakes without HostName
10399 negotiation). If the unrecognized_name alert has to be sent, this by
10400 default is a warning; it becomes fatal with the '-servername_fatal'
10403 *Peter Sylvester, Remy Allais, Christophe Renou*
10405 * Whirlpool hash implementation is added.
10409 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10410 bn(64,32). Because of instruction set limitations it doesn't have
10411 any negative impact on performance. This was done mostly in order
10412 to make it possible to share assembler modules, such as bn_mul_mont
10413 implementations, between 32- and 64-bit builds without hassle.
10417 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10418 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10423 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10424 dedicated Montgomery multiplication procedure, is introduced.
10425 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10426 "64-bit" performance on certain 32-bit targets.
10430 * New option SSL_OP_NO_COMP to disable use of compression selectively
10431 in SSL structures. New SSL ctrl to set maximum send fragment size.
10432 Save memory by setting the I/O buffer sizes dynamically instead of
10433 using the maximum available value.
10437 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10438 in addition to the text details.
10442 * Very, very preliminary EXPERIMENTAL support for printing of general
10443 ASN1 structures. This currently produces rather ugly output and doesn't
10444 handle several customised structures at all.
10448 * Integrated support for PVK file format and some related formats such
10449 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10450 these in the 'rsa' and 'dsa' utilities.
10454 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10458 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10459 place for the (very old) "NETSCAPE" format certificates which are now
10460 handled using new ASN1 code equivalents.
10464 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10465 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10466 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10470 * Modify CRL distribution points extension code to print out previously
10471 unsupported fields. Enhance extension setting code to allow setting of
10476 * Add print and set support for Issuing Distribution Point CRL extension.
10480 * Change 'Configure' script to enable Camellia by default.
10487 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10489 * When rejecting SSL/TLS records due to an incorrect version number, never
10490 update s->server with a new major version number. As of
10491 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10492 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10493 the previous behavior could result in a read attempt at NULL when
10494 receiving specific incorrect SSL/TLS records once record payload
10495 protection is active. ([CVE-2010-0740])
10497 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10499 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10500 could be crashed if the relevant tables were not present (e.g. chrooted).
10502 *Tomas Hoger <thoger@redhat.com>*
10504 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10506 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10508 *Martin Olsson, Neel Mehta*
10510 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10511 accommodate for stack sorting, always a write lock!).
10515 * On some versions of WIN32 Heap32Next is very slow. This can cause
10516 excessive delays in the RAND_poll(): over a minute. As a workaround
10517 include a time check in the inner Heap32Next loop too.
10521 * The code that handled flushing of data in SSL/TLS originally used the
10522 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10523 the problem outlined in PR#1949. The fix suggested there however can
10524 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10525 of Apache). So instead simplify the code to flush unconditionally.
10526 This should be fine since flushing with no data to flush is a no op.
10530 * Handle TLS versions 2.0 and later properly and correctly use the
10531 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10532 off ancient servers have a habit of sticking around for a while...
10536 * Modify compression code so it frees up structures without using the
10537 ex_data callbacks. This works around a problem where some applications
10538 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10539 restarting) then use compression (e.g. SSL with compression) later.
10540 This results in significant per-connection memory leaks and
10541 has caused some security issues including CVE-2008-1678 and
10546 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10547 change when encrypting or decrypting.
10551 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10552 connect and renegotiate with servers which do not support RI.
10553 Until RI is more widely deployed this option is enabled by default.
10557 * Add "missing" ssl ctrls to clear options and mode.
10561 * If client attempts to renegotiate and doesn't support RI respond with
10562 a no_renegotiation alert as required by RFC5746. Some renegotiating
10563 TLS clients will continue a connection gracefully when they receive
10564 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10565 waiting for a server hello which it will never receive. Now we treat a
10566 received no_renegotiation alert as a fatal error. This is because
10567 applications requesting a renegotiation might well expect it to succeed
10568 and would have no code in place to handle the server denying it so the
10569 only safe thing to do is to terminate the connection.
10573 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10574 peer supports secure renegotiation and 0 otherwise. Print out peer
10575 renegotiation support in s_client/s_server.
10579 * Replace the highly broken and deprecated SPKAC certification method with
10580 the updated NID creation version. This should correctly handle UTF8.
10584 * Implement RFC5746. Re-enable renegotiation but require the extension
10585 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10586 turns out to be a bad idea. It has been replaced by
10587 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10588 SSL_CTX_set_options(). This is really not recommended unless you
10589 know what you are doing.
10591 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10593 * Fixes to stateless session resumption handling. Use initial_ctx when
10594 issuing and attempting to decrypt tickets in case it has changed during
10595 servername handling. Use a non-zero length session ID when attempting
10596 stateless session resumption: this makes it possible to determine if
10597 a resumption has occurred immediately after receiving server hello
10598 (several places in OpenSSL subtly assume this) instead of later in
10603 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10604 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10605 fixes for a few places where the return code is not checked
10608 *Julia Lawall <julia@diku.dk>*
10610 * Add --strict-warnings option to Configure script to include devteam
10611 warnings in other configurations.
10615 * Add support for --libdir option and LIBDIR variable in makefiles. This
10616 makes it possible to install openssl libraries in locations which
10617 have names other than "lib", for example "/usr/lib64" which some
10620 *Steve Henson, based on patch from Jeremy Utley*
10622 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10623 X690 8.9.12 and can produce some misleading textual output of OIDs.
10625 *Steve Henson, reported by Dan Kaminsky*
10627 * Delete MD2 from algorithm tables. This follows the recommendation in
10628 several standards that it is not used in new applications due to
10629 several cryptographic weaknesses. For binary compatibility reasons
10630 the MD2 API is still compiled in by default.
10634 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10639 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10640 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10643 *Guenter <lists@gknw.net>*
10645 * Fix the server certificate chain building code to use X509_verify_cert(),
10646 it used to have an ad-hoc builder which was unable to cope with anything
10647 other than a simple chain.
10649 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10651 * Don't check self signed certificate signatures in X509_verify_cert()
10652 by default (a flag can override this): it just wastes time without
10653 adding any security. As a useful side effect self signed root CAs
10654 with non-FIPS digests are now usable in FIPS mode.
10658 * In dtls1_process_out_of_seq_message() the check if the current message
10659 is already buffered was missing. For every new message was memory
10660 allocated, allowing an attacker to perform an denial of service attack
10661 with sending out of seq handshake messages until there is no memory
10662 left. Additionally every future message was buffered, even if the
10663 sequence number made no sense and would be part of another handshake.
10664 So only messages with sequence numbers less than 10 in advance will be
10665 buffered. ([CVE-2009-1378])
10667 *Robin Seggelmann, discovered by Daniel Mentz*
10669 * Records are buffered if they arrive with a future epoch to be
10670 processed after finishing the corresponding handshake. There is
10671 currently no limitation to this buffer allowing an attacker to perform
10672 a DOS attack with sending records with future epochs until there is no
10673 memory left. This patch adds the pqueue_size() function to determine
10674 the size of a buffer and limits the record buffer to 100 entries.
10677 *Robin Seggelmann, discovered by Daniel Mentz*
10679 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10680 parent structure is freed. ([CVE-2009-1379])
10684 * Handle non-blocking I/O properly in SSL_shutdown() call.
10686 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10688 * Add `2.5.4.*` OIDs
10690 *Ilya O. <vrghost@gmail.com>*
10692 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10694 * Disable renegotiation completely - this fixes a severe security
10695 problem ([CVE-2009-3555]) at the cost of breaking all
10696 renegotiation. Renegotiation can be re-enabled by setting
10697 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10698 run-time. This is really not recommended unless you know what
10703 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10705 * Don't set val to NULL when freeing up structures, it is freed up by
10706 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10707 zeroing past the valid field. ([CVE-2009-0789])
10709 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10711 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10712 checked correctly. This would allow some invalid signed attributes to
10713 appear to verify correctly. ([CVE-2009-0591])
10715 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10717 * Reject UniversalString and BMPString types with invalid lengths. This
10718 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10719 a legal length. ([CVE-2009-0590])
10723 * Set S/MIME signing as the default purpose rather than setting it
10724 unconditionally. This allows applications to override it at the store
10729 * Permit restricted recursion of ASN1 strings. This is needed in practice
10730 to handle some structures.
10734 * Improve efficiency of mem_gets: don't search whole buffer each time
10737 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10739 * New -hex option for openssl rand.
10743 * Print out UTF8String and NumericString when parsing ASN1.
10747 * Support NumericString type for name components.
10751 * Allow CC in the environment to override the automatically chosen
10752 compiler. Note that nothing is done to ensure flags work with the
10757 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10759 * Properly check EVP_VerifyFinal() and similar return values
10762 *Ben Laurie, Bodo Moeller, Google Security Team*
10764 * Enable TLS extensions by default.
10768 * Allow the CHIL engine to be loaded, whether the application is
10769 multithreaded or not. (This does not release the developer from the
10770 obligation to set up the dynamic locking callbacks.)
10772 *Sander Temme <sander@temme.net>*
10774 * Use correct exit code if there is an error in dgst command.
10776 *Steve Henson; problem pointed out by Roland Dirlewanger*
10778 * Tweak Configure so that you need to say "experimental-jpake" to enable
10779 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10783 * Add experimental JPAKE support, including demo authentication in
10784 s_client and s_server.
10788 * Set the comparison function in v3_addr_canonize().
10790 *Rob Austein <sra@hactrn.net>*
10792 * Add support for XMPP STARTTLS in s_client.
10794 *Philip Paeps <philip@freebsd.org>*
10796 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10797 to ensure that even with this option, only ciphersuites in the
10798 server's preference list will be accepted. (Note that the option
10799 applies only when resuming a session, so the earlier behavior was
10800 just about the algorithm choice for symmetric cryptography.)
10804 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10806 * Fix NULL pointer dereference if a DTLS server received
10807 ChangeCipherSpec as first record ([CVE-2009-1386]).
10811 * Fix a state transition in s3_srvr.c and d1_srvr.c
10812 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10814 *Nagendra Modadugu*
10816 * The fix in 0.9.8c that supposedly got rid of unsafe
10817 double-checked locking was incomplete for RSA blinding,
10818 addressing just one layer of what turns out to have been
10819 doubly unsafe triple-checked locking.
10821 So now fix this for real by retiring the MONT_HELPER macro
10822 in crypto/rsa/rsa_eay.c.
10824 *Bodo Moeller; problem pointed out by Marius Schilder*
10826 * Various precautionary measures:
10828 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10830 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10831 (NB: This would require knowledge of the secret session ticket key
10832 to exploit, in which case you'd be SOL either way.)
10834 - Change bn_nist.c so that it will properly handle input BIGNUMs
10835 outside the expected range.
10837 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10840 *Neel Mehta, Bodo Moeller*
10842 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10843 the load fails. Useful for distros.
10845 *Ben Laurie and the FreeBSD team*
10847 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10851 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10855 * Expand ENGINE to support engine supplied SSL client certificate functions.
10857 This work was sponsored by Logica.
10861 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10862 keystores. Support for SSL/TLS client authentication too.
10863 Not compiled unless enable-capieng specified to Configure.
10865 This work was sponsored by Logica.
10869 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10870 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10871 attribute creation routines such as certificate requests and PKCS#12
10876 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10878 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10879 handshake which could lead to a client crash as found using the
10880 Codenomicon TLS test suite ([CVE-2008-1672])
10882 *Steve Henson, Mark Cox*
10884 * Fix double free in TLS server name extensions which could lead to
10885 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10889 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10891 Clear the error queue to ensure that error entries left from
10892 older function calls do not interfere with the correct operation.
10894 *Lutz Jaenicke, Erik de Castro Lopo*
10896 * Remove root CA certificates of commercial CAs:
10898 The OpenSSL project does not recommend any specific CA and does not
10899 have any policy with respect to including or excluding any CA.
10900 Therefore, it does not make any sense to ship an arbitrary selection
10901 of root CA certificates with the OpenSSL software.
10905 * RSA OAEP patches to fix two separate invalid memory reads.
10906 The first one involves inputs when 'lzero' is greater than
10907 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10908 before the beginning of from). The second one involves inputs where
10909 the 'db' section contains nothing but zeroes (there is a one-byte
10910 invalid read after the end of 'db').
10912 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10914 * Partial backport from 0.9.9-dev:
10916 Introduce bn_mul_mont (dedicated Montgomery multiplication
10917 procedure) as a candidate for BIGNUM assembler implementation.
10918 While 0.9.9-dev uses assembler for various architectures, only
10919 x86_64 is available by default here in the 0.9.8 branch, and
10920 32-bit x86 is available through a compile-time setting.
10922 To try the 32-bit x86 assembler implementation, use Configure
10923 option "enable-montasm" (which exists only for this backport).
10925 As "enable-montasm" for 32-bit x86 disclaims code stability
10926 anyway, in this constellation we activate additional code
10927 backported from 0.9.9-dev for further performance improvements,
10928 namely BN_from_montgomery_word. (To enable this otherwise,
10929 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10931 *Andy Polyakov (backport partially by Bodo Moeller)*
10933 * Add TLS session ticket callback. This allows an application to set
10934 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10935 values. This is useful for key rollover for example where several key
10936 sets may exist with different names.
10940 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10941 This was broken until now in 0.9.8 releases, such that the only way
10942 a registered ENGINE could be used (assuming it initialises
10943 successfully on the host) was to explicitly set it as the default
10944 for the relevant algorithms. This is in contradiction with 0.9.7
10945 behaviour and the documentation. With this fix, when an ENGINE is
10946 registered into a given algorithm's table of implementations, the
10947 'uptodate' flag is reset so that auto-discovery will be used next
10948 time a new context for that algorithm attempts to select an
10951 *Ian Lister (tweaked by Geoff Thorpe)*
10953 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10954 implementation in the following ways:
10956 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10959 Lack of BER streaming support means one pass streaming processing is
10960 only supported if data is detached: setting the streaming flag is
10961 ignored for embedded content.
10963 CMS support is disabled by default and must be explicitly enabled
10964 with the enable-cms configuration option.
10968 * Update the GMP engine glue to do direct copies between BIGNUM and
10969 mpz_t when openssl and GMP use the same limb size. Otherwise the
10970 existing "conversion via a text string export" trick is still used.
10972 *Paul Sheer <paulsheer@gmail.com>*
10974 * Zlib compression BIO. This is a filter BIO which compressed and
10975 uncompresses any data passed through it.
10979 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10980 RFC3394 compatible AES key wrapping.
10984 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10985 sets string data without copying. X509_ALGOR_set0() and
10986 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10987 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10988 from an X509_ATTRIBUTE structure optionally checking it occurs only
10989 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10994 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10995 to get the expected BN_FLG_CONSTTIME behavior.
10997 *Bodo Moeller (Google)*
11001 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
11002 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
11003 - added some more tests to do_tests.pl
11004 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
11005 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
11006 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
11007 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
11008 - various changes to netware.pl to enable gcc-cross builds on Win32
11010 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
11011 - various changes to fix missing prototype warnings
11012 - fixed x86nasm.pl to create correct asm files for NASM COFF output
11013 - added AES, WHIRLPOOL and CPUID assembler code to build files
11014 - added missing AES assembler make rules to mk1mf.pl
11015 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
11017 *Guenter Knauf <eflash@gmx.net>*
11019 * Implement certificate status request TLS extension defined in RFC3546.
11020 A client can set the appropriate parameters and receive the encoded
11021 OCSP response via a callback. A server can query the supplied parameters
11022 and set the encoded OCSP response in the callback. Add simplified examples
11023 to s_client and s_server.
11027 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
11029 * Fix various bugs:
11030 + Binary incompatibility of ssl_ctx_st structure
11031 + DTLS interoperation with non-compliant servers
11032 + Don't call get_session_cb() without proposed session
11033 + Fix ia64 assembler code
11035 *Andy Polyakov, Steve Henson*
11037 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
11039 * DTLS Handshake overhaul. There were longstanding issues with
11040 OpenSSL DTLS implementation, which were making it impossible for
11041 RFC 4347 compliant client to communicate with OpenSSL server.
11042 Unfortunately just fixing these incompatibilities would "cut off"
11043 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
11044 server keeps tolerating non RFC compliant syntax. The opposite is
11045 not true, 0.9.8f client can not communicate with earlier server.
11046 This update even addresses CVE-2007-4995.
11050 * Changes to avoid need for function casts in OpenSSL: some compilers
11051 (gcc 4.2 and later) reject their use.
11052 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
11055 * Add RFC4507 support to OpenSSL. This includes the corrections in
11056 RFC4507bis. The encrypted ticket format is an encrypted encoded
11057 SSL_SESSION structure, that way new session features are automatically
11060 If a client application caches session in an SSL_SESSION structure
11061 support is transparent because tickets are now stored in the encoded
11064 The SSL_CTX structure automatically generates keys for ticket
11065 protection in servers so again support should be possible
11066 with no application modification.
11068 If a client or server wishes to disable RFC4507 support then the option
11069 SSL_OP_NO_TICKET can be set.
11071 Add a TLS extension debugging callback to allow the contents of any client
11072 or server extensions to be examined.
11074 This work was sponsored by Google.
11078 * Add initial support for TLS extensions, specifically for the server_name
11079 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
11080 have new members for a hostname. The SSL data structure has an
11081 additional member `SSL_CTX *initial_ctx` so that new sessions can be
11082 stored in that context to allow for session resumption, even after the
11083 SSL has been switched to a new SSL_CTX in reaction to a client's
11084 server_name extension.
11086 New functions (subject to change):
11088 SSL_get_servername()
11089 SSL_get_servername_type()
11092 New CTRL codes and macros (subject to change):
11094 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
11095 - SSL_CTX_set_tlsext_servername_callback()
11096 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
11097 - SSL_CTX_set_tlsext_servername_arg()
11098 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
11100 openssl s_client has a new '-servername ...' option.
11102 openssl s_server has new options '-servername_host ...', '-cert2 ...',
11103 '-key2 ...', '-servername_fatal' (subject to change). This allows
11104 testing the HostName extension for a specific single hostname ('-cert'
11105 and '-key' remain fallbacks for handshakes without HostName
11106 negotiation). If the unrecognized_name alert has to be sent, this by
11107 default is a warning; it becomes fatal with the '-servername_fatal'
11110 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
11112 * Add AES and SSE2 assembly language support to VC++ build.
11116 * Mitigate attack on final subtraction in Montgomery reduction.
11120 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
11121 (which previously caused an internal error).
11125 * Squeeze another 10% out of IGE mode when in != out.
11129 * AES IGE mode speedup.
11131 *Dean Gaudet (Google)*
11133 * Add the Korean symmetric 128-bit cipher SEED (see
11134 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
11135 add SEED ciphersuites from RFC 4162:
11137 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
11138 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
11139 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
11140 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
11142 To minimize changes between patchlevels in the OpenSSL 0.9.8
11143 series, SEED remains excluded from compilation unless OpenSSL
11144 is configured with 'enable-seed'.
11146 *KISA, Bodo Moeller*
11148 * Mitigate branch prediction attacks, which can be practical if a
11149 single processor is shared, allowing a spy process to extract
11150 information. For detailed background information, see
11151 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
11152 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
11153 and Necessary Software Countermeasures"). The core of the change
11154 are new versions BN_div_no_branch() and
11155 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
11156 respectively, which are slower, but avoid the security-relevant
11157 conditional branches. These are automatically called by BN_div()
11158 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
11159 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
11160 remove a conditional branch.
11162 BN_FLG_CONSTTIME is the new name for the previous
11163 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
11164 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
11165 in the exponent causes BN_mod_exp_mont() to use the alternative
11166 implementation in BN_mod_exp_mont_consttime().) The old name
11167 remains as a deprecated alias.
11169 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
11170 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
11171 constant-time implementations for more than just exponentiation.
11172 Here too the old name is kept as a deprecated alias.
11174 BN_BLINDING_new() will now use BN_dup() for the modulus so that
11175 the BN_BLINDING structure gets an independent copy of the
11176 modulus. This means that the previous `BIGNUM *m` argument to
11177 BN_BLINDING_new() and to BN_BLINDING_create_param() now
11178 essentially becomes `const BIGNUM *m`, although we can't actually
11179 change this in the header file before 0.9.9. It allows
11180 RSA_setup_blinding() to use BN_with_flags() on the modulus to
11181 enable BN_FLG_CONSTTIME.
11183 *Matthew D Wood (Intel Corp)*
11185 * In the SSL/TLS server implementation, be strict about session ID
11186 context matching (which matters if an application uses a single
11187 external cache for different purposes). Previously,
11188 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
11189 set. This did ensure strict client verification, but meant that,
11190 with applications using a single external cache for quite
11191 different requirements, clients could circumvent ciphersuite
11192 restrictions for a given session ID context by starting a session
11193 in a different context.
11197 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11198 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11199 authentication-only ciphersuites.
11203 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
11204 not complete and could lead to a possible single byte overflow
11205 ([CVE-2007-5135]) [Ben Laurie]
11207 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
11209 * Since AES128 and AES256 (and similarly Camellia128 and
11210 Camellia256) share a single mask bit in the logic of
11211 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11212 kludge to work properly if AES128 is available and AES256 isn't
11213 (or if Camellia128 is available and Camellia256 isn't).
11217 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
11218 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
11219 When a point or a seed is encoded in a BIT STRING, we need to
11220 prevent the removal of trailing zero bits to get the proper DER
11221 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
11222 of a NamedBitList, for which trailing 0 bits need to be removed.)
11226 * Have SSL/TLS server implementation tolerate "mismatched" record
11227 protocol version while receiving ClientHello even if the
11228 ClientHello is fragmented. (The server can't insist on the
11229 particular protocol version it has chosen before the ServerHello
11230 message has informed the client about his choice.)
11234 * Add RFC 3779 support.
11236 *Rob Austein for ARIN, Ben Laurie*
11238 * Load error codes if they are not already present instead of using a
11239 static variable. This allows them to be cleanly unloaded and reloaded.
11240 Improve header file function name parsing.
11244 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
11245 or CAPABILITY handshake as required by RFCs.
11247 *Goetz Babin-Ebell*
11249 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
11251 * Introduce limits to prevent malicious keys being able to
11252 cause a denial of service. ([CVE-2006-2940])
11254 *Steve Henson, Bodo Moeller*
11256 * Fix ASN.1 parsing of certain invalid structures that can result
11257 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11259 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11260 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11262 * Fix SSL client code which could crash if connecting to a
11263 malicious SSLv2 server. ([CVE-2006-4343])
11265 *Tavis Ormandy and Will Drewry, Google Security Team*
11267 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
11268 match only those. Before that, "AES256-SHA" would be interpreted
11269 as a pattern and match "AES128-SHA" too (since AES128-SHA got
11270 the same strength classification in 0.9.7h) as we currently only
11271 have a single AES bit in the ciphersuite description bitmap.
11272 That change, however, also applied to ciphersuite strings such as
11273 "RC4-MD5" that intentionally matched multiple ciphersuites --
11274 namely, SSL 2.0 ciphersuites in addition to the more common ones
11275 from SSL 3.0/TLS 1.0.
11277 So we change the selection algorithm again: Naming an explicit
11278 ciphersuite selects this one ciphersuite, and any other similar
11279 ciphersuite (same bitmap) from *other* protocol versions.
11280 Thus, "RC4-MD5" again will properly select both the SSL 2.0
11281 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
11283 Since SSL 2.0 does not have any ciphersuites for which the
11284 128/256 bit distinction would be relevant, this works for now.
11285 The proper fix will be to use different bits for AES128 and
11286 AES256, which would have avoided the problems from the beginning;
11287 however, bits are scarce, so we can only do this in a new release
11288 (not just a patchlevel) when we can change the SSL_CIPHER
11289 definition to split the single 'unsigned long mask' bitmap into
11290 multiple values to extend the available space.
11294 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
11296 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11297 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11299 * Add AES IGE and biIGE modes.
11303 * Change the Unix randomness entropy gathering to use poll() when
11304 possible instead of select(), since the latter has some
11305 undesirable limitations.
11307 *Darryl Miles via Richard Levitte and Bodo Moeller*
11309 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
11310 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
11311 cannot be implicitly activated as part of, e.g., the "AES" alias.
11312 However, please upgrade to OpenSSL 0.9.9[-dev] for
11313 non-experimental use of the ECC ciphersuites to get TLS extension
11314 support, which is required for curve and point format negotiation
11315 to avoid potential handshake problems.
11319 * Disable rogue ciphersuites:
11321 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11322 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11323 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11325 The latter two were purportedly from
11326 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11329 Also deactivate the remaining ciphersuites from
11330 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11331 unofficial, and the ID has long expired.
11335 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11336 dual-core machines) and other potential thread-safety issues.
11340 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
11341 versions), which is now available for royalty-free use
11342 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
11343 Also, add Camellia TLS ciphersuites from RFC 4132.
11345 To minimize changes between patchlevels in the OpenSSL 0.9.8
11346 series, Camellia remains excluded from compilation unless OpenSSL
11347 is configured with 'enable-camellia'.
11351 * Disable the padding bug check when compression is in use. The padding
11352 bug check assumes the first packet is of even length, this is not
11353 necessarily true if compression is enabled and can result in false
11354 positives causing handshake failure. The actual bug test is ancient
11355 code so it is hoped that implementations will either have fixed it by
11356 now or any which still have the bug do not support compression.
11360 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
11362 * When applying a cipher rule check to see if string match is an explicit
11363 cipher suite and only match that one cipher suite if it is.
11367 * Link in manifests for VC++ if needed.
11369 *Austin Ziegler <halostatue@gmail.com>*
11371 * Update support for ECC-based TLS ciphersuites according to
11372 draft-ietf-tls-ecc-12.txt with proposed changes (but without
11373 TLS extensions, which are supported starting with the 0.9.9
11374 branch, not in the OpenSSL 0.9.8 branch).
11378 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
11379 opaque EVP_CIPHER_CTX handling.
11383 * Fixes and enhancements to zlib compression code. We now only use
11384 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
11385 to conform with the standards mentioned here:
11386 <http://www.zlib.net/DLL_FAQ.txt>
11387 Static zlib linking now works on Windows and the new --with-zlib-include
11388 --with-zlib-lib options to Configure can be used to supply the location
11389 of the headers and library. Gracefully handle case where zlib library
11394 * Several fixes and enhancements to the OID generation code. The old code
11395 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11396 handle numbers larger than ULONG_MAX, truncated printing and had a
11397 non standard OBJ_obj2txt() behaviour.
11401 * Add support for building of engines under engine/ as shared libraries
11402 under VC++ build system.
11406 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11407 Hopefully, we will not see any false combination of paths any more.
11411 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11413 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11414 (part of SSL_OP_ALL). This option used to disable the
11415 countermeasure against man-in-the-middle protocol-version
11416 rollback in the SSL 2.0 server implementation, which is a bad
11417 idea. ([CVE-2005-2969])
11419 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11420 for Information Security, National Institute of Advanced Industrial
11421 Science and Technology [AIST], Japan)*
11423 * Add two function to clear and return the verify parameter flags.
11427 * Keep cipherlists sorted in the source instead of sorting them at
11428 runtime, thus removing the need for a lock.
11432 * Avoid some small subgroup attacks in Diffie-Hellman.
11434 *Nick Mathewson and Ben Laurie*
11436 * Add functions for well-known primes.
11440 * Extended Windows CE support.
11442 *Satoshi Nakamura and Andy Polyakov*
11444 * Initialize SSL_METHOD structures at compile time instead of during
11445 runtime, thus removing the need for a lock.
11449 * Make PKCS7_decrypt() work even if no certificate is supplied by
11450 attempting to decrypt each encrypted key in turn. Add support to
11455 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11457 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11460 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11464 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11465 key into the same file any more.
11469 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11473 * Add -utf8 command line and config file option to 'ca'.
11475 *Stefan <stf@udoma.org*
11477 * Removed the macro des_crypt(), as it seems to conflict with some
11478 libraries. Use DES_crypt().
11482 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11483 involves renaming the source and generated shared-libs for
11484 both. The engines will accept the corrected or legacy ids
11485 ('ncipher' and '4758_cca' respectively) when binding. NB,
11486 this only applies when building 'shared'.
11488 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11490 * Add attribute functions to EVP_PKEY structure. Modify
11491 PKCS12_create() to recognize a CSP name attribute and
11492 use it. Make -CSP option work again in pkcs12 utility.
11496 * Add new functionality to the bn blinding code:
11497 - automatic re-creation of the BN_BLINDING parameters after
11498 a fixed number of uses (currently 32)
11499 - add new function for parameter creation
11500 - introduce flags to control the update behaviour of the
11501 BN_BLINDING parameters
11502 - hide BN_BLINDING structure
11503 Add a second BN_BLINDING slot to the RSA structure to improve
11504 performance when a single RSA object is shared among several
11509 * Add support for DTLS.
11511 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11513 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11514 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11518 * Remove buggy and incomplete DH cert support from
11519 ssl/ssl_rsa.c and ssl/s3_both.c
11523 * Use SHA-1 instead of MD5 as the default digest algorithm for
11524 the `apps/openssl` commands.
11528 * Compile clean with "-Wall -Wmissing-prototypes
11529 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11530 DEBUG_SAFESTACK must also be set.
11534 * Change ./Configure so that certain algorithms can be disabled by default.
11535 The new counterpiece to "no-xxx" is "enable-xxx".
11537 The patented RC5 and MDC2 algorithms will now be disabled unless
11538 "enable-rc5" and "enable-mdc2", respectively, are specified.
11540 (IDEA remains enabled despite being patented. This is because IDEA
11541 is frequently required for interoperability, and there is no license
11542 fee for non-commercial use. As before, "no-idea" can be used to
11543 avoid this algorithm.)
11547 * Add processing of proxy certificates (see RFC 3820). This work was
11548 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11549 EGEE (Enabling Grids for E-science in Europe).
11553 * RC4 performance overhaul on modern architectures/implementations, such
11554 as Intel P4, IA-64 and AMD64.
11558 * New utility extract-section.pl. This can be used specify an alternative
11559 section number in a pod file instead of having to treat each file as
11560 a separate case in Makefile. This can be done by adding two lines to the
11563 =for comment openssl_section:XXX
11565 The blank line is mandatory.
11569 * New arguments -certform, -keyform and -pass for s_client and s_server
11570 to allow alternative format key and certificate files and passphrase
11575 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11576 update associated structures and add various utility functions.
11578 Add new policy related verify parameters, include policy checking in
11579 standard verify code. Enhance 'smime' application with extra parameters
11580 to support policy checking and print out.
11584 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11585 Nehemiah processors. These extensions support AES encryption in hardware
11586 as well as RNG (though RNG support is currently disabled).
11588 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11590 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11594 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11596 *Andy Polyakov and a number of other people*
11598 * Improved PowerPC platform support. Most notably BIGNUM assembler
11599 implementation contributed by IBM.
11601 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11603 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11604 exponent rather than 'unsigned long'. There is a corresponding change to
11605 the new 'rsa_keygen' element of the RSA_METHOD structure.
11607 *Jelte Jansen, Geoff Thorpe*
11609 * Functionality for creating the initial serial number file is now
11610 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11612 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11613 number file to 1, which is bound to cause problems. To avoid
11614 the problems while respecting compatibility between different 0.9.7
11615 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11616 CA.pl for serial number initialization. With the new release 0.9.8,
11617 we can fix the problem directly in the 'ca' utility.)
11621 * Reduced header interdependencies by declaring more opaque objects in
11622 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11623 give fewer recursive includes, which could break lazy source code - so
11624 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11625 developers should define this symbol when building and using openssl to
11626 ensure they track the recommended behaviour, interfaces, [etc], but
11627 backwards-compatible behaviour prevails when this isn't defined.
11631 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11635 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11636 This will generate a random key of the appropriate length based on the
11637 cipher context. The EVP_CIPHER can provide its own random key generation
11638 routine to support keys of a specific form. This is used in the des and
11639 3des routines to generate a key of the correct parity. Update S/MIME
11640 code to use new functions and hence generate correct parity DES keys.
11641 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11642 valid (weak or incorrect parity).
11646 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11647 as looking them up. This is useful when the verified structure may contain
11648 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11649 present unless the new PKCS7_NO_CRL flag is asserted.
11653 * Extend ASN1 oid configuration module. It now additionally accepts the
11656 shortName = some long name, 1.2.3.4
11660 * Reimplemented the BN_CTX implementation. There is now no more static
11661 limitation on the number of variables it can handle nor the depth of the
11662 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11663 information can now expand as required, and rather than having a single
11664 static array of bignums, BN_CTX now uses a linked-list of such arrays
11665 allowing it to expand on demand whilst maintaining the usefulness of
11666 BN_CTX's "bundling".
11670 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11671 to allow all RSA operations to function using a single BN_CTX.
11675 * Preliminary support for certificate policy evaluation and checking. This
11676 is initially intended to pass the tests outlined in "Conformance Testing
11677 of Relying Party Client Certificate Path Processing Logic" v1.07.
11681 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11682 remained unused and not that useful. A variety of other little bignum
11683 tweaks and fixes have also been made continuing on from the audit (see
11688 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11689 associated ASN1, EVP and SSL functions and old ASN1 macros.
11693 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11694 and this should never fail. So the return value from the use of
11695 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11696 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11700 * BN_CTX_get() should return zero-valued bignums, providing the same
11701 initialised value as BN_new().
11703 *Geoff Thorpe, suggested by Ulf Möller*
11705 * Support for inhibitAnyPolicy certificate extension.
11709 * An audit of the BIGNUM code is underway, for which debugging code is
11710 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11711 is considered valid when processing BIGNUMs, and causes execution to
11712 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11713 further steps are taken to deliberately pollute unused data in BIGNUM
11714 structures to try and expose faulty code further on. For now, openssl will
11715 (in its default mode of operation) continue to tolerate the inconsistent
11716 forms that it has tolerated in the past, but authors and packagers should
11717 consider trying openssl and their own applications when compiled with
11718 these debugging symbols defined. It will help highlight potential bugs in
11719 their own code, and will improve the test coverage for OpenSSL itself. At
11720 some point, these tighter rules will become openssl's default to improve
11721 maintainability, though the assert()s and other overheads will remain only
11722 in debugging configurations. See bn.h for more details.
11724 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11726 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11727 that can only be obtained through BN_CTX_new() (which implicitly
11728 initialises it). The presence of this function only made it possible
11729 to overwrite an existing structure (and cause memory leaks).
11733 * Because of the callback-based approach for implementing LHASH as a
11734 template type, lh_insert() adds opaque objects to hash-tables and
11735 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11736 to clean up those corresponding objects before destroying the hash table
11737 (and losing the object pointers). So some over-zealous constifications in
11738 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11739 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11740 prototyped to have "const" restrictions on the object pointers they are
11741 given (and so aren't required to cast them away any more).
11745 * The tmdiff.h API was so ugly and minimal that our own timing utility
11746 (speed) prefers to use its own implementation. The two implementations
11747 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11748 its object type properly exposed (MS_TM) instead of casting to/from
11749 `char *`. This may still change yet if someone realises MS_TM and
11751 aren't necessarily the greatest nomenclatures - but this is what was used
11752 internally to the implementation so I've used that for now.
11756 * Ensure that deprecated functions do not get compiled when
11757 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11758 the self-tests were still using deprecated key-generation functions so
11759 these have been updated also.
11763 * Reorganise PKCS#7 code to separate the digest location functionality
11764 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11765 New function PKCS7_set_digest() to set the digest type for PKCS#7
11766 digestedData type. Add additional code to correctly generate the
11767 digestedData type and add support for this type in PKCS7 initialization
11772 * New function PKCS7_set0_type_other() this initializes a PKCS7
11773 structure of type "other".
11777 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11778 sure the loop does correctly stop and breaking ("division by zero")
11779 modulus operations are not performed. The (pre-generated) prime
11780 table crypto/bn/bn_prime.h was already correct, but it could not be
11781 re-generated on some platforms because of the "division by zero"
11782 situation in the script.
11784 *Ralf S. Engelschall*
11786 * Update support for ECC-based TLS ciphersuites according to
11787 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11788 SHA-1 now is only used for "small" curves (where the
11789 representation of a field element takes up to 24 bytes); for
11790 larger curves, the field element resulting from ECDH is directly
11791 used as premaster secret.
11793 *Douglas Stebila (Sun Microsystems Laboratories)*
11795 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11796 curve secp160r1 to the tests.
11798 *Douglas Stebila (Sun Microsystems Laboratories)*
11800 * Add the possibility to load symbols globally with DSO.
11802 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11804 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11805 control of the error stack.
11809 * Add support for STORE in ENGINE.
11813 * Add the STORE type. The intention is to provide a common interface
11814 to certificate and key stores, be they simple file-based stores, or
11815 HSM-type store, or LDAP stores, or...
11816 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11820 * Add a generic structure called OPENSSL_ITEM. This can be used to
11821 pass a list of arguments to any function as well as provide a way
11822 for a function to pass data back to the caller.
11826 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11827 works like BUF_strdup() but can be used to duplicate a portion of
11828 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11833 * Add the function sk_find_ex() which works like sk_find(), but will
11834 return an index to an element even if an exact match couldn't be
11835 found. The index is guaranteed to point at the element where the
11836 searched-for key would be inserted to preserve sorting order.
11840 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11841 takes an extra flags argument for optional functionality. Currently,
11842 the following flags are defined:
11844 OBJ_BSEARCH_VALUE_ON_NOMATCH
11845 This one gets OBJ_bsearch_ex() to return a pointer to the first
11846 element where the comparing function returns a negative or zero
11849 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11850 This one gets OBJ_bsearch_ex() to return a pointer to the first
11851 element where the comparing function returns zero. This is useful
11852 if there are more than one element where the comparing function
11857 * Make it possible to create self-signed certificates with 'openssl ca'
11858 in such a way that the self-signed certificate becomes part of the
11859 CA database and uses the same mechanisms for serial number generation
11860 as all other certificate signing. The new flag '-selfsign' enables
11861 this functionality. Adapt CA.sh and CA.pl.in.
11865 * Add functionality to check the public key of a certificate request
11866 against a given private. This is useful to check that a certificate
11867 request can be signed by that key (self-signing).
11871 * Make it possible to have multiple active certificates with the same
11872 subject in the CA index file. This is done only if the keyword
11873 'unique_subject' is set to 'no' in the main CA section (default
11874 if 'CA_default') of the configuration file. The value is saved
11875 with the database itself in a separate index attribute file,
11876 named like the index file with '.attr' appended to the name.
11880 * Generate multi-valued AVAs using '+' notation in config files for
11885 * Support for nameConstraints certificate extension.
11889 * Support for policyConstraints certificate extension.
11893 * Support for policyMappings certificate extension.
11897 * Make sure the default DSA_METHOD implementation only uses its
11898 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11899 and change its own handlers to be NULL so as to remove unnecessary
11900 indirection. This lets alternative implementations fallback to the
11901 default implementation more easily.
11905 * Support for directoryName in GeneralName related extensions
11910 * Make it possible to link applications using Makefile.shared.
11911 Make that possible even when linking against static libraries!
11915 * Support for single pass processing for S/MIME signing. This now
11916 means that S/MIME signing can be done from a pipe, in addition
11917 cleartext signing (multipart/signed type) is effectively streaming
11918 and the signed data does not need to be all held in memory.
11920 This is done with a new flag PKCS7_STREAM. When this flag is set
11921 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11922 is done after the data is output (and digests calculated) in
11923 SMIME_write_PKCS7().
11927 * Add full support for -rpath/-R, both in shared libraries and
11928 applications, at least on the platforms where it's known how
11933 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11934 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11935 will now compute a table of multiples of the generator that
11936 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11937 faster (notably in the case of a single point multiplication,
11938 scalar * generator).
11940 *Nils Larsch, Bodo Moeller*
11942 * IPv6 support for certificate extensions. The various extensions
11943 which use the IP:a.b.c.d can now take IPv6 addresses using the
11944 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11949 * Added an ENGINE that implements RSA by performing private key
11950 exponentiations with the GMP library. The conversions to and from
11951 GMP's mpz_t format aren't optimised nor are any montgomery forms
11952 cached, and on x86 it appears OpenSSL's own performance has caught up.
11953 However there are likely to be other architectures where GMP could
11954 provide a boost. This ENGINE is not built in by default, but it can be
11955 specified at Configure time and should be accompanied by the necessary
11956 linker additions, eg;
11957 ./config -DOPENSSL_USE_GMP -lgmp
11961 * "openssl engine" will not display ENGINE/DSO load failure errors when
11962 testing availability of engines with "-t" - the old behaviour is
11963 produced by increasing the feature's verbosity with "-tt".
11967 * ECDSA routines: under certain error conditions uninitialized BN objects
11968 could be freed. Solution: make sure initialization is performed early
11969 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11974 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11975 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11976 software implementations. For DSA and DH, parameter generation can
11977 also be overridden by providing the appropriate method callbacks.
11981 * Change the "progress" mechanism used in key-generation and
11982 primality testing to functions that take a new BN_GENCB pointer in
11983 place of callback/argument pairs. The new API functions have `_ex`
11984 postfixes and the older functions are reimplemented as wrappers for
11985 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11986 declarations of the old functions to help (graceful) attempts to
11987 migrate to the new functions. Also, the new key-generation API
11988 functions operate on a caller-supplied key-structure and return
11989 success/failure rather than returning a key or NULL - this is to
11990 help make "keygen" another member function of RSA_METHOD etc.
11992 Example for using the new callback interface:
11994 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11995 void *my_arg = ...;
11998 BN_GENCB_set(&my_cb, my_callback, my_arg);
12000 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
12001 /* For the meaning of a, b in calls to my_callback(), see the
12002 * documentation of the function that calls the callback.
12003 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
12004 * my_callback should return 1 if it wants BN_is_prime_ex()
12005 * to continue, or 0 to stop.
12010 * Change the ZLIB compression method to be stateful, and make it
12011 available to TLS with the number defined in
12012 draft-ietf-tls-compression-04.txt.
12016 * Add the ASN.1 structures and functions for CertificatePair, which
12017 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
12019 CertificatePair ::= SEQUENCE {
12020 forward [0] Certificate OPTIONAL,
12021 reverse [1] Certificate OPTIONAL,
12022 -- at least one of the pair shall be present -- }
12024 Also implement the PEM functions to read and write certificate
12025 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
12027 This needed to be defined, mostly for the sake of the LDAP
12028 attribute crossCertificatePair, but may prove useful elsewhere as
12033 * Make it possible to inhibit symlinking of shared libraries in
12034 Makefile.shared, for Cygwin's sake.
12038 * Extend the BIGNUM API by creating a function
12039 void BN_set_negative(BIGNUM *a, int neg);
12040 and a macro that behave like
12041 int BN_is_negative(const BIGNUM *a);
12043 to avoid the need to access 'a->neg' directly in applications.
12047 * Implement fast modular reduction for pseudo-Mersenne primes
12048 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
12049 EC_GROUP_new_curve_GFp() will now automatically use this
12052 *Nils Larsch <nla@trustcenter.de>*
12054 * Add new lock type (CRYPTO_LOCK_BN).
12058 * Change the ENGINE framework to automatically load engines
12059 dynamically from specific directories unless they could be
12060 found to already be built in or loaded. Move all the
12061 current engines except for the cryptodev one to a new
12062 directory engines/.
12063 The engines in engines/ are built as shared libraries if
12064 the "shared" options was given to ./Configure or ./config.
12065 Otherwise, they are inserted in libcrypto.a.
12066 /usr/local/ssl/engines is the default directory for dynamic
12067 engines, but that can be overridden at configure time through
12068 the usual use of --prefix and/or --openssldir, and at run
12069 time with the environment variable OPENSSL_ENGINES.
12071 *Geoff Thorpe and Richard Levitte*
12073 * Add Makefile.shared, a helper makefile to build shared
12074 libraries. Adapt Makefile.org.
12078 * Add version info to Win32 DLLs.
12080 *Peter 'Luna' Runestig" <peter@runestig.com>*
12082 * Add new 'medium level' PKCS#12 API. Certificates and keys
12083 can be added using this API to created arbitrary PKCS#12
12084 files while avoiding the low-level API.
12086 New options to PKCS12_create(), key or cert can be NULL and
12087 will then be omitted from the output file. The encryption
12088 algorithm NIDs can be set to -1 for no encryption, the mac
12089 iteration count can be set to 0 to omit the mac.
12091 Enhance pkcs12 utility by making the -nokeys and -nocerts
12092 options work when creating a PKCS#12 file. New option -nomac
12093 to omit the mac, NONE can be set for an encryption algorithm.
12094 New code is modified to use the enhanced PKCS12_create()
12095 instead of the low-level API.
12099 * Extend ASN1 encoder to support indefinite length constructed
12100 encoding. This can output sequences tags and octet strings in
12101 this form. Modify pk7_asn1.c to support indefinite length
12102 encoding. This is experimental and needs additional code to
12103 be useful, such as an ASN1 bio and some enhanced streaming
12106 Extend template encode functionality so that tagging is passed
12107 down to the template encoder.
12111 * Let 'openssl req' fail if an argument to '-newkey' is not
12112 recognized instead of using RSA as a default.
12116 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
12117 As these are not official, they are not included in "ALL";
12118 the "ECCdraft" ciphersuite group alias can be used to select them.
12120 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
12122 * Add ECDH engine support.
12124 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
12126 * Add ECDH in new directory crypto/ecdh/.
12128 *Douglas Stebila (Sun Microsystems Laboratories)*
12130 * Let BN_rand_range() abort with an error after 100 iterations
12131 without success (which indicates a broken PRNG).
12135 * Change BN_mod_sqrt() so that it verifies that the input value
12136 is really the square of the return value. (Previously,
12137 BN_mod_sqrt would show GIGO behaviour.)
12141 * Add named elliptic curves over binary fields from X9.62, SECG,
12142 and WAP/WTLS; add OIDs that were still missing.
12144 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12146 * Extend the EC library for elliptic curves over binary fields
12147 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
12150 EC_GF2m_simple_method
12154 EC_GROUP_new_curve_GF2m
12155 EC_GROUP_set_curve_GF2m
12156 EC_GROUP_get_curve_GF2m
12157 EC_POINT_set_affine_coordinates_GF2m
12158 EC_POINT_get_affine_coordinates_GF2m
12159 EC_POINT_set_compressed_coordinates_GF2m
12161 Point compression for binary fields is disabled by default for
12162 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
12165 As binary polynomials are represented as BIGNUMs, various members
12166 of the EC_GROUP and EC_POINT data structures can be shared
12167 between the implementations for prime fields and binary fields;
12168 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
12169 are essentially identical to their `..._GFp` counterparts.
12170 (For simplicity, the `..._GFp` prefix has been dropped from
12171 various internal method names.)
12173 An internal 'field_div' method (similar to 'field_mul' and
12174 'field_sqr') has been added; this is used only for binary fields.
12176 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12178 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
12179 through methods ('mul', 'precompute_mult').
12181 The generic implementations (now internally called 'ec_wNAF_mul'
12182 and 'ec_wNAF_precomputed_mult') remain the default if these
12183 methods are undefined.
12185 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12187 * New function EC_GROUP_get_degree, which is defined through
12188 EC_METHOD. For curves over prime fields, this returns the bit
12189 length of the modulus.
12191 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12193 * New functions EC_GROUP_dup, EC_POINT_dup.
12194 (These simply call ..._new and ..._copy).
12196 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12198 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
12199 Polynomials are represented as BIGNUMs (where the sign bit is not
12200 used) in the following functions [macros]:
12203 BN_GF2m_sub [= BN_GF2m_add]
12204 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
12205 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
12206 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
12208 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
12209 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
12210 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
12211 BN_GF2m_cmp [= BN_ucmp]
12213 (Note that only the 'mod' functions are actually for fields GF(2^m).
12214 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
12216 For some functions, an the irreducible polynomial defining a
12217 field can be given as an 'unsigned int[]' with strictly
12218 decreasing elements giving the indices of those bits that are set;
12219 i.e., p[] represents the polynomial
12220 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
12222 p[0] > p[1] > ... > p[k] = 0.
12223 This applies to the following functions:
12226 BN_GF2m_mod_mul_arr
12227 BN_GF2m_mod_sqr_arr
12228 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
12229 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
12230 BN_GF2m_mod_exp_arr
12231 BN_GF2m_mod_sqrt_arr
12232 BN_GF2m_mod_solve_quad_arr
12236 Conversion can be performed by the following functions:
12241 bntest.c has additional tests for binary polynomial arithmetic.
12243 Two implementations for BN_GF2m_mod_div() are available.
12244 The default algorithm simply uses BN_GF2m_mod_inv() and
12245 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
12246 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
12247 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
12249 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12251 * Add new error code 'ERR_R_DISABLED' that can be used when some
12252 functionality is disabled at compile-time.
12254 *Douglas Stebila <douglas.stebila@sun.com>*
12256 * Change default behaviour of 'openssl asn1parse' so that more
12257 information is visible when viewing, e.g., a certificate:
12259 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
12260 mode the content of non-printable OCTET STRINGs is output in a
12261 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
12262 avoid the appearance of a printable string.
12264 *Nils Larsch <nla@trustcenter.de>*
12266 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
12268 EC_GROUP_set_asn1_flag()
12269 EC_GROUP_get_asn1_flag()
12270 EC_GROUP_set_point_conversion_form()
12271 EC_GROUP_get_point_conversion_form()
12272 These control ASN1 encoding details:
12273 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
12274 has been set to OPENSSL_EC_NAMED_CURVE.
12275 - Points are encoded in uncompressed form by default; options for
12276 asn1_for are as for point2oct, namely
12277 POINT_CONVERSION_COMPRESSED
12278 POINT_CONVERSION_UNCOMPRESSED
12279 POINT_CONVERSION_HYBRID
12281 Also add 'seed' and 'seed_len' members to EC_GROUP with access
12283 EC_GROUP_set_seed()
12284 EC_GROUP_get0_seed()
12285 EC_GROUP_get_seed_len()
12286 This is used only for ASN1 purposes (so far).
12288 *Nils Larsch <nla@trustcenter.de>*
12290 * Add 'field_type' member to EC_METHOD, which holds the NID
12291 of the appropriate field type OID. The new function
12292 EC_METHOD_get_field_type() returns this value.
12294 *Nils Larsch <nla@trustcenter.de>*
12297 EC_POINT_point2bn()
12298 EC_POINT_bn2point()
12299 EC_POINT_point2hex()
12300 EC_POINT_hex2point()
12301 providing useful interfaces to EC_POINT_point2oct() and
12302 EC_POINT_oct2point().
12304 *Nils Larsch <nla@trustcenter.de>*
12306 * Change internals of the EC library so that the functions
12307 EC_GROUP_set_generator()
12308 EC_GROUP_get_generator()
12309 EC_GROUP_get_order()
12310 EC_GROUP_get_cofactor()
12311 are implemented directly in crypto/ec/ec_lib.c and not dispatched
12312 to methods, which would lead to unnecessary code duplication when
12313 adding different types of curves.
12315 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
12317 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
12318 arithmetic, and such that modified wNAFs are generated
12319 (which avoid length expansion in many cases).
12323 * Add a function EC_GROUP_check_discriminant() (defined via
12324 EC_METHOD) that verifies that the curve discriminant is non-zero.
12326 Add a function EC_GROUP_check() that makes some sanity tests
12327 on a EC_GROUP, its generator and order. This includes
12328 EC_GROUP_check_discriminant().
12330 *Nils Larsch <nla@trustcenter.de>*
12332 * Add ECDSA in new directory crypto/ecdsa/.
12334 Add applications 'openssl ecparam' and 'openssl ecdsa'
12335 (these are based on 'openssl dsaparam' and 'openssl dsa').
12337 ECDSA support is also included in various other files across the
12338 library. Most notably,
12339 - 'openssl req' now has a '-newkey ecdsa:file' option;
12340 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
12341 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
12342 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
12343 them suitable for ECDSA where domain parameters must be
12344 extracted before the specific public key;
12345 - ECDSA engine support has been added.
12347 *Nils Larsch <nla@trustcenter.de>*
12349 * Include some named elliptic curves, and add OIDs from X9.62,
12350 SECG, and WAP/WTLS. Each curve can be obtained from the new
12352 EC_GROUP_new_by_curve_name(),
12353 and the list of available named curves can be obtained with
12354 EC_get_builtin_curves().
12355 Also add a 'curve_name' member to EC_GROUP objects, which can be
12357 EC_GROUP_set_curve_name()
12358 EC_GROUP_get_curve_name()
12360 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
12362 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12363 was actually never needed) and in BN_mul(). The removal in BN_mul()
12364 required a small change in bn_mul_part_recursive() and the addition
12365 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12366 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12367 bn_sub_words() and bn_add_words() except they take arrays with
12372 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
12374 * Cleanse PEM buffers before freeing them since they may contain
12377 *Benjamin Bennett <ben@psc.edu>*
12379 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
12380 a ciphersuite string such as "DEFAULT:RSA" cannot enable
12381 authentication-only ciphersuites.
12385 * Since AES128 and AES256 share a single mask bit in the logic of
12386 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
12387 kludge to work properly if AES128 is available and AES256 isn't.
12391 * Expand security boundary to match 1.1.1 module.
12395 * Remove redundant features: hash file source, editing of test vectors
12396 modify fipsld to use external fips_premain.c signature.
12400 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12401 run algorithm test programs.
12405 * Make algorithm test programs more tolerant of whitespace.
12409 * Have SSL/TLS server implementation tolerate "mismatched" record
12410 protocol version while receiving ClientHello even if the
12411 ClientHello is fragmented. (The server can't insist on the
12412 particular protocol version it has chosen before the ServerHello
12413 message has informed the client about his choice.)
12417 * Load error codes if they are not already present instead of using a
12418 static variable. This allows them to be cleanly unloaded and reloaded.
12422 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12424 * Introduce limits to prevent malicious keys being able to
12425 cause a denial of service. ([CVE-2006-2940])
12427 *Steve Henson, Bodo Moeller*
12429 * Fix ASN.1 parsing of certain invalid structures that can result
12430 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12432 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12433 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12435 * Fix SSL client code which could crash if connecting to a
12436 malicious SSLv2 server. ([CVE-2006-4343])
12438 *Tavis Ormandy and Will Drewry, Google Security Team*
12440 * Change ciphersuite string processing so that an explicit
12441 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12442 will no longer include "AES128-SHA"), and any other similar
12443 ciphersuite (same bitmap) from *other* protocol versions (so that
12444 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12445 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12446 changes from 0.9.8b and 0.9.8d.
12450 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12452 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12453 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12455 * Change the Unix randomness entropy gathering to use poll() when
12456 possible instead of select(), since the latter has some
12457 undesirable limitations.
12459 *Darryl Miles via Richard Levitte and Bodo Moeller*
12461 * Disable rogue ciphersuites:
12463 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12464 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12465 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12467 The latter two were purportedly from
12468 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12471 Also deactivate the remaining ciphersuites from
12472 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12473 unofficial, and the ID has long expired.
12477 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12478 dual-core machines) and other potential thread-safety issues.
12482 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12484 * Adapt fipsld and the build system to link against the validated FIPS
12485 module in FIPS mode.
12489 * Fixes for VC++ 2005 build under Windows.
12493 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12494 from a Windows bash shell such as MSYS. It is autodetected from the
12495 "config" script when run from a VC++ environment. Modify standard VC++
12496 build to use fipscanister.o from the GNU make build.
12500 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12502 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12503 The value now differs depending on if you build for FIPS or not.
12504 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12505 safely run with a non-FIPSed libcrypto, as it may crash because of
12506 the difference induced by this change.
12510 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12512 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12513 (part of SSL_OP_ALL). This option used to disable the
12514 countermeasure against man-in-the-middle protocol-version
12515 rollback in the SSL 2.0 server implementation, which is a bad
12516 idea. ([CVE-2005-2969])
12518 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12519 for Information Security, National Institute of Advanced Industrial
12520 Science and Technology [AIST, Japan)]*
12522 * Minimal support for X9.31 signatures and PSS padding modes. This is
12523 mainly for FIPS compliance and not fully integrated at this stage.
12527 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12528 the exponentiation using a fixed-length exponent. (Otherwise,
12529 the information leaked through timing could expose the secret key
12530 after many signatures; cf. Bleichenbacher's attack on DSA with
12535 * Make a new fixed-window mod_exp implementation the default for
12536 RSA, DSA, and DH private-key operations so that the sequence of
12537 squares and multiplies and the memory access pattern are
12538 independent of the particular secret key. This will mitigate
12539 cache-timing and potential related attacks.
12541 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12542 and this is automatically used by BN_mod_exp_mont() if the new flag
12543 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12544 will use this BN flag for private exponents unless the flag
12545 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12546 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12548 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12550 * Change the client implementation for SSLv23_method() and
12551 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12552 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12553 (Previously, the SSL 2.0 backwards compatible Client Hello
12554 message format would be used even with SSL_OP_NO_SSLv2.)
12558 * Add support for smime-type MIME parameter in S/MIME messages which some
12563 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12564 a threadsafe manner. Modify rsa code to use new function and add calls
12565 to dsa and dh code (which had race conditions before).
12569 * Include the fixed error library code in the C error file definitions
12570 instead of fixing them up at runtime. This keeps the error code
12571 structures constant.
12575 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12577 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12580 * Fixes for newer kerberos headers. NB: the casts are needed because
12581 the 'length' field is signed on one version and unsigned on another
12582 with no (?) obvious way to tell the difference, without these VC++
12583 complains. Also the "definition" of FAR (blank) is no longer included
12584 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12585 some needed definitions.
12589 * Undo Cygwin change.
12593 * Added support for proxy certificates according to RFC 3820.
12594 Because they may be a security thread to unaware applications,
12595 they must be explicitly allowed in run-time. See
12596 docs/HOWTO/proxy_certificates.txt for further information.
12600 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12602 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12603 server and client random values. Previously
12604 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12605 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12607 This change has negligible security impact because:
12609 1. Server and client random values still have 24 bytes of pseudo random
12612 2. Server and client random values are sent in the clear in the initial
12615 3. The master secret is derived using the premaster secret (48 bytes in
12616 size for static RSA ciphersuites) as well as client server and random
12619 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12622 *Stephen Henson, reported by UK NISCC*
12624 * Use Windows randomness collection on Cygwin.
12628 * Fix hang in EGD/PRNGD query when communication socket is closed
12629 prematurely by EGD/PRNGD.
12631 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12633 * Prompt for pass phrases when appropriate for PKCS12 input format.
12637 * Back-port of selected performance improvements from development
12638 branch, as well as improved support for PowerPC platforms.
12642 * Add lots of checks for memory allocation failure, error codes to indicate
12643 failure and freeing up memory if a failure occurs.
12645 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12647 * Add new -passin argument to dgst.
12651 * Perform some character comparisons of different types in X509_NAME_cmp:
12652 this is needed for some certificates that re-encode DNs into UTF8Strings
12653 (in violation of RFC3280) and can't or won't issue name rollover
12658 * Make an explicit check during certificate validation to see that
12659 the CA setting in each certificate on the chain is correct. As a
12660 side effect always do the following basic checks on extensions,
12661 not just when there's an associated purpose to the check:
12663 - if there is an unhandled critical extension (unless the user
12664 has chosen to ignore this fault)
12665 - if the path length has been exceeded (if one is set at all)
12666 - that certain extensions fit the associated purpose (if one has
12671 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12673 * Avoid a race condition when CRLs are checked in a multi threaded
12674 environment. This would happen due to the reordering of the revoked
12675 entries during signature checking and serial number lookup. Now the
12676 encoding is cached and the serial number sort performed under a lock.
12677 Add new STACK function sk_is_sorted().
12681 * Add Delta CRL to the extension code.
12685 * Various fixes to s3_pkt.c so alerts are sent properly.
12687 *David Holmes <d.holmes@f5.com>*
12689 * Reduce the chances of duplicate issuer name and serial numbers (in
12690 violation of RFC3280) using the OpenSSL certificate creation utilities.
12691 This is done by creating a random 64 bit value for the initial serial
12692 number when a serial number file is created or when a self signed
12693 certificate is created using 'openssl req -x509'. The initial serial
12694 number file is created using 'openssl x509 -next_serial' in CA.pl
12695 rather than being initialized to 1.
12699 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12701 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12702 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12704 *Joe Orton, Steve Henson*
12706 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12709 *Joe Orton, Steve Henson*
12711 * Make it possible to have multiple active certificates with the same
12712 subject in the CA index file. This is done only if the keyword
12713 'unique_subject' is set to 'no' in the main CA section (default
12714 if 'CA_default') of the configuration file. The value is saved
12715 with the database itself in a separate index attribute file,
12716 named like the index file with '.attr' appended to the name.
12720 * X509 verify fixes. Disable broken certificate workarounds when
12721 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12722 keyUsage extension present. Don't accept CRLs with unhandled critical
12723 extensions: since verify currently doesn't process CRL extensions this
12724 rejects a CRL with *any* critical extensions. Add new verify error codes
12729 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12730 A clarification of RFC2560 will require the use of OCTET STRINGs and
12731 some implementations cannot handle the current raw format. Since OpenSSL
12732 copies and compares OCSP nonces as opaque blobs without any attempt at
12733 parsing them this should not create any compatibility issues.
12737 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12738 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12739 this HMAC (and other) operations are several times slower than OpenSSL
12744 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12746 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12748 * Use the correct content when signing type "other".
12752 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12754 * Fix various bugs revealed by running the NISCC test suite:
12756 Stop out of bounds reads in the ASN1 code when presented with
12757 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12759 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12761 If verify callback ignores invalid public key errors don't try to check
12762 certificate signature with the NULL public key.
12766 * New -ignore_err option in ocsp application to stop the server
12767 exiting on the first error in a request.
12771 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12772 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12777 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12778 extra data after the compression methods not only for TLS 1.0
12779 but also for SSL 3.0 (as required by the specification).
12781 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12783 * Change X509_certificate_type() to mark the key as exported/exportable
12784 when it's 512 *bits* long, not 512 bytes.
12788 * Change AES_cbc_encrypt() so it outputs exact multiple of
12789 blocks during encryption.
12793 * Various fixes to base64 BIO and non blocking I/O. On write
12794 flushes were not handled properly if the BIO retried. On read
12795 data was not being buffered properly and had various logic bugs.
12796 This also affects blocking I/O when the data being decoded is a
12801 * Various S/MIME bugfixes and compatibility changes:
12802 output correct application/pkcs7 MIME type if
12803 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12804 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12805 of files as .eml work). Correctly handle very long lines in MIME
12810 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12812 * Countermeasure against the Klima-Pokorny-Rosa extension of
12813 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12814 a protocol version number mismatch like a decryption error
12815 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12819 * Turn on RSA blinding by default in the default implementation
12820 to avoid a timing attack. Applications that don't want it can call
12821 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12822 They would be ill-advised to do so in most cases.
12824 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12826 * Change RSA blinding code so that it works when the PRNG is not
12827 seeded (in this case, the secret RSA exponent is abused as
12828 an unpredictable seed -- if it is not unpredictable, there
12829 is no point in blinding anyway). Make RSA blinding thread-safe
12830 by remembering the creator's thread ID in rsa->blinding and
12831 having all other threads use local one-time blinding factors
12832 (this requires more computation than sharing rsa->blinding, but
12833 avoids excessive locking; and if an RSA object is not shared
12834 between threads, blinding will still be very fast).
12838 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12839 ENGINE as defaults for all supported algorithms irrespective of
12840 the 'flags' parameter. 'flags' is now honoured, so applications
12841 should make sure they are passing it correctly.
12845 * Target "mingw" now allows native Windows code to be generated in
12846 the Cygwin environment as well as with the MinGW compiler.
12850 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12852 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12853 via timing by performing a MAC computation even if incorrect
12854 block cipher padding has been found. This is a countermeasure
12855 against active attacks where the attacker has to distinguish
12856 between bad padding and a MAC verification error. ([CVE-2003-0078])
12858 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12859 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12860 Martin Vuagnoux (EPFL, Ilion)*
12862 * Make the no-err option work as intended. The intention with no-err
12863 is not to have the whole error stack handling routines removed from
12864 libcrypto, it's only intended to remove all the function name and
12865 reason texts, thereby removing some of the footprint that may not
12866 be interesting if those errors aren't displayed anyway.
12868 NOTE: it's still possible for any application or module to have its
12869 own set of error texts inserted. The routines are there, just not
12870 used by default when no-err is given.
12874 * Add support for FreeBSD on IA64.
12876 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12878 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12879 Kerberos function mit_des_cbc_cksum(). Before this change,
12880 the value returned by DES_cbc_cksum() was like the one from
12881 mit_des_cbc_cksum(), except the bytes were swapped.
12883 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12885 * Allow an application to disable the automatic SSL chain building.
12886 Before this a rather primitive chain build was always performed in
12887 ssl3_output_cert_chain(): an application had no way to send the
12888 correct chain if the automatic operation produced an incorrect result.
12890 Now the chain builder is disabled if either:
12892 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12894 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12896 The reasoning behind this is that an application would not want the
12897 auto chain building to take place if extra chain certificates are
12898 present and it might also want a means of sending no additional
12899 certificates (for example the chain has two certificates and the
12904 * Add the possibility to build without the ENGINE framework.
12906 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12908 * Under Win32 gmtime() can return NULL: check return value in
12909 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12913 * DSA routines: under certain error conditions uninitialized BN objects
12914 could be freed. Solution: make sure initialization is performed early
12915 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12916 Nils Larsch <nla@trustcenter.de> via PR#459)
12920 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12921 checked on reconnect on the client side, therefore session resumption
12922 could still fail with a "ssl session id is different" error. This
12923 behaviour is masked when SSL_OP_ALL is used due to
12924 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12925 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12926 followup to PR #377.
12930 * IA-32 assembler support enhancements: unified ELF targets, support
12931 for SCO/Caldera platforms, fix for Cygwin shared build.
12935 * Add support for FreeBSD on sparc64. As a consequence, support for
12936 FreeBSD on non-x86 processors is separate from x86 processors on
12937 the config script, much like the NetBSD support.
12939 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12941 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12943 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12946 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12947 code (06) was taken as the first octet of the session ID and the last
12948 octet was ignored consequently. As a result SSLv2 client side session
12949 caching could not have worked due to the session ID mismatch between
12951 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12956 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12957 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12962 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12963 seems that in spite of existing for more than a year, many application
12964 author have done nothing to provide the necessary callbacks, which
12965 means that this particular engine will not work properly anywhere.
12966 This is a very unfortunate situation which forces us, in the name
12967 of usability, to give the hw_ncipher.c a static lock, which is part
12969 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12970 appear in 0.9.8 or later. We EXPECT application authors to have
12971 dealt properly with this when 0.9.8 is released (unless we actually
12972 make such changes in the libcrypto locking code that changes will
12973 have to be made anyway).
12977 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12978 octets have been read, EOF or an error occurs. Without this change
12979 some truncated ASN1 structures will not produce an error.
12983 * Disable Heimdal support, since it hasn't been fully implemented.
12984 Still give the possibility to force the use of Heimdal, but with
12985 warnings and a request that patches get sent to openssl-dev.
12989 * Add the VC-CE target, introduce the WINCE sysname, and add
12990 INSTALL.WCE and appropriate conditionals to make it build.
12992 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12994 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12995 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12996 edit numbers of the version.
12998 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13000 * Introduce safe string copy and catenation functions
13001 (BUF_strlcpy() and BUF_strlcat()).
13003 *Ben Laurie (CHATS) and Richard Levitte*
13005 * Avoid using fixed-size buffers for one-line DNs.
13007 *Ben Laurie (CHATS)*
13009 * Add BUF_MEM_grow_clean() to avoid information leakage when
13010 resizing buffers containing secrets, and use where appropriate.
13012 *Ben Laurie (CHATS)*
13014 * Avoid using fixed size buffers for configuration file location.
13016 *Ben Laurie (CHATS)*
13018 * Avoid filename truncation for various CA files.
13020 *Ben Laurie (CHATS)*
13022 * Use sizeof in preference to magic numbers.
13024 *Ben Laurie (CHATS)*
13026 * Avoid filename truncation in cert requests.
13028 *Ben Laurie (CHATS)*
13030 * Add assertions to check for (supposedly impossible) buffer
13033 *Ben Laurie (CHATS)*
13035 * Don't cache truncated DNS entries in the local cache (this could
13036 potentially lead to a spoofing attack).
13038 *Ben Laurie (CHATS)*
13040 * Fix various buffers to be large enough for hex/decimal
13041 representations in a platform independent manner.
13043 *Ben Laurie (CHATS)*
13045 * Add CRYPTO_realloc_clean() to avoid information leakage when
13046 resizing buffers containing secrets, and use where appropriate.
13048 *Ben Laurie (CHATS)*
13050 * Add BIO_indent() to avoid much slightly worrying code to do
13053 *Ben Laurie (CHATS)*
13055 * Convert sprintf()/BIO_puts() to BIO_printf().
13057 *Ben Laurie (CHATS)*
13059 * buffer_gets() could terminate with the buffer only half
13062 *Ben Laurie (CHATS)*
13064 * Add assertions to prevent user-supplied crypto functions from
13065 overflowing internal buffers by having large block sizes, etc.
13067 *Ben Laurie (CHATS)*
13069 * New OPENSSL_assert() macro (similar to assert(), but enabled
13072 *Ben Laurie (CHATS)*
13074 * Eliminate unused copy of key in RC4.
13076 *Ben Laurie (CHATS)*
13078 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
13080 *Ben Laurie (CHATS)*
13082 * Fix off-by-one error in EGD path.
13084 *Ben Laurie (CHATS)*
13086 * If RANDFILE path is too long, ignore instead of truncating.
13088 *Ben Laurie (CHATS)*
13090 * Eliminate unused and incorrectly sized X.509 structure
13093 *Ben Laurie (CHATS)*
13095 * Eliminate unused and dangerous function knumber().
13097 *Ben Laurie (CHATS)*
13099 * Eliminate unused and dangerous structure, KSSL_ERR.
13101 *Ben Laurie (CHATS)*
13103 * Protect against overlong session ID context length in an encoded
13104 session object. Since these are local, this does not appear to be
13107 *Ben Laurie (CHATS)*
13109 * Change from security patch (see 0.9.6e below) that did not affect
13110 the 0.9.6 release series:
13112 Remote buffer overflow in SSL3 protocol - an attacker could
13113 supply an oversized master key in Kerberos-enabled versions.
13116 *Ben Laurie (CHATS)*
13118 * Change the SSL kerb5 codes to match RFC 2712.
13122 * Make -nameopt work fully for req and add -reqopt switch.
13124 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
13126 * The "block size" for block ciphers in CFB and OFB mode should be 1.
13128 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
13130 * Make sure tests can be performed even if the corresponding algorithms
13131 have been removed entirely. This was also the last step to make
13132 OpenSSL compilable with DJGPP under all reasonable conditions.
13134 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
13136 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
13137 to allow version independent disabling of normally unselected ciphers,
13138 which may be activated as a side-effect of selecting a single cipher.
13140 (E.g., cipher list string "RSA" enables ciphersuites that are left
13141 out of "ALL" because they do not provide symmetric encryption.
13142 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
13144 *Lutz Jaenicke, Bodo Moeller*
13146 * Add appropriate support for separate platform-dependent build
13147 directories. The recommended way to make a platform-dependent
13148 build directory is the following (tested on Linux), maybe with
13151 # Place yourself outside of the OpenSSL source tree. In
13152 # this example, the environment variable OPENSSL_SOURCE
13153 # is assumed to contain the absolute OpenSSL source directory.
13154 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
13155 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
13156 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
13157 mkdir -p `dirname $F`
13158 ln -s $OPENSSL_SOURCE/$F $F
13161 To be absolutely sure not to disturb the source tree, a "make clean"
13162 is a good thing. If it isn't successful, don't worry about it,
13163 it probably means the source directory is very clean.
13167 * Make sure any ENGINE control commands make local copies of string
13168 pointers passed to them whenever necessary. Otherwise it is possible
13169 the caller may have overwritten (or deallocated) the original string
13170 data when a later ENGINE operation tries to use the stored values.
13172 *Götz Babin-Ebell <babinebell@trustcenter.de>*
13174 * Improve diagnostics in file reading and command-line digests.
13176 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
13178 * Add AES modes CFB and OFB to the object database. Correct an
13179 error in AES-CFB decryption.
13183 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
13184 allows existing EVP_CIPHER_CTX structures to be reused after
13185 calling `EVP_*Final()`. This behaviour is used by encryption
13186 BIOs and some applications. This has the side effect that
13187 applications must explicitly clean up cipher contexts with
13188 EVP_CIPHER_CTX_cleanup() or they will leak memory.
13192 * Check the values of dna and dnb in bn_mul_recursive before calling
13193 bn_mul_comba (a non zero value means the a or b arrays do not contain
13194 n2 elements) and fallback to bn_mul_normal if either is not zero.
13198 * Fix escaping of non-ASCII characters when using the -subj option
13199 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
13203 * Make object definitions compliant to LDAP (RFC2256): SN is the short
13204 form for "surname", serialNumber has no short form.
13205 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
13206 therefore remove "mail" short name for "internet 7".
13207 The OID for unique identifiers in X509 certificates is
13208 x500UniqueIdentifier, not uniqueIdentifier.
13209 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
13213 * Add an "init" command to the ENGINE config module and auto initialize
13214 ENGINEs. Without any "init" command the ENGINE will be initialized
13215 after all ctrl commands have been executed on it. If init=1 the
13216 ENGINE is initialized at that point (ctrls before that point are run
13217 on the uninitialized ENGINE and after on the initialized one). If
13218 init=0 then the ENGINE will not be initialized at all.
13222 * Fix the 'app_verify_callback' interface so that the user-defined
13223 argument is actually passed to the callback: In the
13224 SSL_CTX_set_cert_verify_callback() prototype, the callback
13225 declaration has been changed from
13228 int (*cb)(X509_STORE_CTX *,void *);
13229 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
13230 i=s->ctx->app_verify_callback(&ctx)
13231 has been changed into
13232 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
13234 To update applications using SSL_CTX_set_cert_verify_callback(),
13235 a dummy argument can be added to their callback functions.
13237 *D. K. Smetters <smetters@parc.xerox.com>*
13239 * Added the '4758cca' ENGINE to support IBM 4758 cards.
13241 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
13243 * Add and OPENSSL_LOAD_CONF define which will cause
13244 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
13245 This allows older applications to transparently support certain
13246 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
13247 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
13248 load the config file and OPENSSL_add_all_algorithms_conf() which will
13249 always load it have also been added.
13253 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
13254 Adjust NIDs and EVP layer.
13256 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13258 * Config modules support in openssl utility.
13260 Most commands now load modules from the config file,
13261 though in a few (such as version) this isn't done
13262 because it couldn't be used for anything.
13264 In the case of ca and req the config file used is
13265 the same as the utility itself: that is the -config
13266 command line option can be used to specify an
13271 * Move default behaviour from OPENSSL_config(). If appname is NULL
13272 use "openssl_conf" if filename is NULL use default openssl config file.
13276 * Add an argument to OPENSSL_config() to allow the use of an alternative
13277 config section name. Add a new flag to tolerate a missing config file
13278 and move code to CONF_modules_load_file().
13282 * Support for crypto accelerator cards from Accelerated Encryption
13283 Processing, www.aep.ie. (Use engine 'aep')
13284 The support was copied from 0.9.6c [engine] and adapted/corrected
13285 to work with the new engine framework.
13287 *AEP Inc. and Richard Levitte*
13289 * Support for SureWare crypto accelerator cards from Baltimore
13290 Technologies. (Use engine 'sureware')
13291 The support was copied from 0.9.6c [engine] and adapted
13292 to work with the new engine framework.
13296 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
13297 make the newer ENGINE framework commands for the CHIL engine work.
13299 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
13301 * Make it possible to produce shared libraries on ReliantUNIX.
13303 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
13305 * Add the configuration target debug-linux-ppro.
13306 Make 'openssl rsa' use the general key loading routines
13307 implemented in `apps.c`, and make those routines able to
13308 handle the key format FORMAT_NETSCAPE and the variant
13311 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13313 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13315 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13317 * Add -keyform to rsautl, and document -engine.
13319 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
13321 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
13322 BIO_R_NO_SUCH_FILE error code rather than the generic
13323 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
13327 * Add new functions
13328 ERR_peek_last_error
13329 ERR_peek_last_error_line
13330 ERR_peek_last_error_line_data.
13331 These are similar to
13333 ERR_peek_error_line
13334 ERR_peek_error_line_data,
13335 but report on the latest error recorded rather than the first one
13336 still in the error queue.
13338 *Ben Laurie, Bodo Moeller*
13340 * default_algorithms option in ENGINE config module. This allows things
13342 default_algorithms = ALL
13343 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
13347 * Preliminary ENGINE config module.
13351 * New experimental application configuration code.
13355 * Change the AES code to follow the same name structure as all other
13356 symmetric ciphers, and behave the same way. Move everything to
13357 the directory crypto/aes, thereby obsoleting crypto/rijndael.
13359 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13361 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
13363 *Ben Laurie and Theo de Raadt*
13365 * Add option to output public keys in req command.
13367 *Massimiliano Pala madwolf@openca.org*
13369 * Use wNAFs in EC_POINTs_mul() for improved efficiency
13370 (up to about 10% better than before for P-192 and P-224).
13374 * New functions/macros
13376 SSL_CTX_set_msg_callback(ctx, cb)
13377 SSL_CTX_set_msg_callback_arg(ctx, arg)
13378 SSL_set_msg_callback(ssl, cb)
13379 SSL_set_msg_callback_arg(ssl, arg)
13381 to request calling a callback function
13383 void cb(int write_p, int version, int content_type,
13384 const void *buf, size_t len, SSL *ssl, void *arg)
13386 whenever a protocol message has been completely received
13387 (write_p == 0) or sent (write_p == 1). Here 'version' is the
13388 protocol version according to which the SSL library interprets
13389 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
13390 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
13391 the content type as defined in the SSL 3.0/TLS 1.0 protocol
13392 specification (change_cipher_spec(20), alert(21), handshake(22)).
13393 'buf' and 'len' point to the actual message, 'ssl' to the
13394 SSL object, and 'arg' is the application-defined value set by
13395 SSL[_CTX]_set_msg_callback_arg().
13397 'openssl s_client' and 'openssl s_server' have new '-msg' options
13398 to enable a callback that displays all protocol messages.
13402 * Change the shared library support so shared libraries are built as
13403 soon as the corresponding static library is finished, and thereby get
13404 openssl and the test programs linked against the shared library.
13405 This still only happens when the keyword "shard" has been given to
13406 the configuration scripts.
13408 NOTE: shared library support is still an experimental thing, and
13409 backward binary compatibility is still not guaranteed.
13411 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13413 * Add support for Subject Information Access extension.
13415 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13417 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13418 additional bytes when new memory had to be allocated, not just
13419 when reusing an existing buffer.
13423 * New command line and configuration option 'utf8' for the req command.
13424 This allows field values to be specified as UTF8 strings.
13428 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13429 runs for the former and machine-readable output for the latter.
13433 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13434 of the e-mail address in the DN (i.e., it will go into a certificate
13435 extension only). The new configuration file option 'email_in_dn = no'
13436 has the same effect.
13438 *Massimiliano Pala madwolf@openca.org*
13440 * Change all functions with names starting with `des_` to be starting
13441 with `DES_` instead. Add wrappers that are compatible with libdes,
13442 but are named `_ossl_old_des_*`. Finally, add macros that map the
13443 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13444 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13445 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13448 Since we provide two compatibility mappings, the user needs to
13449 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13450 compatibility is desired. The default (i.e., when that macro
13451 isn't defined) is OpenSSL 0.9.6c compatibility.
13453 There are also macros that enable and disable the support of old
13454 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13455 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13456 are defined, the default will apply: to support the old des routines.
13458 In either case, one must include openssl/des.h to get the correct
13459 definitions. Do not try to just include openssl/des_old.h, that
13462 NOTE: This is a major break of an old API into a new one. Software
13463 authors are encouraged to switch to the `DES_` style functions. Some
13464 time in the future, des_old.h and the libdes compatibility functions
13465 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13466 default), and then completely removed.
13470 * Test for certificates which contain unsupported critical extensions.
13471 If such a certificate is found during a verify operation it is
13472 rejected by default: this behaviour can be overridden by either
13473 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13474 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13475 X509_supported_extension() has also been added which returns 1 if a
13476 particular extension is supported.
13480 * Modify the behaviour of EVP cipher functions in similar way to digests
13481 to retain compatibility with existing code.
13485 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13486 compatibility with existing code. In particular the 'ctx' parameter does
13487 not have to be to be initialized before the call to EVP_DigestInit() and
13488 it is tidied up after a call to EVP_DigestFinal(). New function
13489 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13490 EVP_MD_CTX_copy() changed to not require the destination to be
13491 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13492 requires the destination to be valid.
13494 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13495 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13499 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13500 so that complete 'Handshake' protocol structures are kept in memory
13501 instead of overwriting 'msg_type' and 'length' with 'body' data.
13505 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13507 *Massimo Santin via Richard Levitte*
13509 * Major restructuring to the underlying ENGINE code. This includes
13510 reduction of linker bloat, separation of pure "ENGINE" manipulation
13511 (initialisation, etc) from functionality dealing with implementations
13512 of specific crypto interfaces. This change also introduces integrated
13513 support for symmetric ciphers and digest implementations - so ENGINEs
13514 can now accelerate these by providing EVP_CIPHER and EVP_MD
13515 implementations of their own. This is detailed in
13516 [crypto/engine/README.md](crypto/engine/README.md)
13517 as it couldn't be adequately described here. However, there are a few
13518 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13519 were changed in the original introduction of ENGINE code have now
13520 reverted back - the hooking from this code to ENGINE is now a good
13521 deal more passive and at run-time, operations deal directly with
13522 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13523 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13524 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13525 they were not being used by the framework as there is no concept of a
13526 BIGNUM_METHOD and they could not be generalised to the new
13527 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13528 ENGINE_cpy() has been removed as it cannot be consistently defined in
13533 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13537 * Change mkdef.pl to sort symbols that get the same entry number,
13538 and make sure the automatically generated functions `ERR_load_*`
13539 become part of libeay.num as well.
13543 * New function SSL_renegotiate_pending(). This returns true once
13544 renegotiation has been requested (either SSL_renegotiate() call
13545 or HelloRequest/ClientHello received from the peer) and becomes
13546 false once a handshake has been completed.
13547 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13548 sends a HelloRequest, but does not ensure that a handshake takes
13549 place. SSL_renegotiate_pending() is useful for checking if the
13550 client has followed the request.)
13554 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13555 By default, clients may request session resumption even during
13556 renegotiation (if session ID contexts permit); with this option,
13557 session resumption is possible only in the first handshake.
13559 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13560 more bits available for options that should not be part of
13561 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13565 * Add some demos for certificate and certificate request creation.
13569 * Make maximum certificate chain size accepted from the peer application
13570 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13571 "Douglas E. Engert" <deengert@anl.gov>.
13575 * Add support for shared libraries for Unixware-7
13576 (Boyd Lynn Gerber <gerberb@zenez.com>).
13580 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13581 be done prior to destruction. Use this to unload error strings from
13582 ENGINEs that load their own error strings. NB: This adds two new API
13583 functions to "get" and "set" this destroy handler in an ENGINE.
13587 * Alter all existing ENGINE implementations (except "openssl" and
13588 "openbsd") to dynamically instantiate their own error strings. This
13589 makes them more flexible to be built both as statically-linked ENGINEs
13590 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13591 Also, add stub code to each that makes building them as self-contained
13592 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13596 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13597 implementations into applications that are completely implemented in
13598 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13599 commands that can be used to configure what shared-library to load and
13600 to control aspects of the way it is handled. Also, made an update to
13601 the [README-Engine.md](README-Engine.md) file
13602 that brings its information up-to-date and
13603 provides some information and instructions on the "dynamic" ENGINE
13604 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13608 * Make it possible to unload ranges of ERR strings with a new
13609 "ERR_unload_strings" function.
13613 * Add a copy() function to EVP_MD.
13617 * Make EVP_MD routines take a context pointer instead of just the
13618 md_data void pointer.
13622 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13623 that the digest can only process a single chunk of data
13624 (typically because it is provided by a piece of
13625 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13626 is only going to provide a single chunk of data, and hence the
13627 framework needn't accumulate the data for oneshot drivers.
13631 * As with "ERR", make it possible to replace the underlying "ex_data"
13632 functions. This change also alters the storage and management of global
13633 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13634 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13635 index counters. The API functions that use this state have been changed
13636 to take a "class_index" rather than pointers to the class's local STACK
13637 and counter, and there is now an API function to dynamically create new
13638 classes. This centralisation allows us to (a) plug a lot of the
13639 thread-safety problems that existed, and (b) makes it possible to clean
13640 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13641 such data would previously have always leaked in application code and
13642 workarounds were in place to make the memory debugging turn a blind eye
13643 to it. Application code that doesn't use this new function will still
13644 leak as before, but their memory debugging output will announce it now
13645 rather than letting it slide.
13647 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13648 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13649 has a return value to indicate success or failure.
13653 * Make it possible to replace the underlying "ERR" functions such that the
13654 global state (2 LHASH tables and 2 locks) is only used by the "default"
13655 implementation. This change also adds two functions to "get" and "set"
13656 the implementation prior to it being automatically set the first time
13657 any other ERR function takes place. Ie. an application can call "get",
13658 pass the return value to a module it has just loaded, and that module
13659 can call its own "set" function using that value. This means the
13660 module's "ERR" operations will use (and modify) the error state in the
13661 application and not in its own statically linked copy of OpenSSL code.
13665 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13666 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13667 the operation, and provides a more encapsulated way for external code
13668 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13669 to use these functions rather than manually incrementing the counts.
13671 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13675 * Add EVP test program.
13679 * Add symmetric cipher support to ENGINE. Expect the API to change!
13683 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13684 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13685 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13686 These allow a CRL to be built without having to access X509_CRL fields
13687 directly. Modify 'ca' application to use new functions.
13691 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13692 bug workarounds. Rollback attack detection is a security feature.
13693 The problem will only arise on OpenSSL servers when TLSv1 is not
13694 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13695 Software authors not wanting to support TLSv1 will have special reasons
13696 for their choice and can explicitly enable this option.
13698 *Bodo Moeller, Lutz Jaenicke*
13700 * Rationalise EVP so it can be extended: don't include a union of
13701 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13702 (similar to those existing for EVP_CIPHER_CTX).
13707 EVP_MD_CTX_init(&md); /* new function call */
13708 EVP_DigestInit(&md, EVP_sha1());
13709 EVP_DigestUpdate(&md, in, len);
13710 EVP_DigestFinal(&md, out, NULL);
13711 EVP_MD_CTX_cleanup(&md); /* new function call */
13715 * Make DES key schedule conform to the usual scheme, as well as
13716 correcting its structure. This means that calls to DES functions
13717 now have to pass a pointer to a des_key_schedule instead of a
13718 plain des_key_schedule (which was actually always a pointer
13721 des_key_schedule ks;
13723 des_set_key_checked(..., &ks);
13724 des_ncbc_encrypt(..., &ks, ...);
13726 (Note that a later change renames 'des_...' into 'DES_...'.)
13730 * Initial reduction of linker bloat: the use of some functions, such as
13731 PEM causes large amounts of unused functions to be linked in due to
13732 poor organisation. For example pem_all.c contains every PEM function
13733 which has a knock on effect of linking in large amounts of (unused)
13734 ASN1 code. Grouping together similar functions and splitting unrelated
13735 functions prevents this.
13739 * Cleanup of EVP macros.
13743 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13744 correct `_ecb suffix`.
13748 * Add initial OCSP responder support to ocsp application. The
13749 revocation information is handled using the text based index
13750 use by the ca application. The responder can either handle
13751 requests generated internally, supplied in files (for example
13752 via a CGI script) or using an internal minimal server.
13756 * Add configuration choices to get zlib compression for TLS.
13760 * Changes to Kerberos SSL for RFC 2712 compliance:
13761 1. Implemented real KerberosWrapper, instead of just using
13762 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13763 2. Implemented optional authenticator field of KerberosWrapper.
13765 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13766 and authenticator structs; see crypto/krb5/.
13768 Generalized Kerberos calls to support multiple Kerberos libraries.
13769 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13770 via Richard Levitte*
13772 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13773 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13774 values for each of the key sizes rather than having just
13775 parameters (and 'speed' generating keys each time).
13779 * Speed up EVP routines.
13782 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13783 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13784 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13785 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13787 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13788 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13789 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13792 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13794 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13798 * Added the OS2-EMX target.
13800 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13802 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13803 New functions to support `NCONF` routines in extension code.
13804 New function `CONF_set_nconf()`
13805 to allow functions which take an `NCONF` to also handle the old `LHASH`
13806 structure: this means that the old `CONF` compatible routines can be
13807 retained (in particular w.rt. extensions) without having to duplicate the
13808 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13812 * Enhance the general user interface with mechanisms for inner control
13813 and with possibilities to have yes/no kind of prompts.
13817 * Change all calls to low-level digest routines in the library and
13818 applications to use EVP. Add missing calls to HMAC_cleanup() and
13819 don't assume HMAC_CTX can be copied using memcpy().
13821 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13823 * Add the possibility to control engines through control names but with
13824 arbitrary arguments instead of just a string.
13825 Change the key loaders to take a UI_METHOD instead of a callback
13826 function pointer. NOTE: this breaks binary compatibility with earlier
13827 versions of OpenSSL [engine].
13828 Adapt the nCipher code for these new conditions and add a card insertion
13833 * Enhance the general user interface with mechanisms to better support
13834 dialog box interfaces, application-defined prompts, the possibility
13835 to use defaults (for example default passwords from somewhere else)
13836 and interrupts/cancellations.
13840 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13841 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13845 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13846 tidy up some unnecessarily weird code in 'sk_new()').
13848 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13850 * Change the key loading routines for ENGINEs to use the same kind
13851 callback (pem_password_cb) as all other routines that need this
13856 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13857 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13858 than this minimum value is recommended.
13862 * New random seeder for OpenVMS, using the system process statistics
13863 that are easily reachable.
13867 * Windows apparently can't transparently handle global
13868 variables defined in DLLs. Initialisations such as:
13870 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13872 won't compile. This is used by the any applications that need to
13873 declare their own ASN1 modules. This was fixed by adding the option
13874 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13875 needed for static libraries under Win32.
13879 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13880 setting of purpose and trust fields. New X509_STORE trust and
13881 purpose functions and tidy up setting in other SSL functions.
13885 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13886 structure. These are inherited by X509_STORE_CTX when it is
13887 initialised. This allows various defaults to be set in the
13888 X509_STORE structure (such as flags for CRL checking and custom
13889 purpose or trust settings) for functions which only use X509_STORE_CTX
13890 internally such as S/MIME.
13892 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13893 trust settings if they are not set in X509_STORE. This allows X509_STORE
13894 purposes and trust (in S/MIME for example) to override any set by default.
13896 Add command line options for CRL checking to smime, s_client and s_server
13901 * Initial CRL based revocation checking. If the CRL checking flag(s)
13902 are set then the CRL is looked up in the X509_STORE structure and
13903 its validity and signature checked, then if the certificate is found
13904 in the CRL the verify fails with a revoked error.
13906 Various new CRL related callbacks added to X509_STORE_CTX structure.
13908 Command line options added to 'verify' application to support this.
13910 This needs some additional work, such as being able to handle multiple
13911 CRLs with different times, extension based lookup (rather than just
13912 by subject name) and ultimately more complete V2 CRL extension
13917 * Add a general user interface API (crypto/ui/). This is designed
13918 to replace things like des_read_password and friends (backward
13919 compatibility functions using this new API are provided).
13920 The purpose is to remove prompting functions from the DES code
13921 section as well as provide for prompting through dialog boxes in
13922 a window system and the like.
13926 * Add "ex_data" support to ENGINE so implementations can add state at a
13927 per-structure level rather than having to store it globally.
13931 * Make it possible for ENGINE structures to be copied when retrieved by
13932 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13933 This causes the "original" ENGINE structure to act like a template,
13934 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13935 operational state can be localised to each ENGINE structure, despite the
13936 fact they all share the same "methods". New ENGINE structures returned in
13937 this case have no functional references and the return value is the single
13938 structural reference. This matches the single structural reference returned
13939 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13944 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13945 needs to match any other type at all we need to manually clear the
13950 * Changes to the "openssl engine" utility to include;
13951 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13952 about an ENGINE's available control commands.
13953 - executing control commands from command line arguments using the
13954 '-pre' and '-post' switches. '-post' is only used if '-t' is
13955 specified and the ENGINE is successfully initialised. The syntax for
13956 the individual commands are colon-separated, for example;
13957 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13961 * New dynamic control command support for ENGINEs. ENGINEs can now
13962 declare their own commands (numbers), names (strings), descriptions,
13963 and input types for run-time discovery by calling applications. A
13964 subset of these commands are implicitly classed as "executable"
13965 depending on their input type, and only these can be invoked through
13966 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13967 can be based on user input, config files, etc). The distinction is
13968 that "executable" commands cannot return anything other than a boolean
13969 result and can only support numeric or string input, whereas some
13970 discoverable commands may only be for direct use through
13971 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13972 pointers, or other custom uses. The "executable" commands are to
13973 support parameterisations of ENGINE behaviour that can be
13974 unambiguously defined by ENGINEs and used consistently across any
13975 OpenSSL-based application. Commands have been added to all the
13976 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13977 control over shared-library paths without source code alterations.
13981 * Changed all ENGINE implementations to dynamically allocate their
13982 ENGINEs rather than declaring them statically. Apart from this being
13983 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13984 this also allows the implementations to compile without using the
13985 internal engine_int.h header.
13989 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13990 'const' value. Any code that should be able to modify a RAND_METHOD
13991 should already have non-const pointers to it (ie. they should only
13992 modify their own ones).
13996 * Made a variety of little tweaks to the ENGINE code.
13997 - "atalla" and "ubsec" string definitions were moved from header files
13998 to C code. "nuron" string definitions were placed in variables
13999 rather than hard-coded - allowing parameterisation of these values
14000 later on via ctrl() commands.
14001 - Removed unused "#if 0"'d code.
14002 - Fixed engine list iteration code so it uses ENGINE_free() to release
14003 structural references.
14004 - Constified the RAND_METHOD element of ENGINE structures.
14005 - Constified various get/set functions as appropriate and added
14006 missing functions (including a catch-all ENGINE_cpy that duplicates
14007 all ENGINE values onto a new ENGINE except reference counts/state).
14008 - Removed NULL parameter checks in get/set functions. Setting a method
14009 or function to NULL is a way of cancelling out a previously set
14010 value. Passing a NULL ENGINE parameter is just plain stupid anyway
14011 and doesn't justify the extra error symbols and code.
14012 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
14013 flags from engine_int.h to engine.h.
14014 - Changed prototypes for ENGINE handler functions (init(), finish(),
14015 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
14019 * Implement binary inversion algorithm for BN_mod_inverse in addition
14020 to the algorithm using long division. The binary algorithm can be
14021 used only if the modulus is odd. On 32-bit systems, it is faster
14022 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
14023 roughly 5-15% for 256-bit moduli), so we use it only for moduli
14024 up to 450 bits. In 64-bit environments, the binary algorithm
14025 appears to be advantageous for much longer moduli; here we use it
14026 for moduli up to 2048 bits.
14030 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
14031 could not support the combine flag in choice fields.
14035 * Add a 'copy_extensions' option to the 'ca' utility. This copies
14036 extensions from a certificate request to the certificate.
14040 * Allow multiple 'certopt' and 'nameopt' options to be separated
14041 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
14042 file: this allows the display of the certificate about to be
14043 signed to be customised, to allow certain fields to be included
14044 or excluded and extension details. The old system didn't display
14045 multicharacter strings properly, omitted fields not in the policy
14046 and couldn't display additional details such as extensions.
14050 * Function EC_POINTs_mul for multiple scalar multiplication
14051 of an arbitrary number of elliptic curve points
14052 \sum scalars[i]*points[i],
14053 optionally including the generator defined for the EC_GROUP:
14054 scalar*generator + \sum scalars[i]*points[i].
14056 EC_POINT_mul is a simple wrapper function for the typical case
14057 that the point list has just one item (besides the optional
14062 * First EC_METHODs for curves over GF(p):
14064 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
14065 operations and provides various method functions that can also
14066 operate with faster implementations of modular arithmetic.
14068 EC_GFp_mont_method() reuses most functions that are part of
14069 EC_GFp_simple_method, but uses Montgomery arithmetic.
14071 *Bodo Moeller; point addition and point doubling
14072 implementation directly derived from source code provided by
14073 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
14075 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
14076 crypto/ec/ec_lib.c):
14078 Curves are EC_GROUP objects (with an optional group generator)
14079 based on EC_METHODs that are built into the library.
14081 Points are EC_POINT objects based on EC_GROUP objects.
14083 Most of the framework would be able to handle curves over arbitrary
14084 finite fields, but as there are no obvious types for fields other
14085 than GF(p), some functions are limited to that for now.
14089 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
14090 that the file contains a complete HTTP response.
14094 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
14095 change the def and num file printf format specifier from "%-40sXXX"
14096 to "%-39s XXX". The latter will always guarantee a space after the
14097 field while the former will cause them to run together if the field
14098 is 40 of more characters long.
14102 * Constify the cipher and digest 'method' functions and structures
14103 and modify related functions to take constant EVP_MD and EVP_CIPHER
14108 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
14109 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
14113 * Modify `EVP_Digest*()` routines so they now return values. Although the
14114 internal software routines can never fail additional hardware versions
14119 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
14121 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
14122 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
14125 ERR_R_NESTED_ASN1_ERROR
14127 ERR_R_MISSING_ASN1_EOS
14128 were 4 .. 9, conflicting with
14129 ERR_LIB_RSA (= ERR_R_RSA_LIB)
14131 ERR_LIB_PEM (= ERR_R_PEM_LIB).
14132 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
14134 Add new error code 'ERR_R_INTERNAL_ERROR'.
14138 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
14143 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
14144 sets the subject name for a new request or supersedes the
14145 subject name in a given request. Formats that can be parsed are
14146 'CN=Some Name, OU=myOU, C=IT'
14148 'CN=Some Name/OU=myOU/C=IT'.
14150 Add options '-batch' and '-verbose' to 'openssl req'.
14152 *Massimiliano Pala <madwolf@hackmasters.net>*
14154 * Introduce the possibility to access global variables through
14155 functions on platform were that's the best way to handle exporting
14156 global variables in shared libraries. To enable this functionality,
14157 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
14158 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
14159 is normally done by Configure or something similar).
14161 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
14162 in the source file (foo.c) like this:
14164 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
14165 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
14167 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
14168 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
14170 OPENSSL_DECLARE_GLOBAL(int,foo);
14171 #define foo OPENSSL_GLOBAL_REF(foo)
14172 OPENSSL_DECLARE_GLOBAL(double,bar);
14173 #define bar OPENSSL_GLOBAL_REF(bar)
14175 The #defines are very important, and therefore so is including the
14176 header file everywhere where the defined globals are used.
14178 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
14179 of ASN.1 items, but that structure is a bit different.
14181 The largest change is in util/mkdef.pl which has been enhanced with
14182 better and easier to understand logic to choose which symbols should
14183 go into the Windows .def files as well as a number of fixes and code
14184 cleanup (among others, algorithm keywords are now sorted
14185 lexicographically to avoid constant rewrites).
14189 * In BN_div() keep a copy of the sign of 'num' before writing the
14190 result to 'rm' because if rm==num the value will be overwritten
14191 and produce the wrong result if 'num' is negative: this caused
14192 problems with BN_mod() and BN_nnmod().
14196 * Function OCSP_request_verify(). This checks the signature on an
14197 OCSP request and verifies the signer certificate. The signer
14198 certificate is just checked for a generic purpose and OCSP request
14203 * Add OCSP_check_validity() function to check the validity of OCSP
14204 responses. OCSP responses are prepared in real time and may only
14205 be a few seconds old. Simply checking that the current time lies
14206 between thisUpdate and nextUpdate max reject otherwise valid responses
14207 caused by either OCSP responder or client clock inaccuracy. Instead
14208 we allow thisUpdate and nextUpdate to fall within a certain period of
14209 the current time. The age of the response can also optionally be
14210 checked. Two new options -validity_period and -status_age added to
14215 * If signature or public key algorithm is unrecognized print out its
14216 OID rather that just UNKNOWN.
14220 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
14221 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
14222 ID to be generated from the issuer certificate alone which can then be
14223 passed to OCSP_id_issuer_cmp().
14227 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
14228 ASN1 modules to export functions returning ASN1_ITEM pointers
14229 instead of the ASN1_ITEM structures themselves. This adds several
14230 new macros which allow the underlying ASN1 function/structure to
14231 be accessed transparently. As a result code should not use ASN1_ITEM
14232 references directly (such as &X509_it) but instead use the relevant
14233 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
14234 use of the new ASN1 code on platforms where exporting structures
14235 is problematical (for example in shared libraries) but exporting
14236 functions returning pointers to structures is not.
14240 * Add support for overriding the generation of SSL/TLS session IDs.
14241 These callbacks can be registered either in an SSL_CTX or per SSL.
14242 The purpose of this is to allow applications to control, if they wish,
14243 the arbitrary values chosen for use as session IDs, particularly as it
14244 can be useful for session caching in multiple-server environments. A
14245 command-line switch for testing this (and any client code that wishes
14246 to use such a feature) has been added to "s_server".
14248 *Geoff Thorpe, Lutz Jaenicke*
14250 * Modify mkdef.pl to recognise and parse preprocessor conditionals
14251 of the form `#if defined(...) || defined(...) || ...` and
14252 `#if !defined(...) && !defined(...) && ...`. This also avoids
14253 the growing number of special cases it was previously handling.
14257 * Make all configuration macros available for application by making
14258 sure they are available in opensslconf.h, by giving them names starting
14259 with `OPENSSL_` to avoid conflicts with other packages and by making
14260 sure e_os2.h will cover all platform-specific cases together with
14262 Additionally, it is now possible to define configuration/platform-
14263 specific names (called "system identities"). In the C code, these
14264 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
14265 macro with the name beginning with `OPENSSL_SYS_`, which is determined
14266 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
14271 * New option -set_serial to 'req' and 'x509' this allows the serial
14272 number to use to be specified on the command line. Previously self
14273 signed certificates were hard coded with serial number 0 and the
14274 CA options of 'x509' had to use a serial number in a file which was
14279 * New options to 'ca' utility to support V2 CRL entry extensions.
14280 Currently CRL reason, invalidity date and hold instruction are
14281 supported. Add new CRL extensions to V3 code and some new objects.
14285 * New function EVP_CIPHER_CTX_set_padding() this is used to
14286 disable standard block padding (aka PKCS#5 padding) in the EVP
14287 API, which was previously mandatory. This means that the data is
14288 not padded in any way and so the total length much be a multiple
14289 of the block size, otherwise an error occurs.
14293 * Initial (incomplete) OCSP SSL support.
14297 * New function OCSP_parse_url(). This splits up a URL into its host,
14298 port and path components: primarily to parse OCSP URLs. New -url
14299 option to ocsp utility.
14303 * New nonce behavior. The return value of OCSP_check_nonce() now
14304 reflects the various checks performed. Applications can decide
14305 whether to tolerate certain situations such as an absent nonce
14306 in a response when one was present in a request: the ocsp application
14307 just prints out a warning. New function OCSP_add1_basic_nonce()
14308 this is to allow responders to include a nonce in a response even if
14309 the request is nonce-less.
14313 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
14314 skipped when using openssl x509 multiple times on a single input file,
14315 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
14319 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
14320 set string type: to handle setting ASN1_TIME structures. Fix ca
14321 utility to correctly initialize revocation date of CRLs.
14325 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
14326 the clients preferred ciphersuites and rather use its own preferences.
14327 Should help to work around M$ SGC (Server Gated Cryptography) bug in
14328 Internet Explorer by ensuring unchanged hash method during stepup.
14329 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
14333 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
14334 to aes and add a new 'exist' option to print out symbols that don't
14339 * Additional options to ocsp utility to allow flags to be set and
14340 additional certificates supplied.
14344 * Add the option -VAfile to 'openssl ocsp', so the user can give the
14345 OCSP client a number of certificate to only verify the response
14350 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
14351 handle the new API. Currently only ECB, CBC modes supported. Add new
14354 Add TLS AES ciphersuites as described in RFC3268, "Advanced
14355 Encryption Standard (AES) Ciphersuites for Transport Layer
14356 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
14357 not enabled by default and were not part of the "ALL" ciphersuite
14358 alias because they were not yet official; they could be
14359 explicitly requested by specifying the "AESdraft" ciphersuite
14360 group alias. In the final release of OpenSSL 0.9.7, the group
14361 alias is called "AES" and is part of "ALL".)
14363 *Ben Laurie, Steve Henson, Bodo Moeller*
14365 * New function OCSP_copy_nonce() to copy nonce value (if present) from
14366 request to response.
14370 * Functions for OCSP responders. OCSP_request_onereq_count(),
14371 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
14372 extract information from a certificate request. OCSP_response_create()
14373 creates a response and optionally adds a basic response structure.
14374 OCSP_basic_add1_status() adds a complete single response to a basic
14375 response and returns the OCSP_SINGLERESP structure just added (to allow
14376 extensions to be included for example). OCSP_basic_add1_cert() adds a
14377 certificate to a basic response and OCSP_basic_sign() signs a basic
14378 response with various flags. New helper functions ASN1_TIME_check()
14379 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
14380 (converts ASN1_TIME to GeneralizedTime).
14384 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
14385 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
14386 structure from a certificate. X509_pubkey_digest() digests the public_key
14387 contents: this is used in various key identifiers.
14391 * Make sk_sort() tolerate a NULL argument.
14393 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14395 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14396 passed by the function are trusted implicitly. If any of them signed the
14397 response then it is assumed to be valid and is not verified.
14401 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14402 to data. This was previously part of the PKCS7 ASN1 code. This
14403 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14404 *Steve Henson, reported by Kenneth R. Robinette
14405 <support@securenetterm.com>*
14407 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14408 routines: without these tracing memory leaks is very painful.
14409 Fix leaks in PKCS12 and PKCS7 routines.
14413 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14414 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14415 effectively meant GeneralizedTime would never be used. Now it
14416 is initialised to -1 but X509_time_adj() now has to check the value
14417 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14418 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14419 *Steve Henson, reported by Kenneth R. Robinette
14420 <support@securenetterm.com>*
14422 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14423 result in a zero length in the ASN1_INTEGER structure which was
14424 not consistent with the structure when d2i_ASN1_INTEGER() was used
14425 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14426 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14427 where it did not print out a minus for negative ASN1_INTEGER.
14431 * Add summary printout to ocsp utility. The various functions which
14432 convert status values to strings have been renamed to:
14433 OCSP_response_status_str(), OCSP_cert_status_str() and
14434 OCSP_crl_reason_str() and are no longer static. New options
14435 to verify nonce values and to disable verification. OCSP response
14436 printout format cleaned up.
14440 * Add additional OCSP certificate checks. These are those specified
14441 in RFC2560. This consists of two separate checks: the CA of the
14442 certificate being checked must either be the OCSP signer certificate
14443 or the issuer of the OCSP signer certificate. In the latter case the
14444 OCSP signer certificate must contain the OCSP signing extended key
14445 usage. This check is performed by attempting to match the OCSP
14446 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14447 in the OCSP_CERTID structures of the response.
14451 * Initial OCSP certificate verification added to OCSP_basic_verify()
14452 and related routines. This uses the standard OpenSSL certificate
14453 verify routines to perform initial checks (just CA validity) and
14454 to obtain the certificate chain. Then additional checks will be
14455 performed on the chain. Currently the root CA is checked to see
14456 if it is explicitly trusted for OCSP signing. This is used to set
14457 a root CA as a global signing root: that is any certificate that
14458 chains to that CA is an acceptable OCSP signing certificate.
14462 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14463 extensions from a separate configuration file.
14464 As when reading extensions from the main configuration file,
14465 the '-extensions ...' option may be used for specifying the
14468 *Massimiliano Pala <madwolf@comune.modena.it>*
14470 * New OCSP utility. Allows OCSP requests to be generated or
14471 read. The request can be sent to a responder and the output
14472 parsed, outputted or printed in text form. Not complete yet:
14473 still needs to check the OCSP response validity.
14477 * New subcommands for 'openssl ca':
14478 `openssl ca -status <serial>` prints the status of the cert with
14479 the given serial number (according to the index file).
14480 `openssl ca -updatedb` updates the expiry status of certificates
14483 *Massimiliano Pala <madwolf@comune.modena.it>*
14485 * New '-newreq-nodes' command option to CA.pl. This is like
14486 '-newreq', but calls 'openssl req' with the '-nodes' option
14487 so that the resulting key is not encrypted.
14489 *Damien Miller <djm@mindrot.org>*
14491 * New configuration for the GNU Hurd.
14493 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14495 * Initial code to implement OCSP basic response verify. This
14496 is currently incomplete. Currently just finds the signer's
14497 certificate and verifies the signature on the response.
14501 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14502 value of OPENSSLDIR. This is available via the new '-d' option
14503 to 'openssl version', and is also included in 'openssl version -a'.
14507 * Allowing defining memory allocation callbacks that will be given
14508 file name and line number information in additional arguments
14509 (a `const char*` and an int). The basic functionality remains, as
14510 well as the original possibility to just replace malloc(),
14511 realloc() and free() by functions that do not know about these
14512 additional arguments. To register and find out the current
14513 settings for extended allocation functions, the following
14514 functions are provided:
14516 CRYPTO_set_mem_ex_functions
14517 CRYPTO_set_locked_mem_ex_functions
14518 CRYPTO_get_mem_ex_functions
14519 CRYPTO_get_locked_mem_ex_functions
14521 These work the same way as CRYPTO_set_mem_functions and friends.
14522 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14523 extended allocation function is enabled.
14524 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14525 a conventional allocation function is enabled.
14527 *Richard Levitte, Bodo Moeller*
14529 * Finish off removing the remaining LHASH function pointer casts.
14530 There should no longer be any prototype-casting required when using
14531 the LHASH abstraction, and any casts that remain are "bugs". See
14532 the callback types and macros at the head of lhash.h for details
14533 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14537 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14538 If /dev/[u]random devices are not available or do not return enough
14539 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14541 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14542 /etc/entropy will be queried once each in this sequence, querying stops
14543 when enough entropy was collected without querying more sockets.
14547 * Change the Unix RAND_poll() variant to be able to poll several
14548 random devices, as specified by DEVRANDOM, until a sufficient amount
14549 of data has been collected. We spend at most 10 ms on each file
14550 (select timeout) and read in non-blocking mode. DEVRANDOM now
14551 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14552 (previously it was just the string "/dev/urandom"), so on typical
14553 platforms the 10 ms delay will never occur.
14554 Also separate out the Unix variant to its own file, rand_unix.c.
14555 For VMS, there's a currently-empty rand_vms.c.
14559 * Move OCSP client related routines to ocsp_cl.c. These
14560 provide utility functions which an application needing
14561 to issue a request to an OCSP responder and analyse the
14562 response will typically need: as opposed to those which an
14563 OCSP responder itself would need which will be added later.
14565 OCSP_request_sign() signs an OCSP request with an API similar
14566 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14567 response. OCSP_response_get1_basic() extracts basic response
14568 from response. OCSP_resp_find_status(): finds and extracts status
14569 information from an OCSP_CERTID structure (which will be created
14570 when the request structure is built). These are built from lower
14571 level functions which work on OCSP_SINGLERESP structures but
14572 won't normally be used unless the application wishes to examine
14573 extensions in the OCSP response for example.
14575 Replace nonce routines with a pair of functions.
14576 OCSP_request_add1_nonce() adds a nonce value and optionally
14577 generates a random value. OCSP_check_nonce() checks the
14578 validity of the nonce in an OCSP response.
14582 * Change function OCSP_request_add() to OCSP_request_add0_id().
14583 This doesn't copy the supplied OCSP_CERTID and avoids the
14584 need to free up the newly created id. Change return type
14585 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14586 This can then be used to add extensions to the request.
14587 Deleted OCSP_request_new(), since most of its functionality
14588 is now in OCSP_REQUEST_new() (and the case insensitive name
14589 clash) apart from the ability to set the request name which
14590 will be added elsewhere.
14594 * Update OCSP API. Remove obsolete extensions argument from
14595 various functions. Extensions are now handled using the new
14596 OCSP extension code. New simple OCSP HTTP function which
14597 can be used to send requests and parse the response.
14601 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14602 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14603 uses the special reorder version of SET OF to sort the attributes
14604 and reorder them to match the encoded order. This resolves a long
14605 standing problem: a verify on a PKCS7 structure just after signing
14606 it used to fail because the attribute order did not match the
14607 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14608 it uses the received order. This is necessary to tolerate some broken
14609 software that does not order SET OF. This is handled by encoding
14610 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14611 to produce the required SET OF.
14615 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14616 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14617 files to get correct declarations of the ASN.1 item variables.
14621 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14622 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14623 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14624 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14625 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14626 ASN1_ITEM and no wrapper functions.
14630 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14631 replace the old function pointer based I/O routines. Change most of
14632 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14636 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14637 lines, recognize more "algorithms" that can be deselected, and make
14638 it complain about algorithm deselection that isn't recognised.
14642 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14643 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14644 to use new functions. Add NO_ASN1_OLD which can be set to remove
14645 some old style ASN1 functions: this can be used to determine if old
14646 code will still work when these eventually go away.
14650 * New extension functions for OCSP structures, these follow the
14651 same conventions as certificates and CRLs.
14655 * New function X509V3_add1_i2d(). This automatically encodes and
14656 adds an extension. Its behaviour can be customised with various
14657 flags to append, replace or delete. Various wrappers added for
14658 certificates and CRLs.
14662 * Fix to avoid calling the underlying ASN1 print routine when
14663 an extension cannot be parsed. Correct a typo in the
14664 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14668 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14669 entries for variables.
14673 * Add functionality to `apps/openssl.c` for detecting locking
14674 problems: As the program is single-threaded, all we have
14675 to do is register a locking callback using an array for
14676 storing which locks are currently held by the program.
14680 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14681 SSL_get_ex_data_X509_STORE_idx(), which is used in
14682 ssl_verify_cert_chain() and thus can be called at any time
14683 during TLS/SSL handshakes so that thread-safety is essential.
14684 Unfortunately, the ex_data design is not at all suited
14685 for multi-threaded use, so it probably should be abolished.
14689 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14691 *Broadcom, tweaked and integrated by Geoff Thorpe*
14693 * Move common extension printing code to new function
14694 X509V3_print_extensions(). Reorganise OCSP print routines and
14695 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14699 * New function X509_signature_print() to remove duplication in some
14704 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14705 set (this was treated exactly the same as SET OF previously). This
14706 is used to reorder the STACK representing the structure to match the
14707 encoding. This will be used to get round a problem where a PKCS7
14708 structure which was signed could not be verified because the STACK
14709 order did not reflect the encoded order.
14713 * Reimplement the OCSP ASN1 module using the new code.
14717 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14718 for its ASN1 operations. The old style function pointers still exist
14719 for now but they will eventually go away.
14723 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14724 completely replaces the old ASN1 functionality with a table driven
14725 encoder and decoder which interprets an ASN1_ITEM structure describing
14726 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14727 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14728 has also been converted to the new form.
14732 * Change BN_mod_exp_recp so that negative moduli are tolerated
14733 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14734 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14735 for negative moduli.
14739 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14740 of not touching the result's sign bit.
14744 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14749 * Changed the LHASH code to use prototypes for callbacks, and created
14750 macros to declare and implement thin (optionally static) functions
14751 that provide type-safety and avoid function pointer casting for the
14752 type-specific callbacks.
14756 * Added Kerberos Cipher Suites to be used with TLS, as written in
14758 *Veers Staats <staatsvr@asc.hpc.mil>,
14759 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14761 * Reformat the FAQ so the different questions and answers can be divided
14762 in sections depending on the subject.
14766 * Have the zlib compression code load ZLIB.DLL dynamically under
14771 * New function BN_mod_sqrt for computing square roots modulo a prime
14772 (using the probabilistic Tonelli-Shanks algorithm unless
14773 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14774 be handled deterministically).
14776 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14778 * Make BN_mod_inverse faster by explicitly handling small quotients
14779 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14780 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14784 * New function BN_kronecker.
14788 * Fix BN_gcd so that it works on negative inputs; the result is
14789 positive unless both parameters are zero.
14790 Previously something reasonably close to an infinite loop was
14791 possible because numbers could be growing instead of shrinking
14792 in the implementation of Euclid's algorithm.
14796 * Fix BN_is_word() and BN_is_one() macros to take into account the
14797 sign of the number in question.
14799 Fix BN_is_word(a,w) to work correctly for w == 0.
14801 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14802 because its test if the absolute value of 'a' equals 'w'.
14803 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14804 it exists mostly for use in the implementations of BN_is_zero(),
14805 BN_is_one(), and BN_is_word().
14809 * New function BN_swap.
14813 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14814 the exponentiation functions are more likely to produce reasonable
14815 results on negative inputs.
14819 * Change BN_mod_mul so that the result is always non-negative.
14820 Previously, it could be negative if one of the factors was negative;
14821 I don't think anyone really wanted that behaviour.
14825 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14826 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14827 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14828 and add new functions:
14837 BN_mod_lshift1_quick
14839 BN_mod_lshift_quick
14841 These functions always generate non-negative results.
14843 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14844 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14846 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14847 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14848 be reduced modulo `m`.
14850 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14853 The following entry accidentally appeared in the CHANGES file
14854 distributed with OpenSSL 0.9.7. The modifications described in
14855 it do *not* apply to OpenSSL 0.9.7.
14857 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14858 was actually never needed) and in BN_mul(). The removal in BN_mul()
14859 required a small change in bn_mul_part_recursive() and the addition
14860 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14861 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14862 bn_sub_words() and bn_add_words() except they take arrays with
14868 * In 'openssl passwd', verify passwords read from the terminal
14869 unless the '-salt' option is used (which usually means that
14870 verification would just waste user's time since the resulting
14871 hash is going to be compared with some given password hash)
14872 or the new '-noverify' option is used.
14874 This is an incompatible change, but it does not affect
14875 non-interactive use of 'openssl passwd' (passwords on the command
14876 line, '-stdin' option, '-in ...' option) and thus should not
14877 cause any problems.
14881 * Remove all references to RSAref, since there's no more need for it.
14885 * Make DSO load along a path given through an environment variable
14886 (SHLIB_PATH) with shl_load().
14890 * Constify the ENGINE code as a result of BIGNUM constification.
14891 Also constify the RSA code and most things related to it. In a
14892 few places, most notable in the depth of the ASN.1 code, ugly
14893 casts back to non-const were required (to be solved at a later
14898 * Make it so the openssl application has all engines loaded by default.
14902 * Constify the BIGNUM routines a little more.
14906 * Add the following functions:
14908 ENGINE_load_cswift()
14910 ENGINE_load_atalla()
14911 ENGINE_load_nuron()
14912 ENGINE_load_builtin_engines()
14914 That way, an application can itself choose if external engines that
14915 are built-in in OpenSSL shall ever be used or not. The benefit is
14916 that applications won't have to be linked with libdl or other dso
14917 libraries unless it's really needed.
14919 Changed 'openssl engine' to load all engines on demand.
14920 Changed the engine header files to avoid the duplication of some
14921 declarations (they differed!).
14925 * 'openssl engine' can now list capabilities.
14929 * Better error reporting in 'openssl engine'.
14933 * Never call load_dh_param(NULL) in s_server.
14937 * Add engine application. It can currently list engines by name and
14938 identity, and test if they are actually available.
14942 * Improve RPM specification file by forcing symbolic linking and making
14943 sure the installed documentation is also owned by root.root.
14945 *Damien Miller <djm@mindrot.org>*
14947 * Give the OpenSSL applications more possibilities to make use of
14948 keys (public as well as private) handled by engines.
14952 * Add OCSP code that comes from CertCo.
14956 * Add VMS support for the Rijndael code.
14960 * Added untested support for Nuron crypto accelerator.
14964 * Add support for external cryptographic devices. This code was
14965 previously distributed separately as the "engine" branch.
14967 *Geoff Thorpe, Richard Levitte*
14969 * Rework the filename-translation in the DSO code. It is now possible to
14970 have far greater control over how a "name" is turned into a filename
14971 depending on the operating environment and any oddities about the
14972 different shared library filenames on each system.
14976 * Support threads on FreeBSD-elf in Configure.
14980 * Fix for SHA1 assembly problem with MASM: it produces
14981 warnings about corrupt line number information when assembling
14982 with debugging information. This is caused by the overlapping
14985 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14988 NCONF_get_number() has no error checking at all. As a replacement,
14989 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14990 promoted strongly. The old NCONF_get_number is kept around for
14991 binary backward compatibility.
14992 Make it possible for methods to load from something other than a BIO,
14993 by providing a function pointer that is given a name instead of a BIO.
14994 For example, this could be used to load configuration data from an
14999 * Fix for non blocking accept BIOs. Added new I/O special reason
15000 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
15001 with non blocking I/O was not possible because no retry code was
15002 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
15007 * Added the beginnings of Rijndael support.
15011 * Fix for bug in DirectoryString mask setting. Add support for
15012 X509_NAME_print_ex() in 'req' and X509_print_ex() function
15013 to allow certificate printing to more controllable, additional
15014 'certopt' option to 'x509' to allow new printing options to be
15019 * Clean old EAY MD5 hack from e_os.h.
15023 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
15025 * Fix null-pointer assignment in do_change_cipher_spec() revealed
15026 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
15028 *Joe Orton, Steve Henson*
15030 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
15032 * Fix additional bug revealed by the NISCC test suite:
15034 Stop bug triggering large recursion when presented with
15035 certain ASN.1 tags ([CVE-2003-0851])
15039 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
15041 * Fix various bugs revealed by running the NISCC test suite:
15043 Stop out of bounds reads in the ASN1 code when presented with
15044 invalid tags (CVE-2003-0543 and CVE-2003-0544).
15046 If verify callback ignores invalid public key errors don't try to check
15047 certificate signature with the NULL public key.
15051 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
15052 if the server requested one: as stated in TLS 1.0 and SSL 3.0
15057 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
15058 extra data after the compression methods not only for TLS 1.0
15059 but also for SSL 3.0 (as required by the specification).
15061 *Bodo Moeller; problem pointed out by Matthias Loepfe*
15063 * Change X509_certificate_type() to mark the key as exported/exportable
15064 when it's 512 *bits* long, not 512 bytes.
15068 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
15070 * Countermeasure against the Klima-Pokorny-Rosa extension of
15071 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
15072 a protocol version number mismatch like a decryption error
15073 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
15077 * Turn on RSA blinding by default in the default implementation
15078 to avoid a timing attack. Applications that don't want it can call
15079 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
15080 They would be ill-advised to do so in most cases.
15082 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
15084 * Change RSA blinding code so that it works when the PRNG is not
15085 seeded (in this case, the secret RSA exponent is abused as
15086 an unpredictable seed -- if it is not unpredictable, there
15087 is no point in blinding anyway). Make RSA blinding thread-safe
15088 by remembering the creator's thread ID in rsa->blinding and
15089 having all other threads use local one-time blinding factors
15090 (this requires more computation than sharing rsa->blinding, but
15091 avoids excessive locking; and if an RSA object is not shared
15092 between threads, blinding will still be very fast).
15096 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
15098 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
15099 via timing by performing a MAC computation even if incorrect
15100 block cipher padding has been found. This is a countermeasure
15101 against active attacks where the attacker has to distinguish
15102 between bad padding and a MAC verification error. ([CVE-2003-0078])
15104 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
15105 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
15106 Martin Vuagnoux (EPFL, Ilion)*
15108 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
15110 * New function OPENSSL_cleanse(), which is used to cleanse a section of
15111 memory from its contents. This is done with a counter that will
15112 place alternating values in each byte. This can be used to solve
15113 two issues: 1) the removal of calls to memset() by highly optimizing
15114 compilers, and 2) cleansing with other values than 0, since those can
15115 be read through on certain media, for example a swap space on disk.
15119 * Bugfix: client side session caching did not work with external caching,
15120 because the session->cipher setting was not restored when reloading
15121 from the external cache. This problem was masked, when
15122 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
15123 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
15127 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
15128 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
15130 *Zeev Lieber <zeev-l@yahoo.com>*
15132 * Undo an undocumented change introduced in 0.9.6e which caused
15133 repeated calls to OpenSSL_add_all_ciphers() and
15134 OpenSSL_add_all_digests() to be ignored, even after calling
15139 * Change the default configuration reader to deal with last line not
15140 being properly terminated.
15144 * Change X509_NAME_cmp() so it applies the special rules on handling
15145 DN values that are of type PrintableString, as well as RDNs of type
15146 emailAddress where the value has the type ia5String.
15148 *stefank@valicert.com via Richard Levitte*
15150 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
15151 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
15152 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
15153 the bitwise-OR of the two for use by the majority of applications
15154 wanting this behaviour, and update the docs. The documented
15155 behaviour and actual behaviour were inconsistent and had been
15156 changing anyway, so this is more a bug-fix than a behavioural
15159 *Geoff Thorpe, diagnosed by Nadav Har'El*
15161 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
15162 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
15166 * Fix initialization code race conditions in
15167 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
15168 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
15169 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
15170 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
15171 ssl2_get_cipher_by_char(),
15172 ssl3_get_cipher_by_char().
15174 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
15176 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
15177 the cached sessions are flushed, as the remove_cb() might use ex_data
15178 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
15179 (see [openssl.org #212]).
15181 *Geoff Thorpe, Lutz Jaenicke*
15183 * Fix typo in OBJ_txt2obj which incorrectly passed the content
15184 length, instead of the encoding length to d2i_ASN1_OBJECT.
15188 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
15190 * [In 0.9.6g-engine release:]
15191 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
15193 *Lynn Gazis <lgazis@rainbow.com>*
15195 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
15197 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
15198 and get fix the header length calculation.
15199 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
15200 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
15202 * Use proper error handling instead of 'assertions' in buffer
15203 overflow checks added in 0.9.6e. This prevents DoS (the
15204 assertions could call abort()).
15206 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
15208 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
15210 * Add various sanity checks to asn1_get_length() to reject
15211 the ASN1 length bytes if they exceed sizeof(long), will appear
15212 negative or the content length exceeds the length of the
15215 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15217 * Fix cipher selection routines: ciphers without encryption had no flags
15218 for the cipher strength set and where therefore not handled correctly
15219 by the selection routines (PR #130).
15223 * Fix EVP_dsa_sha macro.
15228 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
15229 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
15230 that was added in OpenSSL 0.9.6d.
15232 As the countermeasure turned out to be incompatible with some
15233 broken SSL implementations, the new option is part of SSL_OP_ALL.
15234 SSL_OP_ALL is usually employed when compatibility with weird SSL
15235 implementations is desired (e.g. '-bugs' option to 's_client' and
15236 's_server'), so the new option is automatically set in many
15241 * Changes in security patch:
15243 Changes marked "(CHATS)" were sponsored by the Defense Advanced
15244 Research Projects Agency (DARPA) and Air Force Research Laboratory,
15245 Air Force Materiel Command, USAF, under agreement number
15248 * Add various sanity checks to asn1_get_length() to reject
15249 the ASN1 length bytes if they exceed sizeof(long), will appear
15250 negative or the content length exceeds the length of the
15251 supplied buffer. ([CVE-2002-0659])
15253 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15255 * Assertions for various potential buffer overflows, not known to
15256 happen in practice.
15258 *Ben Laurie (CHATS)*
15260 * Various temporary buffers to hold ASCII versions of integers were
15261 too small for 64 bit platforms. ([CVE-2002-0655])
15262 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
15264 * Remote buffer overflow in SSL3 protocol - an attacker could
15265 supply an oversized session ID to a client. ([CVE-2002-0656])
15267 *Ben Laurie (CHATS)*
15269 * Remote buffer overflow in SSL2 protocol - an attacker could
15270 supply an oversized client master key. ([CVE-2002-0656])
15272 *Ben Laurie (CHATS)*
15274 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
15276 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
15277 encoded as NULL) with id-dsa-with-sha1.
15279 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
15281 * Check various `X509_...()` return values in `apps/req.c`.
15283 *Nils Larsch <nla@trustcenter.de>*
15285 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
15286 an end-of-file condition would erroneously be flagged, when the CRLF
15287 was just at the end of a processed block. The bug was discovered when
15288 processing data through a buffering memory BIO handing the data to a
15289 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
15290 <ptsekov@syntrex.com> and Nedelcho Stanev.
15294 * Implement a countermeasure against a vulnerability recently found
15295 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
15296 before application data chunks to avoid the use of known IVs
15297 with data potentially chosen by the attacker.
15301 * Fix length checks in ssl3_get_client_hello().
15305 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
15306 to prevent ssl3_read_internal() from incorrectly assuming that
15307 ssl3_read_bytes() found application data while handshake
15308 processing was enabled when in fact s->s3->in_read_app_data was
15309 merely automatically cleared during the initial handshake.
15311 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
15313 * Fix object definitions for Private and Enterprise: they were not
15314 recognized in their shortname (=lowercase) representation. Extend
15315 obj_dat.pl to issue an error when using undefined keywords instead
15316 of silently ignoring the problem (Svenning Sorensen
15317 <sss@sss.dnsalias.net>).
15321 * Fix DH_generate_parameters() so that it works for 'non-standard'
15322 generators, i.e. generators other than 2 and 5. (Previously, the
15323 code did not properly initialise the 'add' and 'rem' values to
15324 BN_generate_prime().)
15326 In the new general case, we do not insist that 'generator' is
15327 actually a primitive root: This requirement is rather pointless;
15328 a generator of the order-q subgroup is just as good, if not
15333 * Map new X509 verification errors to alerts. Discovered and submitted by
15334 Tom Wu <tom@arcot.com>.
15338 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
15339 returning non-zero before the data has been completely received
15340 when using non-blocking I/O.
15342 *Bodo Moeller; problem pointed out by John Hughes*
15344 * Some of the ciphers missed the strength entry (SSL_LOW etc).
15346 *Ben Laurie, Lutz Jaenicke*
15348 * Fix bug in SSL_clear(): bad sessions were not removed (found by
15349 Yoram Zahavi <YoramZ@gilian.com>).
15353 * Add information about CygWin 1.3 and on, and preserve proper
15354 configuration for the versions before that.
15356 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
15358 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
15359 check whether we deal with a copy of a session and do not delete from
15360 the cache in this case. Problem reported by "Izhar Shoshani Levi"
15361 <izhar@checkpoint.com>.
15365 * Do not store session data into the internal session cache, if it
15366 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
15367 flag is set). Proposed by Aslam <aslam@funk.com>.
15371 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
15376 * [In 0.9.6d-engine release:]
15377 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
15379 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
15381 * Add the configuration target linux-s390x.
15383 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
15385 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
15386 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
15387 variable as an indication that a ClientHello message has been
15388 received. As the flag value will be lost between multiple
15389 invocations of ssl3_accept when using non-blocking I/O, the
15390 function may not be aware that a handshake has actually taken
15391 place, thus preventing a new session from being added to the
15394 To avoid this problem, we now set s->new_session to 2 instead of
15395 using a local variable.
15397 *Lutz Jaenicke, Bodo Moeller*
15399 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15400 if the SSL_R_LENGTH_MISMATCH error is detected.
15402 *Geoff Thorpe, Bodo Moeller*
15404 * New 'shared_ldflag' column in Configure platform table.
15408 * Fix EVP_CIPHER_mode macro.
15410 *"Dan S. Camper" <dan@bti.net>*
15412 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15413 type, we must throw them away by setting rr->length to 0.
15415 *D P Chang <dpc@qualys.com>*
15417 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15419 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15420 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15421 worked incorrectly for those cases where range = `10..._2` and
15422 `3*range` is two bits longer than range.)
15426 * Only add signing time to PKCS7 structures if it is not already
15431 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15432 OBJ_ld_ce should be OBJ_id_ce.
15433 Also some ip-pda OIDs in crypto/objects/objects.txt were
15434 incorrect (cf. RFC 3039).
15436 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15438 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15439 returns early because it has nothing to do.
15441 *Andy Schneider <andy.schneider@bjss.co.uk>*
15443 * [In 0.9.6c-engine release:]
15444 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15446 *Andy Schneider <andy.schneider@bjss.co.uk>*
15448 * [In 0.9.6c-engine release:]
15449 Add support for Cryptographic Appliance's keyserver technology.
15450 (Use engine 'keyclient')
15452 *Cryptographic Appliances and Geoff Thorpe*
15454 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15455 is called via tools/c89.sh because arguments have to be
15456 rearranged (all '-L' options must appear before the first object
15459 *Richard Shapiro <rshapiro@abinitio.com>*
15461 * [In 0.9.6c-engine release:]
15462 Add support for Broadcom crypto accelerator cards, backported
15465 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15467 * [In 0.9.6c-engine release:]
15468 Add support for SureWare crypto accelerator cards from
15469 Baltimore Technologies. (Use engine 'sureware')
15471 *Baltimore Technologies and Mark Cox*
15473 * [In 0.9.6c-engine release:]
15474 Add support for crypto accelerator cards from Accelerated
15475 Encryption Processing, www.aep.ie. (Use engine 'aep')
15477 *AEP Inc. and Mark Cox*
15479 * Add a configuration entry for gcc on UnixWare.
15481 *Gary Benson <gbenson@redhat.com>*
15483 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15484 messages are stored in a single piece (fixed-length part and
15485 variable-length part combined) and fix various bugs found on the way.
15489 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15490 instead. BIO_gethostbyname() does not know what timeouts are
15491 appropriate, so entries would stay in cache even when they have
15493 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15495 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15496 faced with a pathologically small ClientHello fragment that does
15497 not contain client_version: Instead of aborting with an error,
15498 simply choose the highest available protocol version (i.e.,
15499 TLS 1.0 unless it is disabled). In practice, ClientHello
15500 messages are never sent like this, but this change gives us
15501 strictly correct behaviour at least for TLS.
15505 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15506 never resets s->method to s->ctx->method when called from within
15507 one of the SSL handshake functions.
15509 *Bodo Moeller; problem pointed out by Niko Baric*
15511 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15512 (sent using the client's version number) if client_version is
15513 smaller than the protocol version in use. Also change
15514 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15515 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15516 the client will at least see that alert.
15520 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15525 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15526 client receives HelloRequest while in a handshake.
15528 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15530 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15531 should end in 'break', not 'goto end' which circumvents various
15532 cleanups done in state SSL_ST_OK. But session related stuff
15533 must be disabled for SSL_ST_OK in the case that we just sent a
15536 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15537 before just sending a HelloRequest.
15539 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15541 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15542 reveal whether illegal block cipher padding was found or a MAC
15543 verification error occurred. (Neither SSLerr() codes nor alerts
15544 are directly visible to potential attackers, but the information
15545 may leak via logfiles.)
15547 Similar changes are not required for the SSL 2.0 implementation
15548 because the number of padding bytes is sent in clear for SSL 2.0,
15549 and the extra bytes are just ignored. However ssl/s2_pkt.c
15550 failed to verify that the purported number of padding bytes is in
15555 * Add OpenUNIX-8 support including shared libraries
15556 (Boyd Lynn Gerber <gerberb@zenez.com>).
15560 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15561 'wristwatch attack' using huge encoding parameters (cf.
15562 James H. Manger's CRYPTO 2001 paper). Note that the
15563 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15564 encoding parameters and hence was not vulnerable.
15568 * BN_sqr() bug fix.
15570 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15572 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15573 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15574 followed by modular reduction.
15576 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15578 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15579 equivalent based on BN_pseudo_rand() instead of BN_rand().
15583 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15584 This function was broken, as the check for a new client hello message
15585 to handle SGC did not allow these large messages.
15586 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15590 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15594 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15595 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15599 * Rework the configuration and shared library support for Tru64 Unix.
15600 The configuration part makes use of modern compiler features and
15601 still retains old compiler behavior for those that run older versions
15602 of the OS. The shared library support part includes a variant that
15603 uses the RPATH feature, and is available through the special
15604 configuration target "alpha-cc-rpath", which will never be selected
15607 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15609 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15610 with the same message size as in ssl3_get_certificate_request().
15611 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15612 messages might inadvertently be reject as too long.
15614 *Petr Lampa <lampa@fee.vutbr.cz>*
15616 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15620 * Modified SSL library such that the verify_callback that has been set
15621 specifically for an SSL object with SSL_set_verify() is actually being
15622 used. Before the change, a verify_callback set with this function was
15623 ignored and the verify_callback() set in the SSL_CTX at the time of
15624 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15625 to allow the necessary settings.
15629 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15630 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15631 done automatically (in contradiction to the requirements of the C
15632 standard). This made problems when used from OpenSSH.
15636 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15637 dh->length and always used
15639 BN_rand_range(priv_key, dh->p).
15641 BN_rand_range() is not necessary for Diffie-Hellman, and this
15642 specific range makes Diffie-Hellman unnecessarily inefficient if
15643 dh->length (recommended exponent length) is much smaller than the
15644 length of dh->p. We could use BN_rand_range() if the order of
15645 the subgroup was stored in the DH structure, but we only have
15650 BN_rand(priv_key, l, ...)
15652 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15659 RSA_eay_public_encrypt
15660 RSA_eay_private_decrypt
15661 RSA_eay_private_encrypt (signing)
15662 RSA_eay_public_decrypt (signature verification)
15664 (default implementations for RSA_public_encrypt,
15665 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15666 always reject numbers >= n.
15670 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15671 to synchronize access to 'locking_thread'. This is necessary on
15672 systems where access to 'locking_thread' (an 'unsigned long'
15673 variable) is not atomic.
15677 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15678 *before* setting the 'crypto_lock_rand' flag. The previous code had
15679 a race condition if 0 is a valid thread ID.
15681 *Travis Vitek <vitek@roguewave.com>*
15683 * Add support for shared libraries under Irix.
15685 *Albert Chin-A-Young <china@thewrittenword.com>*
15687 * Add configuration option to build on Linux on both big-endian and
15688 little-endian MIPS.
15690 *Ralf Baechle <ralf@uni-koblenz.de>*
15692 * Add the possibility to create shared libraries on HP-UX.
15696 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15698 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15699 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15700 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15701 PRNG state recovery was possible based on the output of
15702 one PRNG request appropriately sized to gain knowledge on
15703 'md' followed by enough consecutive 1-byte PRNG requests
15704 to traverse all of 'state'.
15706 1. When updating 'md_local' (the current thread's copy of 'md')
15707 during PRNG output generation, hash all of the previous
15708 'md_local' value, not just the half used for PRNG output.
15710 2. Make the number of bytes from 'state' included into the hash
15711 independent from the number of PRNG bytes requested.
15713 The first measure alone would be sufficient to avoid
15714 Markku-Juhani's attack. (Actually it had never occurred
15715 to me that the half of 'md_local' used for chaining was the
15716 half from which PRNG output bytes were taken -- I had always
15717 assumed that the secret half would be used.) The second
15718 measure makes sure that additional data from 'state' is never
15719 mixed into 'md_local' in small portions; this heuristically
15720 further strengthens the PRNG.
15724 * Fix crypto/bn/asm/mips3.s.
15728 * When only the key is given to "enc", the IV is undefined. Print out
15729 an error message in this case.
15733 * Handle special case when X509_NAME is empty in X509 printing routines.
15737 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15738 positive and less than q.
15742 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15743 used: it isn't thread safe and the add_lock_callback should handle
15746 *Paul Rose <Paul.Rose@bridge.com>*
15748 * Verify that incoming data obeys the block size in
15749 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15755 *Ulf Möller, Bodo Möller*
15757 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15758 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15759 when fixing the server behaviour for backwards-compatible 'client
15760 hello' messages. (Note that the attack is impractical against
15761 SSL 3.0 and TLS 1.0 anyway because length and version checking
15762 means that the probability of guessing a valid ciphertext is
15763 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15766 Before 0.9.5, the countermeasure (hide the error by generating a
15767 random 'decryption result') did not work properly because
15768 ERR_clear_error() was missing, meaning that SSL_get_error() would
15769 detect the supposedly ignored error.
15771 Both problems are now fixed.
15775 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15776 (previously it was 1024).
15780 * Fix for compatibility mode trust settings: ignore trust settings
15781 unless some valid trust or reject settings are present.
15785 * Fix for blowfish EVP: its a variable length cipher.
15789 * Fix various bugs related to DSA S/MIME verification. Handle missing
15790 parameters in DSA public key structures and return an error in the
15791 DSA routines if parameters are absent.
15795 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15796 in the current directory if neither $RANDFILE nor $HOME was set.
15797 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15798 caused some confusion to Windows users who haven't defined $HOME.
15799 Thus RAND_file_name() is changed again: e_os.h can define a
15800 DEFAULT_HOME, which will be used if $HOME is not set.
15801 For Windows, we use "C:"; on other platforms, we still require
15802 environment variables.
15804 * Move 'if (!initialized) RAND_poll()' into regions protected by
15805 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15806 having multiple threads call RAND_poll() concurrently.
15810 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15811 combination of a flag and a thread ID variable.
15812 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15813 flag), *other* threads can enter ssleay_add_bytes without obeying
15814 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15815 that they do not hold after the first thread unsets add_do_not_lock).
15819 * Change bctest again: '-x' expressions are not available in all
15820 versions of 'test'.
15824 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15826 * Fix a couple of memory leaks in PKCS7_dataDecode()
15828 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15830 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15831 the default extension for executables, if any. Also, make the perl
15832 scripts that use symlink() to test if it really exists and use "cp"
15833 if it doesn't. All this made OpenSSL compilable and installable in
15838 * Fix for asn1_GetSequence() for indefinite length constructed data.
15839 If SEQUENCE is length is indefinite just set c->slen to the total
15840 amount of data available.
15842 *Steve Henson, reported by shige@FreeBSD.org*
15844 *This change does not apply to 0.9.7.*
15846 * Change bctest to avoid here-documents inside command substitution
15847 (workaround for FreeBSD /bin/sh bug).
15848 For compatibility with Ultrix, avoid shell functions (introduced
15849 in the bctest version that searches along $PATH).
15853 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15854 with des_encrypt() defined on some operating systems, like Solaris
15859 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15860 On the Importance of Eliminating Errors in Cryptographic
15861 Computations, J. Cryptology 14 (2001) 2, 101-119,
15862 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15866 * MIPS assembler BIGNUM division bug fix.
15870 * Disabled incorrect Alpha assembler code.
15874 * Fix PKCS#7 decode routines so they correctly update the length
15875 after reading an EOC for the EXPLICIT tag.
15879 *This change does not apply to 0.9.7.*
15881 * Fix bug in PKCS#12 key generation routines. This was triggered
15882 if a 3DES key was generated with a 0 initial byte. Include
15883 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15884 (but broken) behaviour.
15888 * Enhance bctest to search for a working bc along $PATH and print
15891 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15893 * Fix memory leaks in err.c: free err_data string if necessary;
15894 don't write to the wrong index in ERR_set_error_data.
15898 * Implement ssl23_peek (analogous to ssl23_read), which previously
15903 * Replace rdtsc with `_emit` statements for VC++ version 5.
15905 *Jeremy Cooper <jeremy@baymoo.org>*
15907 * Make it possible to reuse SSLv2 sessions.
15911 * In copy_email() check for >= 0 as a return value for
15912 X509_NAME_get_index_by_NID() since 0 is a valid index.
15914 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15916 * Avoid coredump with unsupported or invalid public keys by checking if
15917 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15918 PKCS7_verify() fails with non detached data.
15922 * Don't use getenv in library functions when run as setuid/setgid.
15923 New function OPENSSL_issetugid().
15927 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15928 due to incorrect handling of multi-threading:
15930 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15932 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15934 3. Count how many times MemCheck_off() has been called so that
15935 nested use can be treated correctly. This also avoids
15936 inband-signalling in the previous code (which relied on the
15937 assumption that thread ID 0 is impossible).
15941 * Add "-rand" option also to s_client and s_server.
15945 * Fix CPU detection on Irix 6.x.
15946 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15947 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15949 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15954 *This change does not apply to 0.9.7.*
15956 * Use the cached encoding of an X509_NAME structure rather than
15957 copying it. This is apparently the reason for the libsafe "errors"
15958 but the code is actually correct.
15962 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15963 Bleichenbacher's DSA attack.
15964 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15965 to be set and top=0 forces the highest bit to be set; top=-1 is new
15966 and leaves the highest bit random.
15968 *Ulf Moeller, Bodo Moeller*
15970 * In the `NCONF_...`-based implementations for `CONF_...` queries
15971 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15972 a temporary CONF structure with the data component set to NULL
15973 (which gives segmentation faults in lh_retrieve).
15974 Instead, use NULL for the CONF pointer in CONF_get_string and
15975 CONF_get_number (which may use environment variables) and directly
15976 return NULL from CONF_get_section.
15980 * Fix potential buffer overrun for EBCDIC.
15984 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15985 keyUsage if basicConstraints absent for a CA.
15989 * Make SMIME_write_PKCS7() write mail header values with a format that
15990 is more generally accepted (no spaces before the semicolon), since
15991 some programs can't parse those values properly otherwise. Also make
15992 sure BIO's that break lines after each write do not create invalid
15997 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15998 macros previously used would not encode an empty SEQUENCE OF
15999 and break the signature.
16003 *This change does not apply to 0.9.7.*
16005 * Zero the premaster secret after deriving the master secret in
16010 * Add some EVP_add_digest_alias registrations (as found in
16011 OpenSSL_add_all_digests()) to SSL_library_init()
16012 aka OpenSSL_add_ssl_algorithms(). This provides improved
16013 compatibility with peers using X.509 certificates
16014 with unconventional AlgorithmIdentifier OIDs.
16018 * Fix for Irix with NO_ASM.
16020 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16022 * ./config script fixes.
16024 *Ulf Moeller, Richard Levitte*
16026 * Fix 'openssl passwd -1'.
16030 * Change PKCS12_key_gen_asc() so it can cope with non null
16031 terminated strings whose length is passed in the passlen
16032 parameter, for example from PEM callbacks. This was done
16033 by adding an extra length parameter to asc2uni().
16035 *Steve Henson, reported by <oddissey@samsung.co.kr>*
16037 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
16038 call failed, free the DSA structure.
16042 * Fix to uni2asc() to cope with zero length Unicode strings.
16043 These are present in some PKCS#12 files.
16047 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
16048 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
16049 when writing a 32767 byte record.
16051 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
16053 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
16054 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
16056 (RSA objects have a reference count access to which is protected
16057 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
16058 so they are meant to be shared between threads.)
16059 *Bodo Moeller, Geoff Thorpe; original patch submitted by
16060 "Reddie, Steven" <Steven.Reddie@ca.com>*
16062 * Fix a deadlock in CRYPTO_mem_leaks().
16066 * Use better test patterns in bntest.
16070 * rand_win.c fix for Borland C.
16074 * BN_rshift bugfix for n == 0.
16078 * Add a 'bctest' script that checks for some known 'bc' bugs
16079 so that 'make test' does not abort just because 'bc' is broken.
16083 * Store verify_result within SSL_SESSION also for client side to
16084 avoid potential security hole. (Re-used sessions on the client side
16085 always resulted in verify_result==X509_V_OK, not using the original
16086 result of the server certificate verification.)
16090 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
16091 SSL3_RT_APPLICATION_DATA, return 0.
16092 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
16097 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
16098 releases, have been re-implemented by renaming the previous
16099 implementations of ssl2_read and ssl3_read to ssl2_read_internal
16100 and ssl3_read_internal, respectively, and adding 'peek' parameters
16101 to them. The new ssl[23]_{read,peek} functions are calls to
16102 ssl[23]_read_internal with the 'peek' flag set appropriately.
16103 A 'peek' parameter has also been added to ssl3_read_bytes, which
16104 does the actual work for ssl3_read_internal.
16108 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
16109 the method-specific "init()" handler. Also clean up ex_data after
16110 calling the method-specific "finish()" handler. Previously, this was
16111 happening the other way round.
16115 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
16116 The previous value, 12, was not always sufficient for BN_mod_exp().
16120 * Make sure that shared libraries get the internal name engine with
16121 the full version number and not just 0. This should mark the
16122 shared libraries as not backward compatible. Of course, this should
16123 be changed again when we can guarantee backward binary compatibility.
16127 * Fix typo in get_cert_by_subject() in by_dir.c
16129 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
16131 * Rework the system to generate shared libraries:
16133 - Make note of the expected extension for the shared libraries and
16134 if there is a need for symbolic links from for example libcrypto.so.0
16135 to libcrypto.so.0.9.7. There is extended info in Configure for
16138 - Make as few rebuilds of the shared libraries as possible.
16140 - Still avoid linking the OpenSSL programs with the shared libraries.
16142 - When installing, install the shared libraries separately from the
16147 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
16149 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
16150 and not in SSL_clear because the latter is also used by the
16151 accept/connect functions; previously, the settings made by
16152 SSL_set_read_ahead would be lost during the handshake.
16154 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
16156 * Correct util/mkdef.pl to be selective about disabled algorithms.
16157 Previously, it would create entries for disabled algorithms no
16162 * Added several new manual pages for SSL_* function.
16166 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
16168 * In ssl23_get_client_hello, generate an error message when faced
16169 with an initial SSL 3.0/TLS record that is too small to contain the
16170 first two bytes of the ClientHello message, i.e. client_version.
16171 (Note that this is a pathologic case that probably has never happened
16172 in real life.) The previous approach was to use the version number
16173 from the record header as a substitute; but our protocol choice
16174 should not depend on that one because it is not authenticated
16175 by the Finished messages.
16179 * More robust randomness gathering functions for Windows.
16181 *Jeffrey Altman <jaltman@columbia.edu>*
16183 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
16184 not set then we don't setup the error code for issuer check errors
16185 to avoid possibly overwriting other errors which the callback does
16186 handle. If an application does set the flag then we assume it knows
16187 what it is doing and can handle the new informational codes
16192 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
16193 a general "ANY" type, as such it should be able to decode anything
16194 including tagged types. However it didn't check the class so it would
16195 wrongly interpret tagged types in the same way as their universal
16196 counterpart and unknown types were just rejected. Changed so that the
16197 tagged and unknown types are handled in the same way as a SEQUENCE:
16198 that is the encoding is stored intact. There is also a new type
16199 "V_ASN1_OTHER" which is used when the class is not universal, in this
16200 case we have no idea what the actual type is so we just lump them all
16205 * On VMS, stdout may very well lead to a file that is written to
16206 in a record-oriented fashion. That means that every write() will
16207 write a separate record, which will be read separately by the
16208 programs trying to read from it. This can be very confusing.
16210 The solution is to put a BIO filter in the way that will buffer
16211 text until a linefeed is reached, and then write everything a
16212 line at a time, so every record written will be an actual line,
16213 not chunks of lines and not (usually doesn't happen, but I've
16214 seen it once) several lines in one record. BIO_f_linebuffer() is
16217 Currently, it's a VMS-only method, because that's where it has
16218 been tested well enough.
16222 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
16223 it can return incorrect results.
16224 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
16225 but it was in 0.9.6-beta[12].)
16229 * Disable the check for content being present when verifying detached
16230 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
16231 include zero length content when signing messages.
16235 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
16236 BIO_ctrl (for BIO pairs).
16240 * Add DSO method for VMS.
16244 * Bug fix: Montgomery multiplication could produce results with the
16249 * Add RPM specification openssl.spec and modify it to build three
16250 packages. The default package contains applications, application
16251 documentation and run-time libraries. The devel package contains
16252 include files, static libraries and function documentation. The
16253 doc package contains the contents of the doc directory. The original
16254 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
16258 * Add a large number of documentation files for many SSL routines.
16260 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16262 * Add a configuration entry for Sony News 4.
16264 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
16266 * Don't set the two most significant bits to one when generating a
16267 random number < q in the DSA library.
16271 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
16272 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
16273 the underlying transport is blocking) if a handshake took place.
16274 (The default behaviour is needed by applications such as s_client
16275 and s_server that use select() to determine when to use SSL_read;
16276 but for applications that know in advance when to expect data, it
16277 just makes things more complicated.)
16281 * Add RAND_egd_bytes(), which gives control over the number of bytes read
16286 * Add a few more EBCDIC conditionals that make `req` and `x509`
16287 work better on such systems.
16289 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16291 * Add two demo programs for PKCS12_parse() and PKCS12_create().
16292 Update PKCS12_parse() so it copies the friendlyName and the
16293 keyid to the certificates aux info.
16297 * Fix bug in PKCS7_verify() which caused an infinite loop
16298 if there was more than one signature.
16300 *Sven Uszpelkat <su@celocom.de>*
16302 * Major change in util/mkdef.pl to include extra information
16303 about each symbol, as well as presenting variables as well
16304 as functions. This change means that there's n more need
16305 to rebuild the .num files when some algorithms are excluded.
16309 * Allow the verify time to be set by an application,
16310 rather than always using the current time.
16314 * Phase 2 verify code reorganisation. The certificate
16315 verify code now looks up an issuer certificate by a
16316 number of criteria: subject name, authority key id
16317 and key usage. It also verifies self signed certificates
16318 by the same criteria. The main comparison function is
16319 X509_check_issued() which performs these checks.
16321 Lot of changes were necessary in order to support this
16322 without completely rewriting the lookup code.
16324 Authority and subject key identifier are now cached.
16326 The LHASH 'certs' is X509_STORE has now been replaced
16327 by a STACK_OF(X509_OBJECT). This is mainly because an
16328 LHASH can't store or retrieve multiple objects with
16329 the same hash value.
16331 As a result various functions (which were all internal
16332 use only) have changed to handle the new X509_STORE
16333 structure. This will break anything that messed round
16334 with X509_STORE internally.
16336 The functions X509_STORE_add_cert() now checks for an
16337 exact match, rather than just subject name.
16339 The X509_STORE API doesn't directly support the retrieval
16340 of multiple certificates matching a given criteria, however
16341 this can be worked round by performing a lookup first
16342 (which will fill the cache with candidate certificates)
16343 and then examining the cache for matches. This is probably
16344 the best we can do without throwing out X509_LOOKUP
16345 entirely (maybe later...).
16347 The X509_VERIFY_CTX structure has been enhanced considerably.
16349 All certificate lookup operations now go via a get_issuer()
16350 callback. Although this currently uses an X509_STORE it
16351 can be replaced by custom lookups. This is a simple way
16352 to bypass the X509_STORE hackery necessary to make this
16353 work and makes it possible to use more efficient techniques
16354 in future. A very simple version which uses a simple
16355 STACK for its trusted certificate store is also provided
16356 using X509_STORE_CTX_trusted_stack().
16358 The verify_cb() and verify() callbacks now have equivalents
16359 in the X509_STORE_CTX structure.
16361 X509_STORE_CTX also has a 'flags' field which can be used
16362 to customise the verify behaviour.
16366 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
16367 excludes S/MIME capabilities.
16371 * When a certificate request is read in keep a copy of the
16372 original encoding of the signed data and use it when outputting
16373 again. Signatures then use the original encoding rather than
16374 a decoded, encoded version which may cause problems if the
16375 request is improperly encoded.
16379 * For consistency with other BIO_puts implementations, call
16380 buffer_write(b, ...) directly in buffer_puts instead of calling
16383 In BIO_puts, increment b->num_write as in BIO_write.
16385 *Peter.Sylvester@EdelWeb.fr*
16387 * Fix BN_mul_word for the case where the word is 0. (We have to use
16388 BN_zero, we may not return a BIGNUM with an array consisting of
16389 words set to zero.)
16393 * Avoid calling abort() from within the library when problems are
16394 detected, except if preprocessor symbols have been defined
16395 (such as REF_CHECK, BN_DEBUG etc.).
16399 * New openssl application 'rsautl'. This utility can be
16400 used for low-level RSA operations. DER public key
16401 BIO/fp routines also added.
16405 * New Configure entry and patches for compiling on QNX 4.
16407 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16409 * A demo state-machine implementation was sponsored by
16410 Nuron (<http://www.nuron.com/>) and is now available in
16411 demos/state_machine.
16415 * New options added to the 'dgst' utility for signature
16416 generation and verification.
16420 * Unrecognized PKCS#7 content types are now handled via a
16421 catch all ASN1_TYPE structure. This allows unsupported
16422 types to be stored as a "blob" and an application can
16423 encode and decode it manually.
16427 * Fix various signed/unsigned issues to make a_strex.c
16428 compile under VC++.
16430 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16432 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16433 length if passed a buffer. ASN1_INTEGER_to_BN failed
16434 if passed a NULL BN and its argument was negative.
16436 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16438 * Modification to PKCS#7 encoding routines to output definite
16439 length encoding. Since currently the whole structures are in
16440 memory there's not real point in using indefinite length
16441 constructed encoding. However if OpenSSL is compiled with
16442 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16446 * Added BIO_vprintf() and BIO_vsnprintf().
16450 * Added more prefixes to parse for in the strings written
16451 through a logging bio, to cover all the levels that are available
16452 through syslog. The prefixes are now:
16454 PANIC, EMERG, EMR => LOG_EMERG
16455 ALERT, ALR => LOG_ALERT
16456 CRIT, CRI => LOG_CRIT
16457 ERROR, ERR => LOG_ERR
16458 WARNING, WARN, WAR => LOG_WARNING
16459 NOTICE, NOTE, NOT => LOG_NOTICE
16460 INFO, INF => LOG_INFO
16461 DEBUG, DBG => LOG_DEBUG
16463 and as before, if none of those prefixes are present at the
16464 beginning of the string, LOG_ERR is chosen.
16466 On Win32, the `LOG_*` levels are mapped according to this:
16468 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16469 LOG_WARNING => EVENTLOG_WARNING_TYPE
16470 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16474 * Made it possible to reconfigure with just the configuration
16475 argument "reconf" or "reconfigure". The command line arguments
16476 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16477 and are retrieved from there when reconfiguring.
16483 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16485 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16489 * The obj_dat.pl script was messing up the sorting of object
16490 names. The reason was that it compared the quoted version
16491 of strings as a result "OCSP" > "OCSP Signing" because
16492 " > SPACE. Changed script to store unquoted versions of
16493 names and add quotes on output. It was also omitting some
16494 names from the lookup table if they were given a default
16495 value (that is if SN is missing it is given the same
16496 value as LN and vice versa), these are now added on the
16497 grounds that if an object has a name we should be able to
16498 look it up. Finally added warning output when duplicate
16499 short or long names are found.
16503 * Changes needed for Tandem NSK.
16505 *Scott Uroff <scott@xypro.com>*
16507 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16508 RSA_padding_check_SSLv23(), special padding was never detected
16509 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16510 version rollback attacks was not effective.
16512 In s23_clnt.c, don't use special rollback-attack detection padding
16513 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16514 client; similarly, in s23_srvr.c, don't do the rollback check if
16515 SSL 2.0 is the only protocol enabled in the server.
16519 * Make it possible to get hexdumps of unprintable data with 'openssl
16520 asn1parse'. By implication, the functions ASN1_parse_dump() and
16521 BIO_dump_indent() are added.
16525 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16526 these print out strings and name structures based on various
16527 flags including RFC2253 support and proper handling of
16528 multibyte characters. Added options to the 'x509' utility
16529 to allow the various flags to be set.
16533 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16534 Also change the functions X509_cmp_current_time() and
16535 X509_gmtime_adj() work with an ASN1_TIME structure,
16536 this will enable certificates using GeneralizedTime in validity
16537 dates to be checked.
16541 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16542 negative public key encodings) on by default,
16543 NO_NEG_PUBKEY_BUG can be set to disable it.
16547 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16548 content octets. An i2c_ASN1_OBJECT is unnecessary because
16549 the encoding can be trivially obtained from the structure.
16553 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16554 not read locks (`CRYPTO_r_[un]lock`).
16558 * A first attempt at creating official support for shared
16559 libraries through configuration. I've kept it so the
16560 default is static libraries only, and the OpenSSL programs
16561 are always statically linked for now, but there are
16562 preparations for dynamic linking in place.
16563 This has been tested on Linux and Tru64.
16567 * Randomness polling function for Win9x, as described in:
16568 Peter Gutmann, Software Generation of Practically Strong
16573 * Fix so PRNG is seeded in req if using an already existing
16578 * New options to smime application. -inform and -outform
16579 allow alternative formats for the S/MIME message including
16580 PEM and DER. The -content option allows the content to be
16581 specified separately. This should allow things like Netscape
16582 form signing output easier to verify.
16586 * Fix the ASN1 encoding of tags using the 'long form'.
16590 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16591 STRING types. These convert content octets to and from the
16592 underlying type. The actual tag and length octets are
16593 already assumed to have been read in and checked. These
16594 are needed because all other string types have virtually
16595 identical handling apart from the tag. By having versions
16596 of the ASN1 functions that just operate on content octets
16597 IMPLICIT tagging can be handled properly. It also allows
16598 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16599 and ASN1_INTEGER are identical apart from the tag.
16603 * Change the handling of OID objects as follows:
16605 - New object identifiers are inserted in objects.txt, following
16606 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16607 - objects.pl is used to process obj_mac.num and create a new
16609 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16612 This is currently kind of a hack, and the perl code in objects.pl
16613 isn't very elegant, but it works as I intended. The simplest way
16614 to check that it worked correctly is to look in obj_dat.h and
16615 check the array nid_objs and make sure the objects haven't moved
16616 around (this is important!). Additions are OK, as well as
16617 consistent name changes.
16621 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16625 * Addition of the command line parameter '-rand file' to 'openssl req'.
16626 The given file adds to whatever has already been seeded into the
16627 random pool through the RANDFILE configuration file option or
16628 environment variable, or the default random state file.
16632 * mkstack.pl now sorts each macro group into lexical order.
16633 Previously the output order depended on the order the files
16634 appeared in the directory, resulting in needless rewriting
16639 * Patches to make OpenSSL compile under Win32 again. Mostly
16640 work arounds for the VC++ problem that it treats func() as
16641 func(void). Also stripped out the parts of mkdef.pl that
16642 added extra typesafe functions: these no longer exist.
16646 * Reorganisation of the stack code. The macros are now all
16647 collected in safestack.h . Each macro is defined in terms of
16648 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16649 DEBUG_SAFESTACK is now handled in terms of function casts,
16650 this has the advantage of retaining type safety without the
16651 use of additional functions. If DEBUG_SAFESTACK is not defined
16652 then the non typesafe macros are used instead. Also modified the
16653 mkstack.pl script to handle the new form. Needs testing to see
16654 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16655 the default if no major problems. Similar behaviour for ASN1_SET_OF
16656 and PKCS12_STACK_OF.
16660 * When some versions of IIS use the 'NET' form of private key the
16661 key derivation algorithm is different. Normally MD5(password) is
16662 used as a 128 bit RC4 key. In the modified case
16663 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16664 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16665 as the old Netscape_RSA functions except they have an additional
16666 'sgckey' parameter which uses the modified algorithm. Also added
16667 an -sgckey command line option to the rsa utility. Thanks to
16668 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16669 algorithm to openssl-dev.
16673 * The evp_local.h macros were using 'c.##kname' which resulted in
16674 invalid expansion on some systems (SCO 5.0.5 for example).
16675 Corrected to 'c.kname'.
16677 *Phillip Porch <root@theporch.com>*
16679 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16680 a STACK of email addresses from a certificate or request, these look
16681 in the subject name and the subject alternative name extensions and
16682 omit any duplicate addresses.
16686 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16687 This makes DSA verification about 2 % faster.
16691 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16692 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16693 plus overhead for 1024 bit moduli).
16694 This makes exponentiations about 0.5 % faster for 1024 bit
16695 exponents (as measured by "openssl speed rsa2048").
16699 * Rename memory handling macros to avoid conflicts with other
16701 Malloc => OPENSSL_malloc
16702 Malloc_locked => OPENSSL_malloc_locked
16703 Realloc => OPENSSL_realloc
16704 Free => OPENSSL_free
16708 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16709 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16713 * CygWin32 support.
16715 *John Jarvie <jjarvie@newsguy.com>*
16717 * The type-safe stack code has been rejigged. It is now only compiled
16718 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16719 by default all type-specific stack functions are "#define"d back to
16720 standard stack functions. This results in more streamlined output
16721 but retains the type-safety checking possibilities of the original
16726 * The STACK code has been cleaned up, and certain type declarations
16727 that didn't make a lot of sense have been brought in line. This has
16728 also involved a cleanup of sorts in safestack.h to more correctly
16729 map type-safe stack functions onto their plain stack counterparts.
16730 This work has also resulted in a variety of "const"ifications of
16731 lots of the code, especially `_cmp` operations which should normally
16732 be prototyped with "const" parameters anyway.
16736 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16737 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16738 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16739 where all of 'md' is used each time the PRNG is used, but 'state'
16740 is used only indexed by a cyclic counter. As entropy may not be
16741 well distributed from the beginning, 'md' is important as a
16742 chaining variable. However, the output function chains only half
16743 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16744 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16745 in all of 'state' being rewritten, with the new values depending
16746 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16750 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16751 the handshake is continued after ssl_verify_cert_chain();
16752 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16753 can lead to 'unexplainable' connection aborts later.
16755 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16757 * Major EVP API cipher revision.
16758 Add hooks for extra EVP features. This allows various cipher
16759 parameters to be set in the EVP interface. Support added for variable
16760 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16761 setting of RC2 and RC5 parameters.
16763 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16766 Remove lots of duplicated code from the EVP library. For example *every*
16767 cipher init() function handles the 'iv' in the same way according to the
16768 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16769 for CFB and OFB modes they zero ctx->num.
16771 New functionality allows removal of S/MIME code RC2 hack.
16773 Most of the routines have the same form and so can be declared in terms
16776 By shifting this to the top level EVP_CipherInit() it can be removed from
16777 all individual ciphers. If the cipher wants to handle IVs or keys
16778 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16781 Change lots of functions like EVP_EncryptUpdate() to now return a
16782 value: although software versions of the algorithms cannot fail
16783 any installed hardware versions can.
16787 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16788 this option is set, tolerate broken clients that send the negotiated
16789 protocol version number instead of the requested protocol version
16794 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16795 i.e. non-zero for export ciphersuites, zero otherwise.
16796 Previous versions had this flag inverted, inconsistent with
16797 rsa_tmp_cb (..._TMP_RSA_CB).
16799 *Bodo Moeller; problem reported by Amit Chopra*
16801 * Add missing DSA library text string. Work around for some IIS
16802 key files with invalid SEQUENCE encoding.
16806 * Add a document (doc/standards.txt) that list all kinds of standards
16807 and so on that are implemented in OpenSSL.
16811 * Enhance c_rehash script. Old version would mishandle certificates
16812 with the same subject name hash and wouldn't handle CRLs at all.
16813 Added -fingerprint option to crl utility, to support new c_rehash
16818 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16822 * Fix for SSL server purpose checking. Server checking was
16823 rejecting certificates which had extended key usage present
16824 but no ssl client purpose.
16826 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16828 * Make PKCS#12 code work with no password. The PKCS#12 spec
16829 is a little unclear about how a blank password is handled.
16830 Since the password in encoded as a BMPString with terminating
16831 double NULL a zero length password would end up as just the
16832 double NULL. However no password at all is different and is
16833 handled differently in the PKCS#12 key generation code. NS
16834 treats a blank password as zero length. MSIE treats it as no
16835 password on export: but it will try both on import. We now do
16836 the same: PKCS12_parse() tries zero length and no password if
16837 the password is set to "" or NULL (NULL is now a valid password:
16838 it wasn't before) as does the pkcs12 application.
16842 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16843 perror when PEM_read_bio_X509_REQ fails, the error message must
16844 be obtained from the error queue.
16848 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16849 it in ERR_remove_state if appropriate, and change ERR_get_state
16850 accordingly to avoid race conditions (this is necessary because
16851 thread_hash is no longer constant once set).
16855 * Bugfix for linux-elf makefile.one.
16859 * RSA_get_default_method() will now cause a default
16860 RSA_METHOD to be chosen if one doesn't exist already.
16861 Previously this was only set during a call to RSA_new()
16862 or RSA_new_method(NULL) meaning it was possible for
16863 RSA_get_default_method() to return NULL.
16867 * Added native name translation to the existing DSO code
16868 that will convert (if the flag to do so is set) filenames
16869 that are sufficiently small and have no path information
16870 into a canonical native form. Eg. "blah" converted to
16871 "libblah.so" or "blah.dll" etc.
16875 * New function ERR_error_string_n(e, buf, len) which is like
16876 ERR_error_string(e, buf), but writes at most 'len' bytes
16877 including the 0 terminator. For ERR_error_string_n, 'buf'
16880 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16882 * CONF library reworked to become more general. A new CONF
16883 configuration file reader "class" is implemented as well as a
16884 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16885 old `CONF_*` functions are still there, but are reimplemented to
16886 work in terms of the new functions. Also, a set of functions
16887 to handle the internal storage of the configuration data is
16888 provided to make it easier to write new configuration file
16889 reader "classes" (I can definitely see something reading a
16890 configuration file in XML format, for example), called `_CONF_*`,
16891 or "the configuration storage API"...
16893 The new configuration file reading functions are:
16895 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16896 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16898 NCONF_default, NCONF_WIN32
16900 NCONF_dump_fp, NCONF_dump_bio
16902 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16903 NCONF_new creates a new CONF object. This works in the same way
16904 as other interfaces in OpenSSL, like the BIO interface.
16905 `NCONF_dump_*` dump the internal storage of the configuration file,
16906 which is useful for debugging. All other functions take the same
16907 arguments as the old `CONF_*` functions with the exception of the
16908 first that must be a `CONF *` instead of a `LHASH *`.
16910 To make it easier to use the new classes with the old `CONF_*` functions,
16911 the function CONF_set_default_method is provided.
16915 * Add '-tls1' option to 'openssl ciphers', which was already
16916 mentioned in the documentation but had not been implemented.
16917 (This option is not yet really useful because even the additional
16918 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16922 * Initial DSO code added into libcrypto for letting OpenSSL (and
16923 OpenSSL-based applications) load shared libraries and bind to
16924 them in a portable way.
16926 *Geoff Thorpe, with contributions from Richard Levitte*
16928 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16930 * Make sure _lrotl and _lrotr are only used with MSVC.
16932 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16933 (the default implementation of RAND_status).
16935 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16936 to '-clrext' (= clear extensions), as intended and documented.
16937 *Bodo Moeller; inconsistency pointed out by Michael Attili
16938 <attili@amaxo.com>*
16940 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16941 was larger than the MD block size.
16943 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16945 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16946 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16947 using the passed key: if the passed key was a private key the result
16948 of X509_print(), for example, would be to print out all the private key
16953 * des_quad_cksum() byte order bug fix.
16954 *Ulf Möller, using the problem description in krb4-0.9.7, where
16955 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16957 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16960 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16962 * For easily testing in shell scripts whether some command
16963 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16964 returns with exit code 0 iff no command of the given name is available.
16965 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16966 the output goes to stdout and nothing is printed to stderr.
16967 Additional arguments are always ignored.
16969 Since for each cipher there is a command of the same name,
16970 the 'no-cipher' compilation switches can be tested this way.
16972 ('openssl no-XXX' is not able to detect pseudo-commands such
16973 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16977 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16981 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16982 is set; it will be thrown away anyway because each handshake creates
16984 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16985 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16986 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16987 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16991 * New s_client option -ign_eof: EOF at stdin is ignored, and
16992 'Q' and 'R' lose their special meanings (quit/renegotiate).
16993 This is part of what -quiet does; unlike -quiet, -ign_eof
16994 does not suppress any output.
16998 * Add compatibility options to the purpose and trust code. The
16999 purpose X509_PURPOSE_ANY is "any purpose" which automatically
17000 accepts a certificate or CA, this was the previous behaviour,
17001 with all the associated security issues.
17003 X509_TRUST_COMPAT is the old trust behaviour: only and
17004 automatically trust self signed roots in certificate store. A
17005 new trust setting X509_TRUST_DEFAULT is used to specify that
17006 a purpose has no associated trust setting and it should instead
17007 use the value in the default purpose.
17011 * Fix the PKCS#8 DSA private key code so it decodes keys again
17012 and fix a memory leak.
17016 * In util/mkerr.pl (which implements 'make errors'), preserve
17017 reason strings from the previous version of the .c file, as
17018 the default to have only downcase letters (and digits) in
17019 automatically generated reasons codes is not always appropriate.
17023 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
17024 using strerror. Previously, ERR_reason_error_string() returned
17025 library names as reason strings for SYSerr; but SYSerr is a special
17026 case where small numbers are errno values, not library numbers.
17030 * Add '-dsaparam' option to 'openssl dhparam' application. This
17031 converts DSA parameters into DH parameters. (When creating parameters,
17032 DSA_generate_parameters is used.)
17036 * Include 'length' (recommended exponent length) in C code generated
17037 by 'openssl dhparam -C'.
17041 * The second argument to set_label in perlasm was already being used
17042 so couldn't be used as a "file scope" flag. Moved to third argument
17047 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
17048 instead of RAND_bytes for encryption IVs and salts.
17052 * Include RAND_status() into RAND_METHOD instead of implementing
17053 it only for md_rand.c Otherwise replacing the PRNG by calling
17054 RAND_set_rand_method would be impossible.
17058 * Don't let DSA_generate_key() enter an infinite loop if the random
17059 number generation fails.
17063 * New 'rand' application for creating pseudo-random output.
17067 * Added configuration support for Linux/IA64
17069 *Rolf Haberrecker <rolf@suse.de>*
17071 * Assembler module support for Mingw32.
17075 * Shared library support for HPUX (in shlib/).
17077 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
17079 * Shared library support for Solaris gcc.
17081 *Lutz Behnke <behnke@trustcenter.de>*
17083 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
17085 * PKCS7_encrypt() was adding text MIME headers twice because they
17086 were added manually and by SMIME_crlf_copy().
17090 * In bntest.c don't call BN_rand with zero bits argument.
17092 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
17094 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
17095 case was implemented. This caused BN_div_recp() to fail occasionally.
17099 * Add an optional second argument to the set_label() in the perl
17100 assembly language builder. If this argument exists and is set
17101 to 1 it signals that the assembler should use a symbol whose
17102 scope is the entire file, not just the current function. This
17103 is needed with MASM which uses the format label:: for this scope.
17105 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
17107 * Change the ASN1 types so they are typedefs by default. Before
17108 almost all types were #define'd to ASN1_STRING which was causing
17109 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
17114 * Change names of new functions to the new get1/get0 naming
17115 convention: After 'get1', the caller owns a reference count
17116 and has to call `..._free`; 'get0' returns a pointer to some
17117 data structure without incrementing reference counters.
17118 (Some of the existing 'get' functions increment a reference
17119 counter, some don't.)
17120 Similarly, 'set1' and 'add1' functions increase reference
17121 counters or duplicate objects.
17125 * Allow for the possibility of temp RSA key generation failure:
17126 the code used to assume it always worked and crashed on failure.
17130 * Fix potential buffer overrun problem in BIO_printf().
17131 *Ulf Möller, using public domain code by Patrick Powell; problem
17132 pointed out by David Sacerdote <das33@cornell.edu>*
17134 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
17135 RAND_egd() and RAND_status(). In the command line application,
17136 the EGD socket can be specified like a seed file using RANDFILE
17141 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
17142 Some CAs (e.g. Verisign) distribute certificates in this form.
17146 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
17147 list to exclude them. This means that no special compilation option
17148 is needed to use anonymous DH: it just needs to be included in the
17153 * Change the EVP_MD_CTX_type macro so its meaning consistent with
17154 EVP_MD_type. The old functionality is available in a new macro called
17155 EVP_MD_md(). Change code that uses it and update docs.
17159 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
17160 where the `void *` argument is replaced by a function pointer argument.
17161 Previously `void *` was abused to point to functions, which works on
17162 many platforms, but is not correct. As these functions are usually
17163 called by macros defined in OpenSSL header files, most source code
17164 should work without changes.
17168 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
17169 sections with information on -D... compiler switches used for
17170 compiling the library so that applications can see them. To enable
17171 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
17172 must be defined. E.g.,
17173 #define OPENSSL_ALGORITHM_DEFINES
17174 #include <openssl/opensslconf.h>
17175 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
17177 *Richard Levitte, Ulf and Bodo Möller*
17179 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
17184 * Change the 'other' type in certificate aux info to a STACK_OF
17185 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
17186 the required ASN1 format: arbitrary types determined by an OID.
17190 * Add some PEM_write_X509_REQ_NEW() functions and a command line
17191 argument to 'req'. This is not because the function is newer or
17192 better than others it just uses the work 'NEW' in the certificate
17193 request header lines. Some software needs this.
17197 * Reorganise password command line arguments: now passwords can be
17198 obtained from various sources. Delete the PEM_cb function and make
17199 it the default behaviour: i.e. if the callback is NULL and the
17200 usrdata argument is not NULL interpret it as a null terminated pass
17201 phrase. If usrdata and the callback are NULL then the pass phrase
17202 is prompted for as usual.
17206 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
17207 the support is automatically enabled. The resulting binaries will
17208 autodetect the card and use it if present.
17210 *Ben Laurie and Compaq Inc.*
17212 * Work around for Netscape hang bug. This sends certificate request
17213 and server done in one record. Since this is perfectly legal in the
17214 SSL/TLS protocol it isn't a "bug" option and is on by default. See
17215 the bugs/SSLv3 entry for more info.
17219 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
17223 * Add -rand argument to smime and pkcs12 applications and read/write
17228 * New 'passwd' tool for crypt(3) and apr1 password hashes.
17232 * Add command line password options to the remaining applications.
17236 * Bug fix for BN_div_recp() for numerators with an even number of
17241 * More tests in bntest.c, and changed test_bn output.
17245 * ./config recognizes MacOS X now.
17249 * Bug fix for BN_div() when the first words of num and divisor are
17250 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
17254 * Add support for various broken PKCS#8 formats, and command line
17255 options to produce them.
17259 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
17260 get temporary BIGNUMs from a BN_CTX.
17264 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
17269 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
17270 include a #define from the old name to the new. The original intent
17271 was that statically linked binaries could for example just call
17272 SSLeay_add_all_ciphers() to just add ciphers to the table and not
17273 link with digests. This never worked because SSLeay_add_all_digests()
17274 and SSLeay_add_all_ciphers() were in the same source file so calling
17275 one would link with the other. They are now in separate source files.
17279 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
17283 * Use a less unusual form of the Miller-Rabin primality test (it used
17284 a binary algorithm for exponentiation integrated into the Miller-Rabin
17285 loop, our standard modexp algorithms are faster).
17289 * Support for the EBCDIC character set completed.
17291 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
17293 * Source code cleanups: use const where appropriate, eliminate casts,
17294 use `void *` instead of `char *` in lhash.
17298 * Bugfix: ssl3_send_server_key_exchange was not restartable
17299 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
17300 this the server could overwrite ephemeral keys that the client
17305 * Turn DSA_is_prime into a macro that calls BN_is_prime,
17306 using 50 iterations of the Rabin-Miller test.
17308 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
17309 iterations of the Rabin-Miller test as required by the appendix
17310 to FIPS PUB 186[-1]) instead of DSA_is_prime.
17311 As BN_is_prime_fasttest includes trial division, DSA parameter
17312 generation becomes much faster.
17314 This implies a change for the callback functions in DSA_is_prime
17315 and DSA_generate_parameters: The callback function is called once
17316 for each positive witness in the Rabin-Miller test, not just
17317 occasionally in the inner loop; and the parameters to the
17318 callback function now provide an iteration count for the outer
17319 loop rather than for the current invocation of the inner loop.
17320 DSA_generate_parameters additionally can call the callback
17321 function with an 'iteration count' of -1, meaning that a
17322 candidate has passed the trial division test (when q is generated
17323 from an application-provided seed, trial division is skipped).
17327 * New function BN_is_prime_fasttest that optionally does trial
17328 division before starting the Rabin-Miller test and has
17329 an additional BN_CTX * argument (whereas BN_is_prime always
17330 has to allocate at least one BN_CTX).
17331 'callback(1, -1, cb_arg)' is called when a number has passed the
17332 trial division stage.
17336 * Fix for bug in CRL encoding. The validity dates weren't being handled
17341 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
17345 * New function BN_pseudo_rand().
17349 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
17350 bignum version of BN_from_montgomery() with the working code from
17351 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
17356 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
17357 made it impossible to use the same SSL_SESSION data structure in
17358 SSL2 clients in multiple threads.
17362 * The return value of RAND_load_file() no longer counts bytes obtained
17363 by stat(). RAND_load_file(..., -1) is new and uses the complete file
17364 to seed the PRNG (previously an explicit byte count was required).
17366 *Ulf Möller, Bodo Möller*
17368 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
17369 used `char *` instead of `void *` and had casts all over the place.
17373 * Make BN_generate_prime() return NULL on error if ret!=NULL.
17377 * Retain source code compatibility for BN_prime_checks macro:
17378 BN_is_prime(..., BN_prime_checks, ...) now uses
17379 BN_prime_checks_for_size to determine the appropriate number of
17380 Rabin-Miller iterations.
17384 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
17385 DH_CHECK_P_NOT_SAFE_PRIME.
17386 (Check if this is true? OpenPGP calls them "strong".)
17390 * Merge the functionality of "dh" and "gendh" programs into a new program
17391 "dhparam". The old programs are retained for now but will handle DH keys
17392 (instead of parameters) in future.
17396 * Make the ciphers, s_server and s_client programs check the return values
17397 when a new cipher list is set.
17401 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17402 ciphers. Before when the 56bit ciphers were enabled the sorting was
17405 The syntax for the cipher sorting has been extended to support sorting by
17406 cipher-strength (using the strength_bits hard coded in the tables).
17407 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17409 Fix a bug in the cipher-command parser: when supplying a cipher command
17410 string with an "undefined" symbol (neither command nor alphanumeric
17411 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17412 an error is flagged.
17414 Due to the strength-sorting extension, the code of the
17415 ssl_create_cipher_list() function was completely rearranged. I hope that
17416 the readability was also increased :-)
17418 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17420 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17421 for the first serial number and places 2 in the serial number file. This
17422 avoids problems when the root CA is created with serial number zero and
17423 the first user certificate has the same issuer name and serial number
17428 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17429 the new code. Add documentation for this stuff.
17433 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17434 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17435 structures and behave in an analogous way to the X509v3 functions:
17436 they shouldn't be called directly but wrapper functions should be used
17439 So we also now have some wrapper functions that call the X509at functions
17440 when passed certificate requests. (TO DO: similar things can be done with
17441 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17442 things. Some of these need some d2i or i2d and print functionality
17443 because they handle more complex structures.)
17447 * Add missing #ifndefs that caused missing symbols when building libssl
17448 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17449 NO_RSA in `ssl/s2*.c`.
17451 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17453 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17454 has a return value which indicates the quality of the random data
17455 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17456 error queue. New function RAND_pseudo_bytes() generates output that is
17457 guaranteed to be unique but not unpredictable. RAND_add is like
17458 RAND_seed, but takes an extra argument for an entropy estimate
17459 (RAND_seed always assumes full entropy).
17463 * Do more iterations of Rabin-Miller probable prime test (specifically,
17464 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17465 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17466 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17467 false-positive rate of at most 2^-80 for random input.
17471 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17475 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17476 in the 0.9.5 release), this returns the chain
17477 from an X509_CTX structure with a dup of the stack and all
17478 the X509 reference counts upped: so the stack will exist
17479 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17482 Also make SSL_SESSION_print() print out the verify return
17487 * Add manpage for the pkcs12 command. Also change the default
17488 behaviour so MAC iteration counts are used unless the new
17489 -nomaciter option is used. This improves file security and
17490 only older versions of MSIE (4.0 for example) need it.
17494 * Honor the no-xxx Configure options when creating .DEF files.
17498 * Add PKCS#10 attributes to field table: challengePassword,
17499 unstructuredName and unstructuredAddress. These are taken from
17500 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17501 international characters are used.
17503 More changes to X509_ATTRIBUTE code: allow the setting of types
17504 based on strings. Remove the 'loc' parameter when adding
17505 attributes because these will be a SET OF encoding which is sorted
17510 * Initial changes to the 'req' utility to allow request generation
17511 automation. This will allow an application to just generate a template
17512 file containing all the field values and have req construct the
17515 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17516 used all over the place including certificate requests and PKCS#7
17517 structures. They are currently handled manually where necessary with
17518 some primitive wrappers for PKCS#7. The new functions behave in a
17519 manner analogous to the X509 extension functions: they allow
17520 attributes to be looked up by NID and added.
17522 Later something similar to the X509V3 code would be desirable to
17523 automatically handle the encoding, decoding and printing of the
17524 more complex types. The string types like challengePassword can
17525 be handled by the string table functions.
17527 Also modified the multi byte string table handling. Now there is
17528 a 'global mask' which masks out certain types. The table itself
17529 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17530 is useful when for example there is only one permissible type
17531 (as in countryName) and using the mask might result in no valid
17536 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17537 SSL_get_peer_finished to allow applications to obtain the latest
17538 Finished messages sent to the peer or expected from the peer,
17539 respectively. (SSL_get_peer_finished is usually the Finished message
17540 actually received from the peer, otherwise the protocol will be aborted.)
17542 As the Finished message are message digests of the complete handshake
17543 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17544 be used for external authentication procedures when the authentication
17545 provided by SSL/TLS is not desired or is not enough.
17549 * Enhanced support for Alpha Linux is added. Now ./config checks if
17550 the host supports BWX extension and if Compaq C is present on the
17551 $PATH. Just exploiting of the BWX extension results in 20-30%
17552 performance kick for some algorithms, e.g. DES and RC4 to mention
17553 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17558 * Add support for MS "fast SGC". This is arguably a violation of the
17559 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17560 weak crypto and after checking the certificate is SGC a second one
17561 with strong crypto. MS SGC stops the first handshake after receiving
17562 the server certificate message and sends a second client hello. Since
17563 a server will typically do all the time consuming operations before
17564 expecting any further messages from the client (server key exchange
17565 is the most expensive) there is little difference between the two.
17567 To get OpenSSL to support MS SGC we have to permit a second client
17568 hello message after we have sent server done. In addition we have to
17569 reset the MAC if we do get this second client hello.
17573 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17574 if a DER encoded private key is RSA or DSA traditional format. Changed
17575 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17576 format DER encoded private key. Newer code should use PKCS#8 format which
17577 has the key type encoded in the ASN1 structure. Added DER private key
17578 support to pkcs8 application.
17582 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17583 ciphersuites has been selected (as required by the SSL 3/TLS 1
17584 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17585 is set, we interpret this as a request to violate the specification
17586 (the worst that can happen is a handshake failure, and 'correct'
17587 behaviour would result in a handshake failure anyway).
17591 * In SSL_CTX_add_session, take into account that there might be multiple
17592 SSL_SESSION structures with the same session ID (e.g. when two threads
17593 concurrently obtain them from an external cache).
17594 The internal cache can handle only one SSL_SESSION with a given ID,
17595 so if there's a conflict, we now throw out the old one to achieve
17600 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17601 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17602 some routines that use cipher OIDs: some ciphers do not have OIDs
17603 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17608 * Simplify the trust setting structure and code. Now we just have
17609 two sequences of OIDs for trusted and rejected settings. These will
17610 typically have values the same as the extended key usage extension
17611 and any application specific purposes.
17613 The trust checking code now has a default behaviour: it will just
17614 check for an object with the same NID as the passed id. Functions can
17615 be provided to override either the default behaviour or the behaviour
17616 for a given id. SSL client, server and email already have functions
17617 in place for compatibility: they check the NID and also return "trusted"
17618 if the certificate is self signed.
17622 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17623 traditional format into an EVP_PKEY structure.
17627 * Add a password callback function PEM_cb() which either prompts for
17628 a password if usr_data is NULL or otherwise assumes it is a null
17629 terminated password. Allow passwords to be passed on command line
17630 environment or config files in a few more utilities.
17634 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17635 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17636 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17637 Update documentation.
17641 * Support for ASN1 "NULL" type. This could be handled before by using
17642 ASN1_TYPE but there wasn't any function that would try to read a NULL
17643 and produce an error if it couldn't. For compatibility we also have
17644 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17645 don't allocate anything because they don't need to.
17649 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17652 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17654 * Rebuild of the memory allocation routines used by OpenSSL code and
17655 possibly others as well. The purpose is to make an interface that
17656 provide hooks so anyone can build a separate set of allocation and
17657 deallocation routines to be used by OpenSSL, for example memory
17658 pool implementations, or something else, which was previously hard
17659 since Malloc(), Realloc() and Free() were defined as macros having
17660 the values malloc, realloc and free, respectively (except for Win32
17661 compilations). The same is provided for memory debugging code.
17662 OpenSSL already comes with functionality to find memory leaks, but
17663 this gives people a chance to debug other memory problems.
17665 With these changes, a new set of functions and macros have appeared:
17667 CRYPTO_set_mem_debug_functions() [F]
17668 CRYPTO_get_mem_debug_functions() [F]
17669 CRYPTO_dbg_set_options() [F]
17670 CRYPTO_dbg_get_options() [F]
17671 CRYPTO_malloc_debug_init() [M]
17673 The memory debug functions are NULL by default, unless the library
17674 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17675 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17676 gives the standard debugging functions that come with OpenSSL) or
17677 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17678 provided by the library user) must be used. When the standard
17679 debugging functions are used, CRYPTO_dbg_set_options can be used to
17680 request additional information:
17681 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17682 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17684 Also, things like CRYPTO_set_mem_functions will always give the
17685 expected result (the new set of functions is used for allocation
17686 and deallocation) at all times, regardless of platform and compiler
17689 To finish it up, some functions that were never use in any other
17690 way than through macros have a new API and new semantic:
17692 CRYPTO_dbg_malloc()
17693 CRYPTO_dbg_realloc()
17696 All macros of value have retained their old syntax.
17698 *Richard Levitte and Bodo Moeller*
17700 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17701 ordering of SMIMECapabilities wasn't in "strength order" and there
17702 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17707 * Some ASN1 types with illegal zero length encoding (INTEGER,
17708 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17710 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17712 * Merge in my S/MIME library for OpenSSL. This provides a simple
17713 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17714 functionality to handle multipart/signed properly) and a utility
17715 called 'smime' to call all this stuff. This is based on code I
17716 originally wrote for Celo who have kindly allowed it to be
17717 included in OpenSSL.
17721 * Add variants des_set_key_checked and des_set_key_unchecked of
17722 des_set_key (aka des_key_sched). Global variable des_check_key
17723 decides which of these is called by des_set_key; this way
17724 des_check_key behaves as it always did, but applications and
17725 the library itself, which was buggy for des_check_key == 1,
17726 have a cleaner way to pick the version they need.
17730 * New function PKCS12_newpass() which changes the password of a
17735 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17736 dynamic mix. In both cases the ids can be used as an index into the
17737 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17738 functions so they accept a list of the field values and the
17739 application doesn't need to directly manipulate the X509_TRUST
17744 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17749 * Modify the way the V3 extension code looks up extensions. This now
17750 works in a similar way to the object code: we have some "standard"
17751 extensions in a static table which is searched with OBJ_bsearch()
17752 and the application can add dynamic ones if needed. The file
17753 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17754 updated whenever a new extension is added to the core code and kept
17755 in ext_nid order. There is a simple program 'tabtest.c' which checks
17756 this. New extensions are not added too often so this file can readily
17757 be maintained manually.
17759 There are two big advantages in doing things this way. The extensions
17760 can be looked up immediately and no longer need to be "added" using
17761 X509V3_add_standard_extensions(): this function now does nothing.
17762 Side note: I get *lots* of email saying the extension code doesn't
17763 work because people forget to call this function.
17764 Also no dynamic allocation is done unless new extensions are added:
17765 so if we don't add custom extensions there is no need to call
17766 X509V3_EXT_cleanup().
17770 * Modify enc utility's salting as follows: make salting the default. Add a
17771 magic header, so unsalted files fail gracefully instead of just decrypting
17772 to garbage. This is because not salting is a big security hole, so people
17773 should be discouraged from doing it.
17777 * Fixes and enhancements to the 'x509' utility. It allowed a message
17778 digest to be passed on the command line but it only used this
17779 parameter when signing a certificate. Modified so all relevant
17780 operations are affected by the digest parameter including the
17781 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17782 DSA key was used because it didn't fix the digest.
17786 * Initial certificate chain verify code. Currently tests the untrusted
17787 certificates for consistency with the verify purpose (which is set
17788 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17790 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17791 this is because it will reject chains with invalid extensions whereas
17792 every previous version of OpenSSL and SSLeay made no checks at all.
17794 Trust code: checks the root CA for the relevant trust settings. Trust
17795 settings have an initial value consistent with the verify purpose: e.g.
17796 if the verify purpose is for SSL client use it expects the CA to be
17797 trusted for SSL client use. However the default value can be changed to
17798 permit custom trust settings: one example of this would be to only trust
17799 certificates from a specific "secure" set of CAs.
17801 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17802 which should be used for version portability: especially since the
17803 verify structure is likely to change more often now.
17805 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17806 to set them. If not set then assume SSL clients will verify SSL servers
17809 Two new options to the verify program: -untrusted allows a set of
17810 untrusted certificates to be passed in and -purpose which sets the
17811 intended purpose of the certificate. If a purpose is set then the
17812 new chain verify code is used to check extension consistency.
17816 * Support for the authority information access extension.
17820 * Modify RSA and DSA PEM read routines to transparently handle
17821 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17822 public keys in a format compatible with certificate
17823 SubjectPublicKeyInfo structures. Unfortunately there were already
17824 functions called *_PublicKey_* which used various odd formats so
17825 these are retained for compatibility: however the DSA variants were
17826 never in a public release so they have been deleted. Changed dsa/rsa
17827 utilities to handle the new format: note no releases ever handled public
17828 keys so we should be OK.
17830 The primary motivation for this change is to avoid the same fiasco
17831 that dogs private keys: there are several incompatible private key
17832 formats some of which are standard and some OpenSSL specific and
17833 require various evil hacks to allow partial transparent handling and
17834 even then it doesn't work with DER formats. Given the option anything
17835 other than PKCS#8 should be dumped: but the other formats have to
17836 stay in the name of compatibility.
17838 With public keys and the benefit of hindsight one standard format
17839 is used which works with EVP_PKEY, RSA or DSA structures: though
17840 it clearly returns an error if you try to read the wrong kind of key.
17842 Added a -pubkey option to the 'x509' utility to output the public key.
17843 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17844 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17845 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17846 that do the same as the `EVP_PKEY_assign_*()` except they up the
17847 reference count of the added key (they don't "swallow" the
17852 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17853 CRLs would fail if the file contained no certificates or no CRLs:
17854 added a new function to read in both types and return the number
17855 read: this means that if none are read it will be an error. The
17856 DER versions of the certificate and CRL reader would always fail
17857 because it isn't possible to mix certificates and CRLs in DER format
17858 without choking one or the other routine. Changed this to just read
17859 a certificate: this is the best we can do. Also modified the code
17860 in `apps/verify.c` to take notice of return codes: it was previously
17861 attempting to read in certificates from NULL pointers and ignoring
17862 any errors: this is one reason why the cert and CRL reader seemed
17863 to work. It doesn't check return codes from the default certificate
17864 routines: these may well fail if the certificates aren't installed.
17868 * Code to support otherName option in GeneralName.
17872 * First update to verify code. Change the verify utility
17873 so it warns if it is passed a self signed certificate:
17874 for consistency with the normal behaviour. X509_verify
17875 has been modified to it will now verify a self signed
17876 certificate if *exactly* the same certificate appears
17877 in the store: it was previously impossible to trust a
17878 single self signed certificate. This means that:
17879 openssl verify ss.pem
17880 now gives a warning about a self signed certificate but
17881 openssl verify -CAfile ss.pem ss.pem
17886 * For servers, store verify_result in SSL_SESSION data structure
17887 (and add it to external session representation).
17888 This is needed when client certificate verifications fails,
17889 but an application-provided verification callback (set by
17890 SSL_CTX_set_cert_verify_callback) allows accepting the session
17891 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17892 but returns 1): When the session is reused, we have to set
17893 ssl->verify_result to the appropriate error code to avoid
17896 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17898 * Fix a bug in the new PKCS#7 code: it didn't consider the
17899 case in PKCS7_dataInit() where the signed PKCS7 structure
17900 didn't contain any existing data because it was being created.
17902 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17904 * Add a salt to the key derivation routines in enc.c. This
17905 forms the first 8 bytes of the encrypted file. Also add a
17906 -S option to allow a salt to be input on the command line.
17910 * New function X509_cmp(). Oddly enough there wasn't a function
17911 to compare two certificates. We do this by working out the SHA1
17912 hash and comparing that. X509_cmp() will be needed by the trust
17917 * SSL_get1_session() is like SSL_get_session(), but increments
17918 the reference count in the SSL_SESSION returned.
17920 *Geoff Thorpe <geoff@eu.c2.net>*
17922 * Fix for 'req': it was adding a null to request attributes.
17923 Also change the X509_LOOKUP and X509_INFO code to handle
17924 certificate auxiliary information.
17928 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17933 * Add the possibility to add extra information to the memory leak
17934 detecting output, to form tracebacks, showing from where each
17935 allocation was originated: CRYPTO_push_info("constant string") adds
17936 the string plus current file name and line number to a per-thread
17937 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17938 is like calling CYRPTO_pop_info() until the stack is empty.
17939 Also updated memory leak detection code to be multi-thread-safe.
17943 * Add options -text and -noout to pkcs7 utility and delete the
17944 encryption options which never did anything. Update docs.
17948 * Add options to some of the utilities to allow the pass phrase
17949 to be included on either the command line (not recommended on
17950 OSes like Unix) or read from the environment. Update the
17951 manpages and fix a few bugs.
17955 * Add a few manpages for some of the openssl commands.
17959 * Fix the -revoke option in ca. It was freeing up memory twice,
17960 leaking and not finding already revoked certificates.
17964 * Extensive changes to support certificate auxiliary information.
17965 This involves the use of X509_CERT_AUX structure and X509_AUX
17966 functions. An X509_AUX function such as PEM_read_X509_AUX()
17967 can still read in a certificate file in the usual way but it
17968 will also read in any additional "auxiliary information". By
17969 doing things this way a fair degree of compatibility can be
17970 retained: existing certificates can have this information added
17971 using the new 'x509' options.
17973 Current auxiliary information includes an "alias" and some trust
17974 settings. The trust settings will ultimately be used in enhanced
17975 certificate chain verification routines: currently a certificate
17976 can only be trusted if it is self signed and then it is trusted
17981 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17982 The problem was that one of the replacement routines had not been working
17983 since SSLeay releases. For now the offending routine has been replaced
17984 with non-optimised assembler. Even so, this now gives around 95%
17985 performance improvement for 1024 bit RSA signs.
17989 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17990 handling. Most clients have the effective key size in bits equal to
17991 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17992 A few however don't do this and instead use the size of the decrypted key
17993 to determine the RC2 key length and the AlgorithmIdentifier to determine
17994 the effective key length. In this case the effective key length can still
17995 be 40 bits but the key length can be 168 bits for example. This is fixed
17996 by manually forcing an RC2 key into the EVP_PKEY structure because the
17997 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17998 the key length and effective key length are equal.
18002 * Add a bunch of functions that should simplify the creation of
18003 X509_NAME structures. Now you should be able to do:
18004 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
18005 and have it automatically work out the correct field type and fill in
18006 the structures. The more adventurous can try:
18007 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
18008 and it will (hopefully) work out the correct multibyte encoding.
18012 * Change the 'req' utility to use the new field handling and multibyte
18013 copy routines. Before the DN field creation was handled in an ad hoc
18014 way in req, ca, and x509 which was rather broken and didn't support
18015 BMPStrings or UTF8Strings. Since some software doesn't implement
18016 BMPStrings or UTF8Strings yet, they can be enabled using the config file
18017 using the dirstring_type option. See the new comment in the default
18018 openssl.cnf for more info.
18022 * Make crypto/rand/md_rand.c more robust:
18023 - Assure unique random numbers after fork().
18024 - Make sure that concurrent threads access the global counter and
18025 md serializably so that we never lose entropy in them
18026 or use exactly the same state in multiple threads.
18027 Access to the large state is not always serializable because
18028 the additional locking could be a performance killer, and
18029 md should be large enough anyway.
18033 * New file `apps/app_rand.c` with commonly needed functionality
18034 for handling the random seed file.
18036 Use the random seed file in some applications that previously did not:
18038 dsaparam -genkey (which also ignored its '-rand' option),
18041 x509 (when signing).
18042 Except on systems with /dev/urandom, it is crucial to have a random
18043 seed file at least for key creation, DSA signing, and for DH exchanges;
18044 for RSA signatures we could do without one.
18046 gendh and gendsa (unlike genrsa) used to read only the first byte
18047 of each file listed in the '-rand' option. The function as previously
18048 found in genrsa is now in app_rand.c and is used by all programs
18049 that support '-rand'.
18053 * In RAND_write_file, use mode 0600 for creating files;
18054 don't just chmod when it may be too late.
18058 * Report an error from X509_STORE_load_locations
18059 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
18063 * New function ASN1_mbstring_copy() this copies a string in either
18064 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
18065 into an ASN1_STRING type. A mask of permissible types is passed
18066 and it chooses the "minimal" type to use or an error if not type
18071 * Add function equivalents to the various macros in asn1.h. The old
18072 macros are retained with an `M_` prefix. Code inside the library can
18073 use the `M_` macros. External code (including the openssl utility)
18074 should *NOT* in order to be "shared library friendly".
18078 * Add various functions that can check a certificate's extensions
18079 to see if it usable for various purposes such as SSL client,
18080 server or S/MIME and CAs of these types. This is currently
18081 VERY EXPERIMENTAL but will ultimately be used for certificate chain
18082 verification. Also added a -purpose flag to x509 utility to
18083 print out all the purposes.
18087 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
18092 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
18093 for, obtain and decode and extension and obtain its critical flag.
18094 This allows all the necessary extension code to be handled in a
18095 single function call.
18099 * RC4 tune-up featuring 30-40% performance improvement on most RISC
18100 platforms. See crypto/rc4/rc4_enc.c for further details.
18104 * New -noout option to asn1parse. This causes no output to be produced
18105 its main use is when combined with -strparse and -out to extract data
18106 from a file (which may not be in ASN.1 format).
18110 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
18111 when producing the local key id.
18113 *Richard Levitte <levitte@stacken.kth.se>*
18115 * New option -dhparam in s_server. This allows a DH parameter file to be
18116 stated explicitly. If it is not stated then it tries the first server
18117 certificate file. The previous behaviour hard coded the filename
18122 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
18123 a public key to be input or output. For example:
18124 openssl rsa -in key.pem -pubout -out pubkey.pem
18125 Also added necessary DSA public key functions to handle this.
18129 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
18130 in the message. This was handled by allowing
18131 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
18133 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
18135 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
18136 to the end of the strings whereas this didn't. This would cause problems
18137 if strings read with d2i_ASN1_bytes() were later modified.
18139 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
18141 * Fix for base64 decode bug. When a base64 bio reads only one line of
18142 data and it contains EOF it will end up returning an error. This is
18143 caused by input 46 bytes long. The cause is due to the way base64
18144 BIOs find the start of base64 encoded data. They do this by trying a
18145 trial decode on each line until they find one that works. When they
18146 do a flag is set and it starts again knowing it can pass all the
18147 data directly through the decoder. Unfortunately it doesn't reset
18148 the context it uses. This means that if EOF is reached an attempt
18149 is made to pass two EOFs through the context and this causes the
18150 resulting error. This can also cause other problems as well. As is
18151 usual with these problems it takes *ages* to find and the fix is
18152 trivial: move one line.
18154 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
18156 * Ugly workaround to get s_client and s_server working under Windows. The
18157 old code wouldn't work because it needed to select() on sockets and the
18158 tty (for keypresses and to see if data could be written). Win32 only
18159 supports select() on sockets so we select() with a 1s timeout on the
18160 sockets and then see if any characters are waiting to be read, if none
18161 are present then we retry, we also assume we can always write data to
18162 the tty. This isn't nice because the code then blocks until we've
18163 received a complete line of data and it is effectively polling the
18164 keyboard at 1s intervals: however it's quite a bit better than not
18165 working at all :-) A dedicated Windows application might handle this
18166 with an event loop for example.
18170 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
18171 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
18172 will be called when RSA_sign() and RSA_verify() are used. This is useful
18173 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
18174 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
18175 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
18176 This necessitated the support of an extra signature type NID_md5_sha1
18177 for SSL signatures and modifications to the SSL library to use it instead
18178 of calling RSA_public_decrypt() and RSA_private_encrypt().
18182 * Add new -verify -CAfile and -CApath options to the crl program, these
18183 will lookup a CRL issuers certificate and verify the signature in a
18184 similar way to the verify program. Tidy up the crl program so it
18185 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
18186 less strict. It will now permit CRL extensions even if it is not
18187 a V2 CRL: this will allow it to tolerate some broken CRLs.
18191 * Initialize all non-automatic variables each time one of the openssl
18192 sub-programs is started (this is necessary as they may be started
18193 multiple times from the "OpenSSL>" prompt).
18195 *Lennart Bang, Bodo Moeller*
18197 * Preliminary compilation option RSA_NULL which disables RSA crypto without
18198 removing all other RSA functionality (this is what NO_RSA does). This
18199 is so (for example) those in the US can disable those operations covered
18200 by the RSA patent while allowing storage and parsing of RSA keys and RSA
18205 * Non-copying interface to BIO pairs.
18206 (still largely untested)
18210 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
18211 ASCII string. This was handled independently in various places before.
18215 * New functions UTF8_getc() and UTF8_putc() that parse and generate
18216 UTF8 strings a character at a time.
18220 * Use client_version from client hello to select the protocol
18221 (s23_srvr.c) and for RSA client key exchange verification
18222 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
18226 * Add various utility functions to handle SPKACs, these were previously
18227 handled by poking round in the structure internals. Added new function
18228 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
18229 print, verify and generate SPKACs. Based on an original idea from
18230 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
18234 * RIPEMD160 is operational on all platforms and is back in 'make test'.
18238 * Allow the config file extension section to be overwritten on the
18239 command line. Based on an original idea from Massimiliano Pala
18240 <madwolf@comune.modena.it>. The new option is called -extensions
18241 and can be applied to ca, req and x509. Also -reqexts to override
18242 the request extensions in req and -crlexts to override the crl extensions
18247 * Add new feature to the SPKAC handling in ca. Now you can include
18248 the same field multiple times by preceding it by "XXXX." for example:
18251 this is the same syntax as used in the req config file.
18255 * Allow certificate extensions to be added to certificate requests. These
18256 are specified in a 'req_extensions' option of the req section of the
18257 config file. They can be printed out with the -text option to req but
18258 are otherwise ignored at present.
18262 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
18263 data read consists of only the final block it would not decrypted because
18264 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
18265 A misplaced 'break' also meant the decrypted final block might not be
18266 copied until the next read.
18270 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
18271 a few extra parameters to the DH structure: these will be useful if
18272 for example we want the value of 'q' or implement X9.42 DH.
18276 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
18277 provides hooks that allow the default DSA functions or functions on a
18278 "per key" basis to be replaced. This allows hardware acceleration and
18279 hardware key storage to be handled without major modification to the
18280 library. Also added low-level modexp hooks and CRYPTO_EX structure and
18281 associated functions.
18285 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
18286 as "read only": it can't be written to and the buffer it points to will
18287 not be freed. Reading from a read only BIO is much more efficient than
18288 a normal memory BIO. This was added because there are several times when
18289 an area of memory needs to be read from a BIO. The previous method was
18290 to create a memory BIO and write the data to it, this results in two
18291 copies of the data and an O(n^2) reading algorithm. There is a new
18292 function BIO_new_mem_buf() which creates a read only memory BIO from
18293 an area of memory. Also modified the PKCS#7 routines to use read only
18298 * Bugfix: ssl23_get_client_hello did not work properly when called in
18299 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
18300 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
18301 but a retry condition occurred while trying to read the rest.
18305 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
18306 NID_pkcs7_encrypted by default: this was wrong since this should almost
18307 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
18308 the encrypted data type: this is a more sensible place to put it and it
18309 allows the PKCS#12 code to be tidied up that duplicated this
18314 * Changed obj_dat.pl script so it takes its input and output files on
18315 the command line. This should avoid shell escape redirection problems
18320 * Initial support for certificate extension requests, these are included
18321 in things like Xenroll certificate requests. Included functions to allow
18322 extensions to be obtained and added.
18326 * -crlf option to s_client and s_server for sending newlines as
18327 CRLF (as required by many protocols).
18331 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
18333 * Install libRSAglue.a when OpenSSL is built with RSAref.
18335 *Ralf S. Engelschall*
18337 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
18339 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
18341 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
18346 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
18347 DH parameters/keys (q is lost during that conversion, but the resulting
18348 DH parameters contain its length).
18350 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
18351 much faster than DH_generate_parameters (which creates parameters
18352 where `p = 2*q + 1`), and also the smaller q makes DH computations
18353 much more efficient (160-bit exponentiation instead of 1024-bit
18354 exponentiation); so this provides a convenient way to support DHE
18355 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
18356 utter importance to use
18357 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18359 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18360 when such DH parameters are used, because otherwise small subgroup
18361 attacks may become possible!
18365 * Avoid memory leak in i2d_DHparams.
18369 * Allow the -k option to be used more than once in the enc program:
18370 this allows the same encrypted message to be read by multiple recipients.
18374 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
18375 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
18376 it will always use the numerical form of the OID, even if it has a short
18381 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
18382 method only got called if p,q,dmp1,dmq1,iqmp components were present,
18383 otherwise bn_mod_exp was called. In the case of hardware keys for example
18384 no private key components need be present and it might store extra data
18385 in the RSA structure, which cannot be accessed from bn_mod_exp.
18386 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
18387 private key operations.
18391 * Added support for SPARC Linux.
18395 * pem_password_cb function type incompatibly changed from
18396 typedef int pem_password_cb(char *buf, int size, int rwflag);
18398 ....(char *buf, int size, int rwflag, void *userdata);
18399 so that applications can pass data to their callbacks:
18400 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18401 additional void * argument, which is just handed through whenever
18402 the password callback is called.
18404 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18406 New function SSL_CTX_set_default_passwd_cb_userdata.
18408 Compatibility note: As many C implementations push function arguments
18409 onto the stack in reverse order, the new library version is likely to
18410 interoperate with programs that have been compiled with the old
18411 pem_password_cb definition (PEM_whatever takes some data that
18412 happens to be on the stack as its last argument, and the callback
18413 just ignores this garbage); but there is no guarantee whatsoever that
18416 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18417 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18418 problems not only on Windows, but also on some Unix platforms.
18419 To avoid problematic command lines, these definitions are now in an
18420 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18421 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18425 * MIPS III/IV assembler module is reimplemented.
18429 * More DES library cleanups: remove references to srand/rand and
18430 delete an unused file.
18434 * Add support for the free Netwide assembler (NASM) under Win32,
18435 since not many people have MASM (ml) and it can be hard to obtain.
18436 This is currently experimental but it seems to work OK and pass all
18437 the tests. Check out INSTALL.W32 for info.
18441 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18442 without temporary keys kept an extra copy of the server key,
18443 and connections with temporary keys did not free everything in case
18448 * New function RSA_check_key and new openssl rsa option -check
18449 for verifying the consistency of RSA keys.
18451 *Ulf Moeller, Bodo Moeller*
18453 * Various changes to make Win32 compile work:
18454 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18455 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18456 comparison" warnings.
18457 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18461 * Add a debugging option to PKCS#5 v2 key generation function: when
18462 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18463 derived keys are printed to stderr.
18467 * Copy the flags in ASN1_STRING_dup().
18469 *Roman E. Pavlov <pre@mo.msk.ru>*
18471 * The x509 application mishandled signing requests containing DSA
18472 keys when the signing key was also DSA and the parameters didn't match.
18474 It was supposed to omit the parameters when they matched the signing key:
18475 the verifying software was then supposed to automatically use the CA's
18476 parameters if they were absent from the end user certificate.
18478 Omitting parameters is no longer recommended. The test was also
18479 the wrong way round! This was probably due to unusual behaviour in
18480 EVP_cmp_parameters() which returns 1 if the parameters match.
18481 This meant that parameters were omitted when they *didn't* match and
18482 the certificate was useless. Certificates signed with 'ca' didn't have
18485 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18487 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18488 The interface is as follows:
18489 Applications can use
18490 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18491 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18492 "off" is now the default.
18493 The library internally uses
18494 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18495 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18496 to disable memory-checking temporarily.
18498 Some inconsistent states that previously were possible (and were
18499 even the default) are now avoided.
18501 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18502 with each memory chunk allocated; this is occasionally more helpful
18503 than just having a counter.
18505 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18507 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18512 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18513 which largely parallels "options", but is for changing API behaviour,
18514 whereas "options" are about protocol behaviour.
18515 Initial "mode" flags are:
18517 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18518 a single record has been written.
18519 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18520 retries use the same buffer location.
18521 (But all of the contents must be
18526 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18529 * Fix problems with no-hmac etc.
18531 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18533 * New functions RSA_get_default_method(), RSA_set_method() and
18534 RSA_get_method(). These allows replacement of RSA_METHODs without having
18535 to mess around with the internals of an RSA structure.
18539 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18540 Also really enable memory leak checks in openssl.c and in some
18543 *Chad C. Mulligan, Bodo Moeller*
18545 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18546 up the length of negative integers. This has now been simplified to just
18547 store the length when it is first determined and use it later, rather
18548 than trying to keep track of where data is copied and updating it to
18550 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18552 * Add a new function PKCS7_signatureVerify. This allows the verification
18553 of a PKCS#7 signature but with the signing certificate passed to the
18554 function itself. This contrasts with PKCS7_dataVerify which assumes the
18555 certificate is present in the PKCS#7 structure. This isn't always the
18556 case: certificates can be omitted from a PKCS#7 structure and be
18557 distributed by "out of band" means (such as a certificate database).
18561 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18562 function prototypes in pem.h, also change util/mkdef.pl to add the
18563 necessary function names.
18567 * mk1mf.pl (used by Windows builds) did not properly read the
18568 options set by Configure in the top level Makefile, and Configure
18569 was not even able to write more than one option correctly.
18570 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18574 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18575 file to be loaded from a BIO or FILE pointer. The BIO version will
18576 for example allow memory BIOs to contain config info.
18580 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18581 Whoever hopes to achieve shared-library compatibility across versions
18582 must use this, not the compile-time macro.
18583 (Exercise 0.9.4: Which is the minimum library version required by
18585 Note: All this applies only to multi-threaded programs, others don't
18590 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18591 through a BIO pair triggered the default case, i.e.
18592 SSLerr(...,SSL_R_UNKNOWN_STATE).
18596 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18597 can use the SSL library even if none of the specific BIOs is
18602 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18603 for the encoded length.
18605 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18607 * Add initial documentation of the X509V3 functions.
18611 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18612 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18613 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18614 secure PKCS#8 private key format with a high iteration count.
18618 * Fix determination of Perl interpreter: A perl or perl5
18619 *directory* in $PATH was also accepted as the interpreter.
18621 *Ralf S. Engelschall*
18623 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18624 wrong with it but it was very old and did things like calling
18625 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18626 unusual formatting.
18630 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18631 to use the new extension code.
18635 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18636 with macros. This should make it easier to change their form, add extra
18637 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18642 * Add to configuration table a new entry that can specify an alternative
18643 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18644 according to Mark Crispin <MRC@Panda.COM>.
18648 * DES CBC did not update the IV. Weird.
18652 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18653 Changing the behaviour of the former might break existing programs --
18654 where IV updating is needed, des_ncbc_encrypt can be used.
18657 * When bntest is run from "make test" it drives bc to check its
18658 calculations, as well as internally checking them. If an internal check
18659 fails, it needs to cause bc to give a non-zero result or make test carries
18660 on without noticing the failure. Fixed.
18664 * DES library cleanups.
18668 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18669 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18670 ciphers. NOTE: although the key derivation function has been verified
18671 against some published test vectors it has not been extensively tested
18672 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18677 * Instead of "mkdir -p", which is not fully portable, use new
18678 Perl script "util/mkdir-p.pl".
18682 * Rewrite the way password based encryption (PBE) is handled. It used to
18683 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18684 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18685 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18686 the 'parameter' field of the AlgorithmIdentifier is passed to the
18687 underlying key generation function so it must do its own ASN1 parsing.
18688 This has also changed the EVP_PBE_CipherInit() function which now has a
18689 'parameter' argument instead of literal salt and iteration count values
18690 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18694 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18695 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18696 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18697 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18698 value was just used as a "magic string" and not used directly its
18699 value doesn't matter.
18703 * Introduce some semblance of const correctness to BN. Shame C doesn't
18708 * "linux-sparc64" configuration (ultrapenguin).
18710 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18711 "linux-sparc" configuration.
18713 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18715 * config now generates no-xxx options for missing ciphers.
18719 * Support the EBCDIC character set (work in progress).
18720 File ebcdic.c not yet included because it has a different license.
18722 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18724 * Support BS2000/OSD-POSIX.
18726 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18728 * Make callbacks for key generation use `void *` instead of `char *`.
18732 * Make S/MIME samples compile (not yet tested).
18736 * Additional typesafe stacks.
18740 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18744 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18746 * New configuration variant "sco5-gcc".
18748 * Updated some demos.
18750 *Sean O Riordain, Wade Scholine*
18752 * Add missing BIO_free at exit of pkcs12 application.
18756 * Fix memory leak in conf.c.
18760 * Updates for Win32 to assembler version of MD5.
18764 * Set #! path to perl in `apps/der_chop` to where we found it
18765 instead of using a fixed path.
18769 * SHA library changes for irix64-mips4-cc.
18773 * Improvements for VMS support.
18777 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18779 * Bignum library bug fix. IRIX 6 passes "make test" now!
18780 This also avoids the problems with SC4.2 and unpatched SC5.
18782 *Andy Polyakov <appro@fy.chalmers.se>*
18784 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18785 These are required because of the typesafe stack would otherwise break
18786 existing code. If old code used a structure member which used to be STACK
18787 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18788 sk_num or sk_value it would produce an error because the num, data members
18789 are not present in STACK_OF. Now it just produces a warning. sk_set
18790 replaces the old method of assigning a value to sk_value
18791 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18792 that does this will no longer work (and should use sk_set instead) but
18793 this could be regarded as a "questionable" behaviour anyway.
18797 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18798 correctly handle encrypted S/MIME data.
18802 * Change type of various DES function arguments from des_cblock
18803 (which means, in function argument declarations, pointer to char)
18804 to des_cblock * (meaning pointer to array with 8 char elements),
18805 which allows the compiler to do more typechecking; it was like
18806 that back in SSLeay, but with lots of ugly casts.
18808 Introduce new type const_des_cblock.
18812 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18813 problems: find RecipientInfo structure that matches recipient certificate
18814 and initialise the ASN1 structures properly based on passed cipher.
18818 * Belatedly make the BN tests actually check the results.
18822 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18823 to and from BNs: it was completely broken. New compilation option
18824 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18825 key elements as negative integers.
18829 * Reorganize and speed up MD5.
18831 *Andy Polyakov <appro@fy.chalmers.se>*
18835 *Richard Levitte <richard@levitte.org>*
18837 * New option -out to asn1parse to allow the parsed structure to be
18838 output to a file. This is most useful when combined with the -strparse
18839 option to examine the output of things like OCTET STRINGS.
18843 * Make SSL library a little more fool-proof by not requiring any longer
18844 that `SSL_set_{accept,connect}_state` be called before
18845 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18846 in many applications because usually everything *appeared* to work as
18847 intended anyway -- now it really works as intended).
18851 * Move openssl.cnf out of lib/.
18855 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18856 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18857 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18859 *Ralf S. Engelschall*
18861 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18862 handle PKCS#7 enveloped data properly.
18864 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18866 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18867 copying pointers. The cert_st handling is changed by this in
18868 various ways (and thus what used to be known as ctx->default_cert
18869 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18870 any longer when s->cert does not give us what we need).
18871 ssl_cert_instantiate becomes obsolete by this change.
18872 As soon as we've got the new code right (possibly it already is?),
18873 we have solved a couple of bugs of the earlier code where s->cert
18874 was used as if it could not have been shared with other SSL structures.
18876 Note that using the SSL API in certain dirty ways now will result
18877 in different behaviour than observed with earlier library versions:
18878 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18879 does not influence s as it used to.
18881 In order to clean up things more thoroughly, inside SSL_SESSION
18882 we don't use CERT any longer, but a new structure SESS_CERT
18883 that holds per-session data (if available); currently, this is
18884 the peer's certificate chain and, for clients, the server's certificate
18885 and temporary key. CERT holds only those values that can have
18886 meaningful defaults in an SSL_CTX.
18890 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18891 from the internal representation. Various PKCS#7 fixes: remove some
18892 evil casts and set the enc_dig_alg field properly based on the signing
18897 * Allow PKCS#12 password to be set from the command line or the
18898 environment. Let 'ca' get its config file name from the environment
18899 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18904 * Allow certificate policies extension to use an IA5STRING for the
18905 organization field. This is contrary to the PKIX definition but
18906 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18911 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18912 without disallowing inline assembler and the like for non-pedantic builds.
18916 * Support Borland C++ builder.
18918 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18924 * SHA-1 cleanups and performance enhancements.
18926 *Andy Polyakov <appro@fy.chalmers.se>*
18928 * Sparc v8plus assembler for the bignum library.
18930 *Andy Polyakov <appro@fy.chalmers.se>*
18932 * Accept any -xxx and +xxx compiler options in Configure.
18936 * Update HPUX configuration.
18940 * Add missing `sk_<type>_unshift()` function to safestack.h
18942 *Ralf S. Engelschall*
18944 * New function SSL_CTX_use_certificate_chain_file that sets the
18945 "extra_cert"s in addition to the certificate. (This makes sense
18946 only for "PEM" format files, as chains as a whole are not
18951 * Support verify_depth from the SSL API.
18952 x509_vfy.c had what can be considered an off-by-one-error:
18953 Its depth (which was not part of the external interface)
18954 was actually counting the number of certificates in a chain;
18955 now it really counts the depth.
18959 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18960 instead of X509err, which often resulted in confusing error
18961 messages since the error codes are not globally unique
18962 (e.g. an alleged error in ssl3_accept when a certificate
18963 didn't match the private key).
18965 * New function SSL_CTX_set_session_id_context that allows to set a default
18966 value (so that you don't need SSL_set_session_id_context for each
18967 connection using the SSL_CTX).
18971 * OAEP decoding bug fix.
18975 * Support INSTALL_PREFIX for package builders, as proposed by
18980 * New Configure options "threads" and "no-threads". For systems
18981 where the proper compiler options are known (currently Solaris
18982 and Linux), "threads" is the default.
18986 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18990 * Install various scripts to $(OPENSSLDIR)/misc, not to
18991 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18992 such as /usr/local/bin.
18996 * "make linux-shared" to build shared libraries.
18998 *Niels Poppe <niels@netbox.org>*
19000 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
19004 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
19005 extension adding in x509 utility.
19009 * Remove NOPROTO sections and error code comments.
19013 * Partial rewrite of the DEF file generator to now parse the ANSI
19018 * New Configure options --prefix=DIR and --openssldir=DIR.
19022 * Complete rewrite of the error code script(s). It is all now handled
19023 by one script at the top level which handles error code gathering,
19024 header rewriting and C source file generation. It should be much better
19025 than the old method: it now uses a modified version of Ulf's parser to
19026 read the ANSI prototypes in all header files (thus the old K&R definitions
19027 aren't needed for error creation any more) and do a better job of
19028 translating function codes into names. The old 'ASN1 error code embedded
19029 in a comment' is no longer necessary and it doesn't use .err files which
19030 have now been deleted. Also the error code call doesn't have to appear all
19031 on one line (which resulted in some large lines...).
19035 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
19039 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
19040 0 (which usually indicates a closed connection), but continue reading.
19044 * Fix some race conditions.
19048 * Add support for CRL distribution points extension. Add Certificate
19049 Policies and CRL distribution points documentation.
19053 * Move the autogenerated header file parts to crypto/opensslconf.h.
19057 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
19058 8 of keying material. Merlin has also confirmed interop with this fix
19059 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
19061 *Merlin Hughes <merlin@baltimore.ie>*
19063 * Fix lots of warnings.
19065 *Richard Levitte <levitte@stacken.kth.se>*
19067 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
19068 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
19070 *Richard Levitte <levitte@stacken.kth.se>*
19072 * Fix problems with sizeof(long) == 8.
19074 *Andy Polyakov <appro@fy.chalmers.se>*
19076 * Change functions to ANSI C.
19080 * Fix typos in error codes.
19082 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
19084 * Remove defunct assembler files from Configure.
19088 * SPARC v8 assembler BIGNUM implementation.
19090 *Andy Polyakov <appro@fy.chalmers.se>*
19092 * Support for Certificate Policies extension: both print and set.
19093 Various additions to support the r2i method this uses.
19097 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
19098 return a const string when you are expecting an allocated buffer.
19102 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
19103 types DirectoryString and DisplayText.
19107 * Add code to allow r2i extensions to access the configuration database,
19108 add an LHASH database driver and add several ctx helper functions.
19112 * Fix an evil bug in bn_expand2() which caused various BN functions to
19113 fail when they extended the size of a BIGNUM.
19117 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
19118 support typesafe stack.
19122 * Fix typo in SSL_[gs]et_options().
19124 *Nils Frostberg <nils@medcom.se>*
19126 * Delete various functions and files that belonged to the (now obsolete)
19127 old X509V3 handling code.
19131 * New Configure option "rsaref".
19135 * Don't auto-generate pem.h.
19139 * Introduce type-safe ASN.1 SETs.
19143 * Convert various additional casted stacks to type-safe STACK_OF() variants.
19145 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
19147 * Introduce type-safe STACKs. This will almost certainly break lots of code
19148 that links with OpenSSL (well at least cause lots of warnings), but fear
19149 not: the conversion is trivial, and it eliminates loads of evil casts. A
19150 few STACKed things have been converted already. Feel free to convert more.
19151 In the fullness of time, I'll do away with the STACK type altogether.
19155 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
19156 specified in `<certfile>` by updating the entry in the index.txt file.
19157 This way one no longer has to edit the index.txt file manually for
19158 revoking a certificate. The -revoke option does the gory details now.
19160 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
19162 * Fix `openssl crl -noout -text` combination where `-noout` killed the
19163 `-text` option at all and this way the `-noout -text` combination was
19164 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
19166 *Ralf S. Engelschall*
19168 * Make sure a corresponding plain text error message exists for the
19169 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
19170 verify callback function determined that a certificate was revoked.
19172 *Ralf S. Engelschall*
19174 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
19175 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
19176 all available ciphers including rc5, which was forgotten until now.
19177 In order to let the testing shell script know which algorithms
19178 are available, a new (up to now undocumented) command
19179 `openssl list-cipher-commands` is used.
19183 * Bugfix: s_client occasionally would sleep in select() when
19184 it should have checked SSL_pending() first.
19188 * New functions DSA_do_sign and DSA_do_verify to provide access to
19189 the raw DSA values prior to ASN.1 encoding.
19193 * Tweaks to Configure
19195 *Niels Poppe <niels@netbox.org>*
19197 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
19202 * New variables $(RANLIB) and $(PERL) in the Makefiles.
19206 * New config option to avoid instructions that are illegal on the 80386.
19207 The default code is faster, but requires at least a 486.
19211 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
19212 SSL2_SERVER_VERSION (not used at all) macros, which are now the
19213 same as SSL2_VERSION anyway.
19217 * New "-showcerts" option for s_client.
19221 * Still more PKCS#12 integration. Add pkcs12 application to openssl
19222 application. Various cleanups and fixes.
19226 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
19227 modify error routines to work internally. Add error codes and PBE init
19228 to library startup routines.
19232 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
19233 packing functions to asn1 and evp. Changed function names and error
19234 codes along the way.
19238 * PKCS12 integration: and so it begins... First of several patches to
19239 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
19240 objects to objects.h
19244 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
19245 and display support for Thawte strong extranet extension.
19249 * Add LinuxPPC support.
19251 *Jeff Dubrule <igor@pobox.org>*
19253 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
19254 bn_div_words in alpha.s.
19256 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
19258 * Make sure the RSA OAEP test is skipped under -DRSAref because
19259 OAEP isn't supported when OpenSSL is built with RSAref.
19261 *Ulf Moeller <ulf@fitug.de>*
19263 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
19264 so they no longer are missing under -DNOPROTO.
19266 *Soren S. Jorvang <soren@t.dk>*
19268 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
19270 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
19271 doesn't work when the session is reused. Coming soon!
19275 * Fix a security hole, that allows sessions to be reused in the wrong
19276 context thus bypassing client cert protection! All software that uses
19277 client certs and session caches in multiple contexts NEEDS PATCHING to
19278 allow session reuse! A fuller solution is in the works.
19280 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
19282 * Some more source tree cleanups (removed obsolete files
19283 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
19284 permission on "config" script to be executable) and a fix for the INSTALL
19287 *Ulf Moeller <ulf@fitug.de>*
19289 * Remove some legacy and erroneous uses of malloc, free instead of
19292 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
19294 * Make rsa_oaep_test return non-zero on error.
19296 *Ulf Moeller <ulf@fitug.de>*
19298 * Add support for native Solaris shared libraries. Configure
19299 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
19300 if someone would make that last step automatic.
19302 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
19304 * ctx_size was not built with the right compiler during "make links". Fixed.
19308 * Change the meaning of 'ALL' in the cipher list. It now means "everything
19309 except NULL ciphers". This means the default cipher list will no longer
19310 enable NULL ciphers. They need to be specifically enabled e.g. with
19311 the string "DEFAULT:eNULL".
19315 * Fix to RSA private encryption routines: if p < q then it would
19316 occasionally produce an invalid result. This will only happen with
19317 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
19321 * Be less restrictive and allow also `perl util/perlpath.pl
19322 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
19323 because this way one can also use an interpreter named `perl5` (which is
19324 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
19325 installed as `perl`).
19327 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19329 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
19331 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19333 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
19334 advapi32.lib to Win32 build and change the pem test comparison
19335 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
19336 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
19337 and crypto/des/ede_cbcm_enc.c.
19341 * DES quad checksum was broken on big-endian architectures. Fixed.
19345 * Comment out two functions in bio.h that aren't implemented. Fix up the
19346 Win32 test batch file so it (might) work again. The Win32 test batch file
19347 is horrible: I feel ill....
19351 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
19352 in e_os.h. Audit of header files to check ANSI and non ANSI
19353 sections: 10 functions were absent from non ANSI section and not exported
19354 from Windows DLLs. Fixed up libeay.num for new functions.
19358 * Make `openssl version` output lines consistent.
19360 *Ralf S. Engelschall*
19362 * Fix Win32 symbol export lists for BIO functions: Added
19363 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
19364 to ms/libeay{16,32}.def.
19366 *Ralf S. Engelschall*
19368 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
19369 fine under Unix and passes some trivial tests I've now added. But the
19370 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
19371 added to make sure no one expects that this stuff really works in the
19372 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
19373 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
19376 *Ralf S. Engelschall*
19378 * Fix the generation of two part addresses in perl.
19380 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
19382 * Add config entry for Linux on MIPS.
19384 *John Tobey <jtobey@channel1.com>*
19386 * Make links whenever Configure is run, unless we are on Windoze.
19390 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
19391 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19396 * Add a useful kludge to allow package maintainers to specify compiler and
19397 other platforms details on the command line without having to patch the
19398 Configure script every time: One now can use
19399 `perl Configure <id>:<details>`,
19400 i.e. platform ids are allowed to have details appended
19401 to them (separated by colons). This is treated as there would be a static
19402 pre-configured entry in Configure's %table under key `<id>` with value
19403 `<details>` and `perl Configure <id>` is called. So, when you want to
19404 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19405 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19406 now, which overrides the FreeBSD-elf entry on-the-fly.
19408 *Ralf S. Engelschall*
19410 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19414 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19415 on the `perl Configure ...` command line. This way one can compile
19416 OpenSSL libraries with Position Independent Code (PIC) which is needed
19417 for linking it into DSOs.
19419 *Ralf S. Engelschall*
19421 * Remarkably, export ciphers were totally broken and no-one had noticed!
19426 * Cleaned up the LICENSE document: The official contact for any license
19427 questions now is the OpenSSL core team under openssl-core@openssl.org.
19428 And add a paragraph about the dual-license situation to make sure people
19429 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19430 to the OpenSSL toolkit.
19432 *Ralf S. Engelschall*
19434 * General source tree makefile cleanups: Made `making xxx in yyy...`
19435 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19436 Additionally cleaned up the `make links` target: Remove unnecessary
19437 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19438 to speed processing and no longer clutter the display with confusing
19439 stuff. Instead only the actually done links are displayed.
19441 *Ralf S. Engelschall*
19443 * Permit null encryption ciphersuites, used for authentication only. It used
19444 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19445 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19450 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19451 signed attributes when verifying signatures (this would break them),
19452 the detached data encoding was wrong and public keys obtained using
19453 X509_get_pubkey() weren't freed.
19457 * Add text documentation for the BUFFER functions. Also added a work around
19458 to a Win95 console bug. This was triggered by the password read stuff: the
19459 last character typed gets carried over to the next fread(). If you were
19460 generating a new cert request using 'req' for example then the last
19461 character of the passphrase would be CR which would then enter the first
19466 * Added the new 'Includes OpenSSL Cryptography Software' button as
19467 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19468 button and can be used by applications based on OpenSSL to show the
19469 relationship to the OpenSSL project.
19471 *Ralf S. Engelschall*
19473 * Remove confusing variables in function signatures in files
19474 ssl/ssl_lib.c and ssl/ssl.h.
19476 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19478 * Don't install bss_file.c under PREFIX/include/
19480 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19482 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19483 functions that return function pointers and has support for NT specific
19484 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19485 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19486 unsigned to signed types: this was killing the Win32 compile.
19490 * Add new certificate file to stack functions,
19491 SSL_add_dir_cert_subjects_to_stack() and
19492 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19493 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19494 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19495 This means that Apache-SSL and similar packages don't have to mess around
19496 to add as many CAs as they want to the preferred list.
19500 * Experiment with doxygen documentation. Currently only partially applied to
19502 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19503 openssl.doxy as the configuration file.
19507 * Get rid of remaining C++-style comments which strict C compilers hate.
19509 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19511 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19512 compiled in by default: it has problems with large keys.
19516 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19517 DH private keys and/or callback functions which directly correspond to
19518 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19519 is needed for applications which have to configure certificates on a
19520 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19522 For the RSA certificate situation is makes no difference, but
19523 for the DSA certificate situation this fixes the "no shared cipher"
19524 problem where the OpenSSL cipher selection procedure failed because the
19525 temporary keys were not overtaken from the context and the API provided
19526 no way to reconfigure them.
19527 The new functions now let applications reconfigure the stuff and they
19528 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19529 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19530 non-public-API function ssl_cert_instantiate() is used as a helper
19531 function and also to reduce code redundancy inside ssl_rsa.c.
19533 *Ralf S. Engelschall*
19535 * Move s_server -dcert and -dkey options out of the undocumented feature
19536 area because they are useful for the DSA situation and should be
19537 recognized by the users.
19539 *Ralf S. Engelschall*
19541 * Fix the cipher decision scheme for export ciphers: the export bits are
19542 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19543 SSL_EXP_MASK. So, the original variable has to be used instead of the
19544 already masked variable.
19546 *Richard Levitte <levitte@stacken.kth.se>*
19548 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19550 *Richard Levitte <levitte@stacken.kth.se>*
19552 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19553 from `int` to `unsigned int` because it is a length and initialized by
19554 EVP_DigestFinal() which expects an `unsigned int *`.
19556 *Richard Levitte <levitte@stacken.kth.se>*
19558 * Don't hard-code path to Perl interpreter on shebang line of Configure
19559 script. Instead use the usual Shell->Perl transition trick.
19561 *Ralf S. Engelschall*
19563 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19564 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19565 -noout -modulus` as it's already the case for `openssl rsa -noout
19566 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19567 currently the public key is printed (a decision which was already done by
19568 `openssl dsa -modulus` in the past) which serves a similar purpose.
19569 Additionally the NO_RSA no longer completely removes the whole -modulus
19570 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19573 *Ralf S. Engelschall*
19575 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19576 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19578 *Arne Ansper <arne@ats.cyber.ee>*
19580 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19581 to be added. Now both 'req' and 'ca' can use new objects defined in the
19586 * Add cool BIO that does syslog (or event log on NT).
19588 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19590 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19591 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19592 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19593 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19597 * Add preliminary config info for new extension code.
19601 * Make RSA_NO_PADDING really use no padding.
19603 *Ulf Moeller <ulf@fitug.de>*
19605 * Generate errors when private/public key check is done.
19609 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19610 for some CRL extensions and new objects added.
19614 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19615 key usage extension and fuller support for authority key id.
19619 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19620 padding method for RSA, which is recommended for new applications in PKCS
19621 #1 v2.0 (RFC 2437, October 1998).
19622 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19623 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19624 against Bleichbacher's attack on RSA.
19625 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19628 * Updates to the new SSL compression code
19630 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19632 * Fix so that the version number in the master secret, when passed
19633 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19634 (because the server will not accept higher), that the version number
19635 is 0x03,0x01, not 0x03,0x00
19637 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19639 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19640 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19641 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19645 * Support for RAW extensions where an arbitrary extension can be
19646 created by including its DER encoding. See `apps/openssl.cnf` for
19651 * Make sure latest Perl versions don't interpret some generated C array
19652 code as Perl array code in the crypto/err/err_genc.pl script.
19654 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19656 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19657 not many people have the assembler. Various Win32 compilation fixes and
19658 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19659 build instructions.
19663 * Modify configure script 'Configure' to automatically create crypto/date.h
19664 file under Win32 and also build pem.h from pem.org. New script
19665 util/mkfiles.pl to create the MINFO file on environments that can't do a
19666 'make files': perl util/mkfiles.pl >MINFO should work.
19670 * Major rework of DES function declarations, in the pursuit of correctness
19671 and purity. As a result, many evil casts evaporated, and some weirdness,
19672 too. You may find this causes warnings in your code. Zapping your evil
19673 casts will probably fix them. Mostly.
19677 * Fix for a typo in asn1.h. Bug fix to object creation script
19678 obj_dat.pl. It considered a zero in an object definition to mean
19679 "end of object": none of the objects in objects.h have any zeros
19680 so it wasn't spotted.
19682 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19684 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19685 Masking (CBCM). In the absence of test vectors, the best I have been able
19686 to do is check that the decrypt undoes the encrypt, so far. Send me test
19687 vectors if you have them.
19691 * Correct calculation of key length for export ciphers (too much space was
19692 allocated for null ciphers). This has not been tested!
19696 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19697 message is now correct (it understands "crypto" and "ssl" on its
19698 command line). There is also now an "update" option. This will update
19699 the util/ssleay.num and util/libeay.num files with any new functions.
19701 perl util/mkdef.pl crypto ssl update
19702 it will update them.
19706 * Overhauled the Perl interface:
19707 - ported BN stuff to OpenSSL's different BN library
19708 - made the perl/ source tree CVS-aware
19709 - renamed the package from SSLeay to OpenSSL (the files still contain
19710 their history because I've copied them in the repository)
19711 - removed obsolete files (the test scripts will be replaced
19712 by better Test::Harness variants in the future)
19714 *Ralf S. Engelschall*
19716 * First cut for a very conservative source tree cleanup:
19717 1. merge various obsolete readme texts into doc/ssleay.txt
19718 where we collect the old documents and readme texts.
19719 2. remove the first part of files where I'm already sure that we no
19720 longer need them because of three reasons: either they are just temporary
19721 files which were left by Eric or they are preserved original files where
19722 I've verified that the diff is also available in the CVS via "cvs diff
19723 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19724 the crypto/md/ stuff).
19726 *Ralf S. Engelschall*
19728 * More extension code. Incomplete support for subject and issuer alt
19729 name, issuer and authority key id. Change the i2v function parameters
19730 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19731 what that's for :-) Fix to ASN1 macro which messed up
19732 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19736 * Preliminary support for ENUMERATED type. This is largely copied from the
19741 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19743 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19745 * Make sure `make rehash` target really finds the `openssl` program.
19747 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19749 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19750 like to hear about it if this slows down other processors.
19754 * Add CygWin32 platform information to Configure script.
19756 *Alan Batie <batie@aahz.jf.intel.com>*
19758 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19760 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19762 * New program nseq to manipulate netscape certificate sequences
19766 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19771 * Fixes to BN code. Previously the default was to define BN_RECURSION
19772 but the BN code had some problems that would cause failures when
19773 doing certificate verification and some other functions.
19775 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19777 * Add ASN1 and PEM code to support netscape certificate sequences.
19781 * Add ASN1 and PEM code to support netscape certificate sequences.
19785 * Add several PKIX and private extended key usage OIDs.
19789 * Modify the 'ca' program to handle the new extension code. Modify
19790 openssl.cnf for new extension format, add comments.
19794 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19795 and add a sample to openssl.cnf so req -x509 now adds appropriate
19800 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19801 error code, add initial support to X509_print() and x509 application.
19805 * Takes a deep breath and start adding X509 V3 extension support code. Add
19806 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19807 stuff is currently isolated and isn't even compiled yet.
19811 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19812 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19813 Removed the versions check from X509 routines when loading extensions:
19814 this allows certain broken certificates that don't set the version
19815 properly to be processed.
19819 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19820 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19821 can still be regenerated with "make depend".
19825 * Spelling mistake in C version of CAST-128.
19827 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19829 * Changes to the error generation code. The perl script err-code.pl
19830 now reads in the old error codes and retains the old numbers, only
19831 adding new ones if necessary. It also only changes the .err files if new
19832 codes are added. The makefiles have been modified to only insert errors
19833 when needed (to avoid needlessly modifying header files). This is done
19834 by only inserting errors if the .err file is newer than the auto generated
19835 C file. To rebuild all the error codes from scratch (the old behaviour)
19836 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19837 or delete all the .err files.
19841 * CAST-128 was incorrectly implemented for short keys. The C version has
19842 been fixed, but is untested. The assembler versions are also fixed, but
19843 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19844 to regenerate it if needed.
19845 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19846 Hagino <itojun@kame.net>*
19848 * File was opened incorrectly in randfile.c.
19850 *Ulf Möller <ulf@fitug.de>*
19852 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19853 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19854 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19855 al: it's just almost always a UTCTime. Note this patch adds new error
19856 codes so do a "make errors" if there are problems.
19860 * Correct Linux 1 recognition in config.
19862 *Ulf Möller <ulf@fitug.de>*
19864 * Remove pointless MD5 hash when using DSA keys in ca.
19866 *Anonymous <nobody@replay.com>*
19868 * Generate an error if given an empty string as a cert directory. Also
19869 generate an error if handed NULL (previously returned 0 to indicate an
19870 error, but didn't set one).
19872 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19874 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19878 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19879 parameters. This was causing a warning which killed off the Win32 compile.
19883 * Remove C++ style comments from crypto/bn/bn_local.h.
19885 *Neil Costigan <neil.costigan@celocom.com>*
19887 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19888 based on a text string, looking up short and long names and finally
19889 "dot" format. The "dot" format stuff didn't work. Added new function
19890 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19891 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19892 OID is not part of the table.
19896 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19897 X509_LOOKUP_by_alias().
19901 * Sort openssl functions by name.
19905 * Get the `gendsa` command working and add it to the `list` command. Remove
19906 encryption from sample DSA keys (in case anyone is interested the password
19911 * Make *all* `*_free` functions accept a NULL pointer.
19913 *Frans Heymans <fheymans@isaserver.be>*
19915 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19918 *Anonymous <nobody@replay.com>*
19920 * s_server should send the CAfile as acceptable CAs, not its own cert.
19922 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19924 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19926 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19928 * Temp key "for export" tests were wrong in s3_srvr.c.
19930 *Anonymous <nobody@replay.com>*
19932 * Add prototype for temp key callback functions
19933 SSL_CTX_set_tmp_{rsa,dh}_callback().
19937 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19938 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19942 * X509_name_add_entry() freed the wrong thing after an error.
19944 *Arne Ansper <arne@ats.cyber.ee>*
19946 * rsa_eay.c would attempt to free a NULL context.
19948 *Arne Ansper <arne@ats.cyber.ee>*
19950 * BIO_s_socket() had a broken should_retry() on Windoze.
19952 *Arne Ansper <arne@ats.cyber.ee>*
19954 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19956 *Arne Ansper <arne@ats.cyber.ee>*
19958 * Make sure the already existing X509_STORE->depth variable is initialized
19959 in X509_STORE_new(), but document the fact that this variable is still
19960 unused in the certificate verification process.
19962 *Ralf S. Engelschall*
19964 * Fix the various library and `apps/` files to free up pkeys obtained from
19965 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19969 * Fix reference counting in X509_PUBKEY_get(). This makes
19970 demos/maurice/example2.c work, amongst others, probably.
19972 *Steve Henson and Ben Laurie*
19974 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19975 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19976 are no longer created. This way we have a single and consistent command
19977 line interface `openssl <command>`, similar to `cvs <command>`.
19979 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19981 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19982 BIT STRING wrapper always have zero unused bits.
19986 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19990 * Make the top-level INSTALL documentation easier to understand.
19994 * Makefiles updated to exit if an error occurs in a sub-directory
19995 make (including if user presses ^C) [Paul Sutton]
19997 * Make Montgomery context stuff explicit in RSA data structure.
20001 * Fix build order of pem and err to allow for generated pem.h.
20005 * Fix renumbering bug in X509_NAME_delete_entry().
20009 * Enhanced the err-ins.pl script so it makes the error library number
20010 global and can add a library name. This is needed for external ASN1 and
20011 other error libraries.
20015 * Fixed sk_insert which never worked properly.
20019 * Fix ASN1 macros so they can handle indefinite length constructed
20020 EXPLICIT tags. Some non standard certificates use these: they can now
20025 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
20026 into a single doc/ssleay.txt bundle. This way the information is still
20027 preserved but no longer messes up this directory. Now it's new room for
20028 the new set of documentation files.
20030 *Ralf S. Engelschall*
20032 * SETs were incorrectly DER encoded. This was a major pain, because they
20033 shared code with SEQUENCEs, which aren't coded the same. This means that
20034 almost everything to do with SETs or SEQUENCEs has either changed name or
20035 number of arguments.
20037 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
20039 * Fix test data to work with the above.
20043 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
20044 was already fixed by Eric for 0.9.1 it seems.
20046 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
20048 * Autodetect FreeBSD3.
20052 * Fix various bugs in Configure. This affects the following platforms:
20056 unixware-2.0-pentium
20061 * Eliminate generated files from CVS. Reorder tests to regenerate files
20062 before they are needed.
20066 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
20070 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
20072 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
20073 changed SSLeay to OpenSSL in version strings.
20075 *Ralf S. Engelschall*
20077 * Some fixups to the top-level documents.
20081 * Fixed the nasty bug where rsaref.h was not found under compile-time
20082 because the symlink to include/ was missing.
20084 *Ralf S. Engelschall*
20086 * Incorporated the popular no-RSA/DSA-only patches
20087 which allow to compile an RSA-free SSLeay.
20089 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
20091 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
20092 when "ssleay" is still not found.
20094 *Ralf S. Engelschall*
20096 * Added more platforms to Configure: Cray T3E, HPUX 11,
20098 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
20100 * Updated the README file.
20102 *Ralf S. Engelschall*
20104 * Added various .cvsignore files in the CVS repository subdirs
20105 to make a "cvs update" really silent.
20107 *Ralf S. Engelschall*
20109 * Recompiled the error-definition header files and added
20110 missing symbols to the Win32 linker tables.
20112 *Ralf S. Engelschall*
20114 * Cleaned up the top-level documents;
20115 o new files: CHANGES and LICENSE
20116 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
20117 o merged COPYRIGHT into LICENSE
20118 o removed obsolete TODO file
20119 o renamed MICROSOFT to INSTALL.W32
20121 *Ralf S. Engelschall*
20123 * Removed dummy files from the 0.9.1b source tree:
20124 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
20125 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
20126 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
20127 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
20128 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
20130 *Ralf S. Engelschall*
20132 * Added various platform portability fixes.
20136 * The Genesis of the OpenSSL rpject:
20137 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
20138 Young and Tim J. Hudson created while they were working for C2Net until
20141 *The OpenSSL Project*
20143 ### Changes between 0.9.0b and 0.9.1b [not released]
20145 * Updated a few CA certificates under certs/
20149 * Changed some BIGNUM api stuff.
20153 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
20154 DGUX x86, Linux Alpha, etc.
20158 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
20159 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
20164 * Add -strparse option to asn1pars program which parses nested
20167 *Dr Stephen Henson <shenson@bigfoot.com>*
20169 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
20173 * DSA fix for "ca" program.
20177 * Added "-genkey" option to "dsaparam" program.
20181 * Added RIPE MD160 (rmd160) message digest.
20185 * Added -a (all) option to "ssleay version" command.
20189 * Added PLATFORM define which is the id given to Configure.
20193 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
20197 * Extended the ASN.1 parser routines.
20201 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
20205 * Added a BN_CTX to the BN library.
20209 * Fixed the weak key values in DES library
20213 * Changed API in EVP library for cipher aliases.
20217 * Added support for RC2/64bit cipher.
20221 * Converted the lhash library to the crypto/mem.c functions.
20225 * Added more recognized ASN.1 object ids.
20229 * Added more RSA padding checks for SSL/TLS.
20233 * Added BIO proxy/filter functionality.
20237 * Added extra_certs to SSL_CTX which can be used
20238 send extra CA certificates to the client in the CA cert chain sending
20239 process. It can be configured with SSL_CTX_add_extra_chain_cert().
20243 * Now Fortezza is denied in the authentication phase because
20244 this is key exchange mechanism is not supported by SSLeay at all.
20248 * Additional PKCS1 checks.
20252 * Support the string "TLSv1" for all TLS v1 ciphers.
20256 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
20257 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
20261 * Fixed a few memory leaks.
20265 * Fixed various code and comment typos.
20269 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
20270 bytes sent in the client random.
20272 *Edward Bishop <ebishop@spyglass.com>*
20276 [CVE-2023-4807]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-4807
20277 [CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817
20278 [CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446
20279 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
20280 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
20281 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
20282 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
20283 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
20284 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
20285 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
20286 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
20287 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
20288 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
20289 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
20290 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
20291 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
20292 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
20293 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
20294 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
20295 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
20296 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
20297 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
20298 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
20299 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
20300 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
20301 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
20302 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
20303 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
20304 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
20305 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
20306 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
20307 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
20308 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
20309 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
20310 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
20311 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
20312 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
20313 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
20314 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
20315 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
20316 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
20317 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
20318 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
20319 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
20320 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
20321 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
20322 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
20323 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
20324 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
20325 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
20326 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
20327 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
20328 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
20329 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
20330 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
20331 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
20332 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
20333 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
20334 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
20335 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
20336 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
20337 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
20338 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
20339 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
20340 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
20341 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
20342 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
20343 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
20344 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
20345 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
20346 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
20347 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
20348 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
20349 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
20350 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
20351 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
20352 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
20353 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
20354 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
20355 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
20356 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
20357 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
20358 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
20359 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
20360 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
20361 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
20362 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
20363 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
20364 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
20365 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
20366 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
20367 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
20368 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
20369 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
20370 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
20371 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
20372 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
20373 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
20374 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
20375 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
20376 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
20377 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
20378 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
20379 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
20380 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
20381 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
20382 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
20383 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
20384 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
20385 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
20386 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
20387 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
20388 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
20389 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
20390 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20391 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20392 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20393 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20394 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20395 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20396 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20397 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20398 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20399 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20400 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20401 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20402 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20403 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20404 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20405 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20406 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20407 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20408 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20409 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20410 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20411 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20412 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20413 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20414 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20415 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20416 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20417 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20418 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20419 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20420 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20421 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20422 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20423 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20424 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20425 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20426 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20427 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20428 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20429 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20430 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20431 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20432 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20433 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20434 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20435 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20436 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20437 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20438 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20439 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20440 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20441 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20442 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20443 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20444 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20445 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20446 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20447 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20448 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20449 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20450 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20451 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20452 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20453 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20454 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20455 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20456 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20457 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20458 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655