4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.2](#openssl-32)
14 - [OpenSSL 3.1](#openssl-31)
15 - [OpenSSL 3.0](#openssl-30)
16 - [OpenSSL 1.1.1](#openssl-111)
17 - [OpenSSL 1.1.0](#openssl-110)
18 - [OpenSSL 1.0.2](#openssl-102)
19 - [OpenSSL 1.0.1](#openssl-101)
20 - [OpenSSL 1.0.0](#openssl-100)
21 - [OpenSSL 0.9.x](#openssl-09x)
26 ### Changes between 3.1 and 3.2 [xx XXX xxxx]
28 * Added EC_GROUP_to_params which creates an OSSL_PARAM array
29 from a given EC_GROUP.
33 * Added support for pluggable (provider-based) TLS signature algorithms.
34 This enables TLS 1.3 authentication operations with algorithms embedded
35 in providers not included by default in OpenSSL. In combination with
36 the already available pluggable KEM and X.509 support, this enables
37 for example suitable providers to deliver post-quantum or quantum-safe
38 cryptography to OpenSSL users.
42 * Added support for Hybrid Public Key Encryption (HPKE) as defined
43 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
44 Message Layer Security (MLS) and other IETF specifications.
45 HPKE can also be used by other applications that require
46 encrypting "to" an ECDH public key. External APIs are defined in
47 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
51 * Add support for certificate compression (RFC8879), including
52 library support for Brotli and Zstandard compression.
56 * Add the ability to add custom attributes to PKCS12 files. Add a new API
57 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
58 for a user specified callback and optional argument.
59 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
60 added to the existing STACK_OF attrs.
64 * Major refactor of the libssl record layer
68 * Added a new BIO_s_dgram_mem() to read/write datagrams to memory
72 * Add a mac salt length option for the pkcs12 command.
76 * Add more SRTP protection profiles from RFC8723 and RFC8269.
80 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
82 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
84 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
85 supported and enabled.
89 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
90 to the list of ciphersuites providing Perfect Forward Secrecy as
91 required by SECLEVEL >= 3.
93 *Dmitry Belyavskiy, Nicola Tuveri*
95 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
96 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
97 SSL_get0_iana_groups() function-like macro, retrieves the list of
98 supported groups sent by the peer.
99 The function SSL_client_hello_get_extension_order() populates
100 a caller-supplied array with the list of extension types present in the
101 ClientHello, in order of appearance.
105 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
106 to make it possible to use empty passphrase strings.
110 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
111 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
112 of 160 bits and above and less than 224 bits were previously accepted by
113 default but are now no longer allowed. By default TLS compression was
114 already disabled in previous OpenSSL versions. At security level 2 it cannot
119 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
124 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
125 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
126 will need to load the legacy crypto provider.
129 * CCM8 cipher suites in TLS have been downgraded to security level zero
130 because they use a short authentication tag which lowers their strength.
134 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
139 * Add X.509 certificate codeSigning purpose and related checks on key usage and
140 extended key usage of the leaf certificate according to the CA/Browser Forum.
144 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
145 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
146 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
147 X.509 version 3 if the certificate information includes X.509 extensions.
151 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
152 such as adding a trace facility for debugging certificate chain building.
156 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
157 in particular supporting requests for central key generation, generalized
158 polling, and various types of genm/genp exchanges defined in CMP Updates.
162 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
163 like correcting the TLS and proxy support and adding tracing for debugging.
167 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
171 * `CMS_add0_cert()` and `CMS_add1_cert()` no more throw an error
172 if a certificate to be added is already present.
173 * `CMS_sign_ex()` and `CMS_sign()` now ignore any duplicate certificates
174 in their `certs` argument and no longer throw an error for them.
178 * Fixed and extended `util/check-format.pl` for checking adherence to the
179 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
180 The checks are meanwhile more complete and yield fewer false positives.
184 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
185 sending and receiving multiple messages in a single call. An implementation
186 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
190 * The `SSL_CERT_PATH` and `SSL_CERT_URI` environment variables are introduced.
191 `SSL_CERT_URI` can be used to specify a URI for a root certificate store. The
192 `SSL_CERT_PATH` environment variable specifies a delimiter-separated list of
193 paths which are searched for root certificates.
195 The existing `SSL_CERT_DIR` environment variable is deprecated.
196 `SSL_CERT_DIR` was previously used to specify either a delimiter-separated
197 list of paths or an URI, which is ambiguous. Setting `SSL_CERT_PATH` causes
198 `SSL_CERT_DIR` to be ignored for the purposes of determining root certificate
199 directories, and setting `SSL_CERT_URI` causes `SSL_CERT_DIR` to be ignored
200 for the purposes of determining root certificate stores.
204 * Support for loading root certificates from the Windows certificate store
205 has been added. The support is in the form of a store which recognises the
206 URI string of `org.openssl.winstore://`. This store is enabled by default and
207 can be disabled using the new compile-time option `no-winstore`.
211 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
212 kernel versions that support KTLS have a known bug in CCM processing. That
213 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
214 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
219 * Zerocopy KTLS sendfile() support on Linux.
223 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
224 decryption as a protection against Bleichenbacher-like attacks.
225 The RSA decryption API will now return a randomly generated deterministic
226 message instead of an error in case it detects an error when checking
227 padding during PKCS#1 v1.5 decryption. This is a general protection against
228 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
230 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
231 on the RSA decryption context.
235 * Support for Argon2d, Argon2i, Argon2id KDFs has been added along with
236 basic thread pool implementation for select platforms.
243 ### Changes between 3.1.0 and 3.1.1 [xx XXX xxxx]
245 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
246 that it does not enable policy checking. Thanks to David Benjamin for
247 discovering this issue.
252 * Fixed an issue where invalid certificate policies in leaf certificates are
253 silently ignored by OpenSSL and other certificate policy checks are skipped
254 for that certificate. A malicious CA could use this to deliberately assert
255 invalid certificate policies in order to circumvent policy checking on the
256 certificate altogether.
261 * Limited the number of nodes created in a policy tree to mitigate
262 against CVE-2023-0464. The default limit is set to 1000 nodes, which
263 should be sufficient for most installations. If required, the limit
264 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
265 time define to a desired maximum number of nodes or zero to allow
271 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
273 * Add FIPS provider configuration option to enforce the
274 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
275 The option '-ems-check' can optionally be supplied to
276 'openssl fipsinstall'.
280 * The FIPS provider includes a few non-approved algorithms for
281 backward compatibility purposes and the "fips=yes" property query
282 must be used for all algorithm fetches to ensure FIPS compliance.
284 The algorithms that are included but not approved are Triple DES ECB,
285 Triple DES CBC and EdDSA.
289 * Added support for KMAC in KBKDF.
293 * RNDR and RNDRRS support in provider functions to provide
294 random number generation for Arm CPUs (aarch64).
298 * s_client and s_server apps now explicitly say when the TLS version
299 does not include the renegotiation mechanism. This avoids confusion
300 between that scenario versus when the TLS version includes secure
301 renegotiation but the peer lacks support for it.
305 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
307 *Tomasz Kantecki, Andrey Matyukov*
309 * The various OBJ_* functions have been made thread safe.
313 * Parallel dual-prime 1536/2048-bit modular exponentiation for
314 AVX512_IFMA capable processors.
316 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
318 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
319 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
320 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
321 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
322 `OPENSSL_NO_DEPRECATED_3_1`.
324 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
325 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
326 definitions for these functions regardless of whether
327 `OPENSSL_NO_DEPRECATED_3_1` is defined.
329 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
330 functions regardless of whether they are using them. It is recommended that
331 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
335 * When generating safe-prime DH parameters set the recommended private key
336 length equivalent to minimum key lengths as in RFC 7919.
340 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
341 maximum size that is smaller or equal to the digest length to comply with
342 FIPS 186-4 section 5. This is implemented by a new option
343 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
344 `rsa_pss_saltlen` parameter, which is now the default. Signature
345 verification is not affected by this change and continues to work as before.
352 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
353 listed here are only a brief description.
354 The migration guide contains more detailed information related to new features,
355 breaking changes, and mappings for the large list of deprecated functions.
357 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
359 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
361 * Fixed NULL dereference during PKCS7 data verification.
363 A NULL pointer can be dereferenced when signatures are being
364 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
365 algorithm used for the signature is known to the OpenSSL library but
366 the implementation of the hash algorithm is not available the digest
367 initialization will fail. There is a missing check for the return
368 value from the initialization function which later leads to invalid
369 usage of the digest API most likely leading to a crash.
372 PKCS7 data is processed by the SMIME library calls and also by the
373 time stamp (TS) library calls. The TLS implementation in OpenSSL does
374 not call these functions however third party applications would be
375 affected if they call these functions to verify signatures on untrusted
380 * Fixed X.400 address type confusion in X.509 GeneralName.
382 There is a type confusion vulnerability relating to X.400 address processing
383 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
384 but the public structure definition for GENERAL_NAME incorrectly specified
385 the type of the x400Address field as ASN1_TYPE. This field is subsequently
386 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
389 When CRL checking is enabled (i.e. the application sets the
390 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
391 pass arbitrary pointers to a memcmp call, enabling them to read memory
392 contents or enact a denial of service.
397 * Fixed NULL dereference validating DSA public key.
399 An invalid pointer dereference on read can be triggered when an
400 application tries to check a malformed DSA public key by the
401 EVP_PKEY_public_check() function. This will most likely lead
402 to an application crash. This function can be called on public
403 keys supplied from untrusted sources which could allow an attacker
404 to cause a denial of service attack.
406 The TLS implementation in OpenSSL does not call this function
407 but applications might call the function if there are additional
408 security requirements imposed by standards such as FIPS 140-3.
411 *Shane Lontis, Tomáš Mráz*
413 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
415 An invalid pointer dereference on read can be triggered when an
416 application tries to load malformed PKCS7 data with the
417 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
419 The result of the dereference is an application crash which could
420 lead to a denial of service attack. The TLS implementation in OpenSSL
421 does not call this function however third party applications might
422 call these functions on untrusted data.
427 * Fixed Use-after-free following BIO_new_NDEF.
429 The public API function BIO_new_NDEF is a helper function used for
430 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
431 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
432 be called directly by end user applications.
434 The function receives a BIO from the caller, prepends a new BIO_f_asn1
435 filter BIO onto the front of it to form a BIO chain, and then returns
436 the new head of the BIO chain to the caller. Under certain conditions,
437 for example if a CMS recipient public key is invalid, the new filter BIO
438 is freed and the function returns a NULL result indicating a failure.
439 However, in this case, the BIO chain is not properly cleaned up and the
440 BIO passed by the caller still retains internal pointers to the previously
441 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
442 then a use-after-free will occur. This will most likely result in a crash.
445 *Viktor Dukhovni, Matt Caswell*
447 * Fixed Double free after calling PEM_read_bio_ex.
449 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
450 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
451 data. If the function succeeds then the "name_out", "header" and "data"
452 arguments are populated with pointers to buffers containing the relevant
453 decoded data. The caller is responsible for freeing those buffers. It is
454 possible to construct a PEM file that results in 0 bytes of payload data.
455 In this case PEM_read_bio_ex() will return a failure code but will populate
456 the header argument with a pointer to a buffer that has already been freed.
457 If the caller also frees this buffer then a double free will occur. This
458 will most likely lead to a crash.
460 The functions PEM_read_bio() and PEM_read() are simple wrappers around
461 PEM_read_bio_ex() and therefore these functions are also directly affected.
463 These functions are also called indirectly by a number of other OpenSSL
464 functions including PEM_X509_INFO_read_bio_ex() and
465 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
466 internal uses of these functions are not vulnerable because the caller does
467 not free the header argument if PEM_read_bio_ex() returns a failure code.
470 *Kurt Roeckx, Matt Caswell*
472 * Fixed Timing Oracle in RSA Decryption.
474 A timing based side channel exists in the OpenSSL RSA Decryption
475 implementation which could be sufficient to recover a plaintext across
476 a network in a Bleichenbacher style attack. To achieve a successful
477 decryption an attacker would have to be able to send a very large number
478 of trial messages for decryption. The vulnerability affects all RSA padding
479 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
482 *Dmitry Belyavsky, Hubert Kario*
484 * Fixed X.509 Name Constraints Read Buffer Overflow.
486 A read buffer overrun can be triggered in X.509 certificate verification,
487 specifically in name constraint checking. The read buffer overrun might
488 result in a crash which could lead to a denial of service attack.
489 In a TLS client, this can be triggered by connecting to a malicious
490 server. In a TLS server, this can be triggered if the server requests
491 client authentication and a malicious client connects.
496 * Fixed X.509 Policy Constraints Double Locking security issue.
498 If an X.509 certificate contains a malformed policy constraint and
499 policy processing is enabled, then a write lock will be taken twice
500 recursively. On some operating systems (most widely: Windows) this
501 results in a denial of service when the affected process hangs. Policy
502 processing being enabled on a publicly facing server is not considered
503 to be a common setup.
508 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
509 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
510 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
511 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
512 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
513 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
514 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
515 for legacy EC and SM2 keys is also changed similarly to honor the
516 equivalent conversion format flag as specified in the underlying
517 `EC_KEY` object being exported to a provider, when this function is
518 called through `EVP_PKEY_export()`.
522 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
524 * Fixed two buffer overflows in punycode decoding functions.
526 A buffer overrun can be triggered in X.509 certificate verification,
527 specifically in name constraint checking. Note that this occurs after
528 certificate chain signature verification and requires either a CA to
529 have signed the malicious certificate or for the application to continue
530 certificate verification despite failure to construct a path to a trusted
533 In a TLS client, this can be triggered by connecting to a malicious
534 server. In a TLS server, this can be triggered if the server requests
535 client authentication and a malicious client connects.
537 An attacker can craft a malicious email address to overflow
538 an arbitrary number of bytes containing the `.` character (decimal 46)
539 on the stack. This buffer overflow could result in a crash (causing a
543 An attacker can craft a malicious email address to overflow four
544 attacker-controlled bytes on the stack. This buffer overflow could
545 result in a crash (causing a denial of service) or potentially remote code
546 execution depending on stack layout for any given platform/compiler.
551 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
552 parameters in OpenSSL code.
553 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
554 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
555 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
556 Using these invalid names may cause algorithms to use slower methods
557 that ignore the CRT parameters.
561 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
566 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
567 data to be signed before signing the certificate.
571 * Added RIPEMD160 to the default provider.
575 * Ensured that the key share group sent or accepted for the key exchange
576 is allowed for the protocol version.
580 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
582 * OpenSSL supports creating a custom cipher via the legacy
583 EVP_CIPHER_meth_new() function and associated function calls. This function
584 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
585 to use the new provider mechanism in order to implement custom ciphers.
587 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
588 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
589 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
590 and decryption initialisation functions). Instead of using the custom cipher
591 directly it incorrectly tries to fetch an equivalent cipher from the
592 available providers. An equivalent cipher is found based on the NID passed to
593 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
594 given cipher. However it is possible for an application to incorrectly pass
595 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
596 is used in this way the OpenSSL encryption/decryption initialisation function
597 will match the NULL cipher as being equivalent and will fetch this from the
598 available providers. This will succeed if the default provider has been
599 loaded (or if a third party provider has been loaded that offers this
600 cipher). Using the NULL cipher means that the plaintext is emitted as the
603 Applications are only affected by this issue if they call
604 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
605 encryption/decryption initialisation function. Applications that only use
606 SSL/TLS are not impacted by this issue.
611 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
616 * Fixed the linux-mips64 Configure target which was missing the
617 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
622 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
627 * Correctly handle a retransmitted ClientHello in DTLS
631 * Fixed detection of ktls support in cross-compile environment on Linux
635 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
640 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
641 report correct results in some cases
645 * Fix UWP builds by defining VirtualLock
649 * For known safe primes use the minimum key length according to RFC 7919.
650 Longer private key sizes unnecessarily raise the cycles needed to compute the
651 shared secret without any increase of the real security. This fixes a
652 regression from 1.1.1 where these shorter keys were generated for the known
657 * Added the loongarch64 target
661 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
662 only passed to the FIPS provider and not to the default or legacy provider.
666 * Fixed reported performance degradation on aarch64. Restored the
667 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
668 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
669 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
670 The new algorithm is still used for 32 bit targets.
674 * Added a missing header for memcmp that caused compilation failure on some
679 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
681 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
682 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
683 This issue makes the RSA implementation with 2048 bit private keys
684 incorrect on such machines and memory corruption will happen during
685 the computation. As a consequence of the memory corruption an attacker
686 may be able to trigger a remote code execution on the machine performing
689 SSL/TLS servers or other servers using 2048 bit RSA private keys running
690 on machines supporting AVX512IFMA instructions of the X86_64 architecture
691 are affected by this issue.
696 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
697 implementation would not encrypt the entirety of the data under some
698 circumstances. This could reveal sixteen bytes of data that was
699 preexisting in the memory that wasn't written. In the special case of
700 "in place" encryption, sixteen bytes of the plaintext would be revealed.
702 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
703 they are both unaffected.
706 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
708 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
710 * In addition to the c_rehash shell command injection identified in
711 CVE-2022-1292, further bugs where the c_rehash script does not
712 properly sanitise shell metacharacters to prevent command injection have been
715 When the CVE-2022-1292 was fixed it was not discovered that there
716 are other places in the script where the file names of certificates
717 being hashed were possibly passed to a command executed through the shell.
719 This script is distributed by some operating systems in a manner where
720 it is automatically executed. On such operating systems, an attacker
721 could execute arbitrary commands with the privileges of the script.
723 Use of the c_rehash script is considered obsolete and should be replaced
724 by the OpenSSL rehash command line tool.
727 *Daniel Fiala, Tomáš Mráz*
729 * Case insensitive string comparison no longer uses locales. It has instead
730 been directly implemented.
734 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
736 * Case insensitive string comparison is reimplemented via new locale-agnostic
737 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
738 comparison. The previous implementation had problems when the Turkish locale
743 * Fixed a bug in the c_rehash script which was not properly sanitising shell
744 metacharacters to prevent command injection. This script is distributed by
745 some operating systems in a manner where it is automatically executed. On
746 such operating systems, an attacker could execute arbitrary commands with the
747 privileges of the script.
749 Use of the c_rehash script is considered obsolete and should be replaced
750 by the OpenSSL rehash command line tool.
755 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
756 certificate on an OCSP response. The bug caused the function in the case
757 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
758 response (meaning a successful verification) even in the case where the
759 response signing certificate fails to verify.
761 It is anticipated that most users of `OCSP_basic_verify` will not use the
762 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
763 a negative value (indicating a fatal error) in the case of a certificate
764 verification failure. The normal expected return value in this case would be
767 This issue also impacts the command line OpenSSL "ocsp" application. When
768 verifying an ocsp response with the "-no_cert_checks" option the command line
769 application will report that the verification is successful even though it
770 has in fact failed. In this case the incorrect successful response will also
771 be accompanied by error messages showing the failure and contradicting the
772 apparently successful result.
777 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
778 AAD data as the MAC key. This made the MAC key trivially predictable.
780 An attacker could exploit this issue by performing a man-in-the-middle attack
781 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
782 that the modified data would still pass the MAC integrity check.
784 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
785 endpoint will always be rejected by the recipient and the connection will
786 fail at that point. Many application protocols require data to be sent from
787 the client to the server first. Therefore, in such a case, only an OpenSSL
788 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
790 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
791 sent in both directions. In this case both clients and servers could be
792 affected, regardless of the application protocol.
794 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
795 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
796 the handshake when using this ciphersuite.
798 The confidentiality of data is not impacted by this issue, i.e. an attacker
799 cannot decrypt data that has been encrypted using this ciphersuite - they can
802 In order for this attack to work both endpoints must legitimately negotiate
803 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
804 OpenSSL 3.0, and is not available within the default provider or the default
805 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
806 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
807 following must have occurred:
809 1) OpenSSL must have been compiled with the (non-default) compile time option
810 enable-weak-ssl-ciphers
812 2) OpenSSL must have had the legacy provider explicitly loaded (either
813 through application code or via configuration)
815 3) The ciphersuite must have been explicitly added to the ciphersuite list
817 4) The libssl security level must have been set to 0 (default is 1)
819 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
821 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
822 others that both endpoints have in common
827 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
828 occupied by the removed hash table entries.
830 This function is used when decoding certificates or keys. If a long lived
831 process periodically decodes certificates or keys its memory usage will
832 expand without bounds and the process might be terminated by the operating
833 system causing a denial of service. Also traversing the empty hash table
834 entries will take increasingly more time.
836 Typically such long lived processes might be TLS clients or TLS servers
837 configured to accept client certificate authentication.
840 *Hugo Landau, Aliaksei Levin*
842 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
843 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
844 statistics are no longer supported. For compatibility, these statistics are
845 still listed in the output but are now always reported as zero.
849 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
851 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
852 for non-prime moduli.
854 Internally this function is used when parsing certificates that contain
855 elliptic curve public keys in compressed form or explicit elliptic curve
856 parameters with a base point encoded in compressed form.
858 It is possible to trigger the infinite loop by crafting a certificate that
859 has invalid explicit curve parameters.
861 Since certificate parsing happens prior to verification of the certificate
862 signature, any process that parses an externally supplied certificate may thus
863 be subject to a denial of service attack. The infinite loop can also be
864 reached when parsing crafted private keys as they can contain explicit
865 elliptic curve parameters.
867 Thus vulnerable situations include:
869 - TLS clients consuming server certificates
870 - TLS servers consuming client certificates
871 - Hosting providers taking certificates or private keys from customers
872 - Certificate authorities parsing certification requests from subscribers
873 - Anything else which parses ASN.1 elliptic curve parameters
875 Also any other applications that use the BN_mod_sqrt() where the attacker
876 can control the parameter values are vulnerable to this DoS issue.
881 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
882 to the list of ciphersuites providing Perfect Forward Secrecy as
883 required by SECLEVEL >= 3.
885 *Dmitry Belyavskiy, Nicola Tuveri*
887 * Made the AES constant time code for no-asm configurations
888 optional due to the resulting 95% performance degradation.
889 The AES constant time code can be enabled, for no assembly
890 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
894 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
899 * The negative return value handling of the certificate verification callback
900 was reverted. The replacement is to set the verification retry state with
901 the SSL_set_retry_verify() function.
905 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
907 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
908 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
909 verify a certificate supplied by a server. That function may return a
910 negative return value to indicate an internal error (for example out of
911 memory). Such a negative return value is mishandled by OpenSSL and will cause
912 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
913 success and a subsequent call to SSL_get_error() to return the value
914 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
915 returned by OpenSSL if the application has previously called
916 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
917 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
918 totally unexpected and applications may not behave correctly as a result. The
919 exact behaviour will depend on the application but it could result in
920 crashes, infinite loops or other similar incorrect responses.
922 This issue is made more serious in combination with a separate bug in OpenSSL
923 3.0 that will cause X509_verify_cert() to indicate an internal error when
924 processing a certificate chain. This will occur where a certificate does not
925 include the Subject Alternative Name extension but where a Certificate
926 Authority has enforced name constraints. This issue can occur even with valid
932 * Corrected a few file name and file reference bugs in the build,
933 installation and setup scripts, which lead to installation verification
934 failures. Slightly enhanced the installation verification script.
938 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
943 * Fixed PVK encoder to properly query for the passphrase.
947 * Multiple fixes in the OSSL_HTTP API functions.
951 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
952 OSSL_PARAM_INTEGER data type and return error on negative numbers
953 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
954 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
958 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
962 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
966 * Multiple threading fixes.
970 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
974 * Allow fetching an operation from the provider that owns an unexportable key
975 as a fallback if that is still allowed by the property query.
979 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
981 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
986 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
987 S390X capability vector to zero. This simplifies testing of different code
988 paths on S390X architecture.
992 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
993 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
994 SP 800-38D". The communication will fail at this point.
998 * The EC_GROUP_clear_free() function is deprecated as there is nothing
999 confidential in EC_GROUP data.
1003 * The byte order mark (BOM) character is ignored if encountered at the
1004 beginning of a PEM-formatted file.
1008 * Added CMS support for the Russian GOST algorithms.
1012 * Due to move of the implementation of cryptographic operations
1013 to the providers, validation of various operation parameters can
1014 be postponed until the actual operation is executed where previously
1015 it happened immediately when an operation parameter was set.
1017 For example when setting an unsupported curve with
1018 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
1019 fail but later keygen operations with the EVP_PKEY_CTX will fail.
1021 *OpenSSL team members and many third party contributors*
1023 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
1024 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
1025 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1026 instead to retrieve these algorithms from a provider.
1030 * On build targets where the multilib postfix is set in the build
1031 configuration the libdir directory was changing based on whether
1032 the lib directory with the multilib postfix exists on the system
1033 or not. This unpredictable behavior was removed and eventual
1034 multilib postfix is now always added to the default libdir. Use
1035 `--libdir=lib` to override the libdir if adding the postfix is
1040 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1041 no longer interoperable with OpenSSL 1.1.1.
1045 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1046 function codes, this function can only cause problems for calling
1051 * Add a configurable flag to output date formats as ISO 8601. Does not
1052 change the default date format.
1056 * Version of MSVC earlier than 1300 could get link warnings, which could
1057 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1058 Support for this flag has been removed.
1062 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1063 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1064 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1065 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1066 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1070 * The signatures of the functions to get and set options on SSL and
1071 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1072 Some source code changes may be required.
1076 * The public definitions of conf_method_st and conf_st have been
1077 deprecated. They will be made opaque in a future release.
1079 *Rich Salz and Tomáš Mráz*
1081 * Client-initiated renegotiation is disabled by default. To allow it, use
1082 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1083 flag, or the "ClientRenegotiation" config parameter as appropriate.
1087 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1088 or modify relative pathname inclusion.
1092 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1093 validated. Please consult the README-FIPS and
1094 README-PROVIDERS files, as well as the migration guide.
1096 *OpenSSL team members and many third party contributors*
1098 * For the key types DH and DHX the allowed settable parameters are now different.
1102 * The openssl commands that read keys, certificates, and CRLs now
1103 automatically detect the PEM or DER format of the input files.
1105 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1107 * Added enhanced PKCS#12 APIs which accept a library context.
1111 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1115 * Added support for Kernel TLS (KTLS).
1117 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1119 * Support for RFC 5746 secure renegotiation is now required by default for
1120 SSL or TLS connections to succeed.
1124 * The signature of the `copy` functional parameter of the
1125 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1126 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1127 the signature of the `pub_decode` functional parameter of the
1128 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1129 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1133 * The error return values from some control calls (ctrl) have changed.
1137 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1141 * Many functions in the EVP_ namespace that are getters of values from
1142 implementations or contexts were renamed to include get or get0 in their
1143 names. Old names are provided as macro aliases for compatibility and
1148 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1149 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1150 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1155 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1158 * The output from the command line applications may have minor
1163 * The output from numerous "printing" may have minor changes.
1167 * Windows thread synchronization uses read/write primitives (SRWLock) when
1168 supported by the OS, otherwise CriticalSection continues to be used.
1172 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1173 work on read only BIO source/sinks that do not support these functions.
1174 This allows piping or redirection of a file BIO using stdin to be buffered
1175 into memory. This is used internally in OSSL_DECODER_from_bio().
1179 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1180 this function would return one of the values OSSL_STORE_INFO_NAME,
1181 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1182 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1183 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1184 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1185 using this function should be amended to handle the changed return value.
1189 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1190 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1191 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1192 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1193 contains more than one certificate identifier: This means that all
1194 certificates referenced there MUST be part of the validation chain.
1198 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1199 RC5, DESX and DES have been moved to the legacy provider.
1203 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1204 RIPEMD-160 have been moved to the legacy provider.
1208 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1213 * The deprecated functions EVP_PKEY_get0_RSA(),
1214 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1215 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1216 well as the similarly named "get1" functions behave differently in
1221 * A number of functions handling low-level keys or engines were deprecated
1222 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1223 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1224 EVP_PKEY_get0_siphash().
1228 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1229 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1230 will need to load the legacy crypto provider. This includes these PBE
1231 algorithms which use this KDF:
1232 - NID_pbeWithMD2AndDES_CBC
1233 - NID_pbeWithMD5AndDES_CBC
1234 - NID_pbeWithSHA1AndRC2_CBC
1235 - NID_pbeWithMD2AndRC2_CBC
1236 - NID_pbeWithMD5AndRC2_CBC
1237 - NID_pbeWithSHA1AndDES_CBC
1241 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1242 BIO_debug_callback() functions.
1246 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1247 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1251 * The RAND_METHOD APIs have been deprecated.
1255 * The SRP APIs have been deprecated.
1259 * Add a compile time option to prevent the caching of provider fetched
1260 algorithms. This is enabled by including the no-cached-fetch option
1261 at configuration time.
1265 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1266 count of PKCS12_DEFAULT_ITER.
1268 *Tomáš Mráz and Sahana Prasad*
1270 * The openssl speed command does not use low-level API calls anymore.
1274 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1277 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1279 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1283 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1284 providers may supply their own group implementations (using either the "key
1285 exchange" or the "key encapsulation" methods) which will automatically be
1286 detected and used by libssl.
1288 *Matt Caswell, Nicola Tuveri*
1290 * The undocumented function X509_certificate_type() has been deprecated;
1294 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1298 * Removed RSA padding mode for SSLv23 (which was only used for
1299 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1300 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1305 * Deprecated the obsolete X9.31 RSA key generation related functions.
1307 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1308 is not allowed to return a value > 1, this is no more taken as failure.
1310 *Viktor Dukhovni and David von Oheimb*
1312 * Deprecated the obsolete X9.31 RSA key generation related functions
1313 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1314 BN_X931_generate_prime_ex().
1318 * The default key generation method for the regular 2-prime RSA keys was
1319 changed to the FIPS 186-4 B.3.6 method.
1323 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1327 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1331 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1332 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1334 *Rich Salz, Richard Levitte, and David von Oheimb*
1336 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1340 * Deprecated `OCSP_parse_url()`.
1344 * Validation of SM2 keys has been separated from the validation of regular EC
1349 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1350 switches: a validation failure triggers an early exit, returning a failure
1351 exit status to the parent process.
1355 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1356 to ignore unknown ciphers.
1360 * The `-cipher-commands` and `-digest-commands` options
1361 of the command line utility `list` have been deprecated.
1362 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1366 * Added convenience functions for generating asymmetric key pairs:
1367 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1368 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1372 * All of the low-level EC_KEY functions have been deprecated.
1374 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1376 * Deprecated all the libcrypto and libssl error string loading
1381 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1382 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1387 * The `-crypt` option to the `passwd` command line tool has been removed.
1391 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1396 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1400 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1401 EVP_PKEY_get1_tls_encodedpoint().
1405 * The security callback, which can be customised by application code, supports
1406 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1407 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1411 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1412 interface. Their functionality remains unchanged.
1416 * Added new option for 'openssl list', '-providers', which will display the
1417 list of loaded providers, their names, version and status. It optionally
1418 displays their gettable parameters.
1422 * Removed EVP_PKEY_set_alias_type().
1426 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1427 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1431 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1432 parameters are still checked for type safety at compile time via helper
1437 * Remove the RAND_DRBG API
1439 *Paul Dale and Matthias St. Pierre*
1441 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1442 as well as actual hostnames.
1446 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1447 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1448 conversely, silently ignore DTLS protocol version bounds when configuring
1449 TLS-based contexts. The commands can be repeated to set bounds of both
1450 types. The same applies with the corresponding "min_protocol" and
1451 "max_protocol" command-line switches, in case some application uses both TLS
1454 SSL_CTX instances that are created for a fixed protocol version (e.g.
1455 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1456 attempts to apply bounds to these protocol versions would result in an
1457 error. Now only the "version-flexible" SSL_CTX instances are subject to
1458 limits in configuration files in command-line options.
1462 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1467 * Reworked the recorded ERR codes to make better space for system errors.
1468 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1469 given code is a system error (true) or an OpenSSL error (false).
1473 * Reworked the test perl framework to better allow parallel testing.
1475 *Nicola Tuveri and David von Oheimb*
1477 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1478 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1482 * 'Configure' has been changed to figure out the configuration target if
1483 none is given on the command line. Consequently, the 'config' script is
1484 now only a mere wrapper. All documentation is changed to only mention
1487 *Rich Salz and Richard Levitte*
1489 * Added a library context `OSSL_LIB_CTX` that applications as well as
1490 other libraries can use to form a separate context within which
1491 libcrypto operations are performed.
1495 * Added various `_ex` functions to the OpenSSL API that support using
1496 a non-default `OSSL_LIB_CTX`.
1500 * Handshake now fails if Extended Master Secret extension is dropped
1505 * Dropped interactive mode from the `openssl` program.
1509 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1511 *David von Oheimb and Shane Lontis*
1513 * Deprecated `EC_METHOD_get_field_type()`.
1517 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1518 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1519 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1523 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1527 * Add CAdES-BES signature verification support, mostly derived
1528 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1530 *Filipe Raimundo da Silva*
1532 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1536 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1537 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1541 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1545 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1546 EC_KEY_precompute_mult().
1550 * Deprecated EC_POINTs_mul().
1554 * Removed FIPS_mode() and FIPS_mode_set().
1558 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1562 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1563 EC_POINT_get_Jprojective_coordinates_GFp().
1567 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1568 arrays to be more easily constructed via a series of utility functions.
1569 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1570 the various push functions and finally convert to a passable OSSL_PARAM
1571 array using OSSL_PARAM_BLD_to_param().
1575 * The security strength of SHA1 and MD5 based signatures in TLS has been
1580 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1581 contain a provider side internal key.
1585 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1589 * Project text documents not yet having a proper file name extension
1590 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1591 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1595 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1596 have been converted to Markdown with the goal to produce documents
1597 which not only look pretty when viewed online in the browser, but
1598 remain well readable inside a plain text editor.
1600 To achieve this goal, a 'minimalistic' Markdown style has been applied
1601 which avoids formatting elements that interfere too much with the
1602 reading flow in the text file. For example, it
1604 * avoids [ATX headings][] and uses [setext headings][] instead
1605 (which works for `<h1>` and `<h2>` headings only).
1606 * avoids [inline links][] and uses [reference links][] instead.
1607 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1609 [ATX headings]: https://github.github.com/gfm/#atx-headings
1610 [setext headings]: https://github.github.com/gfm/#setext-headings
1611 [inline links]: https://github.github.com/gfm/#inline-link
1612 [reference links]: https://github.github.com/gfm/#reference-link
1613 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1614 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1616 *Matthias St. Pierre*
1618 * The test suite is changed to preserve results of each test recipe.
1619 A new directory test-runs/ with subdirectories named like the
1620 test recipes are created in the build tree for this purpose.
1624 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1625 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1626 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1628 *David von Oheimb, Martin Peylo*
1630 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1631 It supports arbitrary request and response content types, GET redirection,
1632 TLS, connections via HTTP(S) proxies, connections and exchange via
1633 user-defined BIOs (allowing implicit connections), persistent connections,
1634 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1635 The legacy OCSP-focused (and only partly documented) API
1636 is retained for backward compatibility, while most of it is deprecated.
1640 * Added `util/check-format.pl`, a tool for checking adherence to the
1641 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1642 The checks performed are incomplete and yield some false positives.
1643 Still the tool should be useful for detecting most typical glitches.
1647 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1648 If domain name resolution yields multiple IP addresses all of them are tried
1649 after `connect()` failures.
1653 * All of the low-level RSA functions have been deprecated.
1657 * X509 certificates signed using SHA1 are no longer allowed at security
1662 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1663 modified to use PKEY APIs. These commands are now in maintenance mode
1664 and no new features will be added to them.
1668 * The command line utility rsautl has been deprecated.
1672 * The command line utilities genrsa and rsa have been modified to use PKEY
1673 APIs. They now write PKCS#8 keys by default. These commands are now in
1674 maintenance mode and no new features will be added to them.
1678 * All of the low-level DH functions have been deprecated.
1680 *Paul Dale and Matt Caswell*
1682 * All of the low-level DSA functions have been deprecated.
1686 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1687 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1691 * Deprecated low-level ECDH and ECDSA functions.
1695 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1699 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1700 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1701 a new formulation to include all the things it can be used for,
1702 as well as words of caution.
1706 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1710 * All of the low-level HMAC functions have been deprecated.
1712 *Paul Dale and David von Oheimb*
1714 * Over two thousand fixes were made to the documentation, including:
1715 - Common options (such as -rand/-writerand, TLS version control, etc)
1716 were refactored and point to newly-enhanced descriptions in openssl.pod.
1717 - Added style conformance for all options (with help from Richard Levitte),
1718 documented all reported missing options, added a CI build to check
1719 that all options are documented and that no unimplemented options
1721 - Documented some internals, such as all use of environment variables.
1722 - Addressed all internal broken L<> references.
1726 * All of the low-level CMAC functions have been deprecated.
1730 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1731 functions have been deprecated.
1733 *Paul Dale and David von Oheimb*
1735 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1736 set of functions. The documentation mentioned negative values for some
1737 errors, but this was never the case, so the mention of negative values
1740 Code that followed the documentation and thereby check with something
1741 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1745 * All of the low-level cipher functions have been deprecated.
1747 *Matt Caswell and Paul Dale*
1749 * Removed include/openssl/opensslconf.h.in and replaced it with
1750 include/openssl/configuration.h.in, which differs in not including
1751 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1752 was added to include both.
1754 This allows internal hacks where one might need to modify the set
1755 of configured macros, for example this if deprecated symbols are
1756 still supposed to be available internally:
1758 #include <openssl/configuration.h>
1760 #undef OPENSSL_NO_DEPRECATED
1761 #define OPENSSL_SUPPRESS_DEPRECATED
1763 #include <openssl/macros.h>
1765 This should not be used by applications that use the exported
1766 symbols, as that will lead to linking errors.
1770 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1771 used in exponentiation with 512-bit moduli. No EC algorithms are
1772 affected. Analysis suggests that attacks against 2-prime RSA1024,
1773 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1774 difficult to perform and are not believed likely. Attacks against DH512
1775 are considered just feasible. However, for an attack the target would
1776 have to re-use the DH512 private key, which is not recommended anyway.
1777 Also applications directly using the low-level API BN_mod_exp may be
1778 affected if they use BN_FLG_CONSTTIME.
1783 * Most memory-debug features have been deprecated, and the functionality
1784 replaced with no-ops.
1788 * Added documentation for the STACK API.
1792 * Introduced a new method type and API, OSSL_ENCODER, to represent
1793 generic encoders. These do the same sort of job that PEM writers
1794 and d2i functions do, but with support for methods supplied by
1795 providers, and the possibility for providers to support other
1800 * Introduced a new method type and API, OSSL_DECODER, to represent
1801 generic decoders. These do the same sort of job that PEM readers
1802 and i2d functions do, but with support for methods supplied by
1803 providers, and the possibility for providers to support other
1808 * Added a .pragma directive to the syntax of configuration files, to
1809 allow varying behavior in a supported and predictable manner.
1810 Currently added pragma:
1814 This allows dollar signs to be a keyword character unless it's
1815 followed by a opening brace or parenthesis. This is useful for
1816 platforms where dollar signs are commonly used in names, such as
1817 volume names and system directory names on VMS.
1821 * Added functionality to create an EVP_PKEY from user data.
1825 * Change the interpretation of the '--api' configuration option to
1826 mean that this is a desired API compatibility level with no
1827 further meaning. The previous interpretation, that this would
1828 also mean to remove all deprecated symbols up to and including
1829 the given version, no requires that 'no-deprecated' is also used
1830 in the configuration.
1832 When building applications, the desired API compatibility level
1833 can be set with the OPENSSL_API_COMPAT macro like before. For
1834 API compatibility version below 3.0, the old style numerical
1835 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1836 For version 3.0 and on, the value is expected to be the decimal
1837 value calculated from the major and minor version like this:
1839 MAJOR * 10000 + MINOR * 100
1843 -DOPENSSL_API_COMPAT=30000 For 3.0
1844 -DOPENSSL_API_COMPAT=30200 For 3.2
1846 To hide declarations that are deprecated up to and including the
1847 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1848 given when building the application as well.
1852 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1853 access to certificate and CRL stores via URIs and OSSL_STORE
1856 This adds the following functions:
1858 - X509_LOOKUP_store()
1859 - X509_STORE_load_file()
1860 - X509_STORE_load_path()
1861 - X509_STORE_load_store()
1862 - SSL_add_store_cert_subjects_to_stack()
1863 - SSL_CTX_set_default_verify_store()
1864 - SSL_CTX_load_verify_file()
1865 - SSL_CTX_load_verify_dir()
1866 - SSL_CTX_load_verify_store()
1870 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1871 The presence of this system service is determined at run-time.
1875 * Added functionality to create an EVP_PKEY context based on data
1876 for methods from providers. This takes an algorithm name and a
1877 property query string and simply stores them, with the intent
1878 that any operation that uses this context will use those strings
1879 to fetch the needed methods implicitly, thereby making the port
1880 of application written for pre-3.0 OpenSSL easier.
1884 * The undocumented function NCONF_WIN32() has been deprecated; for
1885 conversion details see the HISTORY section of doc/man5/config.pod
1889 * Introduced the new functions EVP_DigestSignInit_ex() and
1890 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1891 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1892 pages for further details.
1896 * Over two thousand fixes were made to the documentation, including:
1897 adding missing command flags, better style conformance, documentation
1900 *Rich Salz, Richard Levitte*
1902 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1903 X25519, X448, Ed25519 and Ed448.
1907 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1912 * Deprecated the public definition of `ERR_STATE` as well as the function
1913 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1918 * Added ERR functionality to give callers access to the stored function
1919 names that have replaced the older function code based functions.
1921 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1922 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1923 ERR_peek_error_all() and ERR_peek_last_error_all().
1925 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1926 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1927 ERR_func_error_string().
1931 * Extended testing to be verbose for failing tests only. The make variables
1932 VERBOSE_FAILURE or VF can be used to enable this:
1934 $ make VF=1 test # Unix
1935 $ mms /macro=(VF=1) test ! OpenVMS
1936 $ nmake VF=1 test # Windows
1940 * Added the `-copy_extensions` option to the `x509` command for use with
1941 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1942 all extensions in the request are copied to the certificate or vice versa.
1944 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1946 * Added the `-copy_extensions` option to the `req` command for use with
1947 `-x509`. When given with the `copy` or `copyall` argument,
1948 all extensions in the certification request are copied to the certificate.
1952 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1953 they generate are by default RFC 5280 compliant in the following sense:
1954 There is a subjectKeyIdentifier extension with a hash value of the public key
1955 and for not self-signed certs there is an authorityKeyIdentifier extension
1956 with a keyIdentifier field or issuer information identifying the signing key.
1957 This is done unless some configuration overrides the new default behavior,
1958 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1962 * Added several checks to `X509_verify_cert()` according to requirements in
1963 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1964 (which may be done by using the CLI option `-x509_strict`):
1965 * The basicConstraints of CA certificates must be marked critical.
1966 * CA certificates must explicitly include the keyUsage extension.
1967 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1968 * The issuer name of any certificate must not be empty.
1969 * The subject name of CA certs, certs with keyUsage crlSign,
1970 and certs without subjectAlternativeName must not be empty.
1971 * If a subjectAlternativeName extension is given it must not be empty.
1972 * The signatureAlgorithm field and the cert signature must be consistent.
1973 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1974 must not be marked critical.
1975 * The authorityKeyIdentifier must be given for X.509v3 certs
1976 unless they are self-signed.
1977 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1981 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1982 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1986 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1987 used even when parsing explicit parameters, when loading a encoded key
1988 or calling `EC_GROUP_new_from_ecpkparameters()`/
1989 `EC_GROUP_new_from_ecparameters()`.
1990 This prevents bypass of security hardening and performance gains,
1991 especially for curves with specialized EC_METHODs.
1992 By default, if a key encoded with explicit parameters is loaded and later
1993 encoded, the output is still encoded with explicit parameters, even if
1994 internally a "named" EC_GROUP is used for computation.
1998 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1999 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2000 NULL. After this change, only the cofactor parameter can be NULL. It also
2001 does some minimal sanity checks on the passed order.
2006 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2007 An attack is simple, if the first CMS_recipientInfo is valid but the
2008 second CMS_recipientInfo is chosen ciphertext. If the second
2009 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2010 encryption key will be replaced by garbage, and the message cannot be
2011 decoded, but if the RSA decryption fails, the correct encryption key is
2012 used and the recipient will not notice the attack.
2013 As a work around for this potential attack the length of the decrypted
2014 key must be equal to the cipher default key length, in case the
2015 certificate is not given and all recipientInfo are tried out.
2016 The old behaviour can be re-enabled in the CMS code by setting the
2017 CMS_DEBUG_DECRYPT flag.
2021 * Early start up entropy quality from the DEVRANDOM seed source has been
2022 improved for older Linux systems. The RAND subsystem will wait for
2023 /dev/random to be producing output before seeding from /dev/urandom.
2024 The seeded state is stored for future library initialisations using
2025 a system global shared memory segment. The shared memory identifier
2026 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2027 the desired value. The default identifier is 114.
2031 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2032 when primes for RSA keys are computed.
2033 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2034 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2035 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2036 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2037 This avoids possible fingerprinting of newly generated RSA modules.
2041 * Correct the extended master secret constant on EBCDIC systems. Without this
2042 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2043 negotiate EMS will fail. Unfortunately this also means that TLS connections
2044 between EBCDIC systems with this fix, and EBCDIC systems without this
2045 fix will fail if they negotiate EMS.
2049 * Changed the library initialisation so that the config file is now loaded
2050 by default. This was already the case for libssl. It now occurs for both
2051 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2052 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2056 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2057 where the former acts as a replacement for `ERR_put_error()`, and the
2058 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2059 `ERR_raise_data()` adds more flexibility by taking a format string and
2060 an arbitrary number of arguments following it, to be processed with
2065 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2066 to check if a named provider is loaded and available. When called, it
2067 will also activate all fallback providers if such are still present.
2071 * Enforce a minimum DH modulus size of 512 bits.
2075 * Changed DH parameters to generate the order q subgroup instead of 2q.
2076 Previously generated DH parameters are still accepted by DH_check
2077 but DH_generate_key works around that by clearing bit 0 of the
2078 private key for those. This avoids leaking bit 0 of the private key.
2082 * Significantly reduce secure memory usage by the randomness pools.
2086 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2091 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2092 algorithms. An implementation of a key exchange algorithm can be obtained
2093 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2094 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2095 the older EVP_PKEY_derive_init() function. See the man pages for the new
2096 functions for further details.
2100 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2104 * Removed the function names from error messages and deprecated the
2109 * Removed NextStep support and the macro OPENSSL_UNISTD
2113 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2114 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2115 Also removed "export var as function" capability; we do not export
2116 variables, only functions.
2120 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2121 an error and 1 indicating success. In previous versions of OpenSSL this
2122 was a void type. If a key was set longer than the maximum possible this
2127 * Support SM2 signing and verification schemes with X509 certificate.
2131 * Use SHA256 as the default digest for TS query in the `ts` app.
2135 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2139 * Default cipher lists/suites are now available via a function, the
2140 #defines are deprecated.
2144 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2145 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2146 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2150 * Join the directories crypto/x509 and crypto/x509v3
2154 * Added command 'openssl kdf' that uses the EVP_KDF API.
2158 * Added command 'openssl mac' that uses the EVP_MAC API.
2162 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2163 as default directories. Also added the command 'openssl info'
2164 for scripting purposes.
2168 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2173 * Add prediction resistance to the DRBG reseeding process.
2177 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2178 mandated by IEEE Std 1619-2018.
2182 * Added newline escaping functionality to a filename when using openssl dgst.
2183 This output format is to replicate the output format found in the `*sum`
2184 checksum programs. This aims to preserve backward compatibility.
2186 *Matt Eaton, Richard Levitte, and Paul Dale*
2188 * Removed the heartbeat message in DTLS feature, as it has very
2189 little usage and doesn't seem to fulfill a valuable purpose.
2190 The configuration option is now deprecated.
2194 * Changed the output of 'openssl {digestname} < file' to display the
2195 digest name in its output.
2199 * Added a new generic trace API which provides support for enabling
2200 instrumentation through trace output.
2202 *Richard Levitte & Matthias St. Pierre*
2204 * Added build tests for C++. These are generated files that only do one
2205 thing, to include one public OpenSSL head file each. This tests that
2206 the public header files can be usefully included in a C++ application.
2208 This test isn't enabled by default. It can be enabled with the option
2209 'enable-buildtest-c++'.
2213 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2217 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2221 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2225 * Added KMAC to EVP_MAC.
2229 * Added property based algorithm implementation selection framework to
2234 * Added SCA hardening for modular field inversion in EC_GROUP through
2235 a new dedicated field_inv() pointer in EC_METHOD.
2236 This also addresses a leakage affecting conversions from projective
2237 to affine coordinates.
2239 *Billy Bob Brumley, Nicola Tuveri*
2241 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2242 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2243 those algorithms that were already supported through the EVP_PKEY API
2244 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2245 and scrypt are now wrappers that call EVP_KDF.
2249 * Build devcrypto engine as a dynamic engine.
2251 *Eneas U de Queiroz*
2253 * Add keyed BLAKE2 to EVP_MAC.
2257 * Fix a bug in the computation of the endpoint-pair shared secret used
2258 by DTLS over SCTP. This breaks interoperability with older versions
2259 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2260 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2261 interoperability with such broken implementations. However, enabling
2262 this switch breaks interoperability with correct implementations.
2264 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2265 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2269 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2273 * Changed the license to the Apache License v2.0.
2277 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2279 - Major releases (indicated by incrementing the MAJOR release number)
2280 may introduce incompatible API/ABI changes.
2281 - Minor releases (indicated by incrementing the MINOR release number)
2282 may introduce new features but retain API/ABI compatibility.
2283 - Patch releases (indicated by incrementing the PATCH number)
2284 are intended for bug fixes and other improvements of existing
2285 features only (like improving performance or adding documentation)
2286 and retain API/ABI compatibility.
2290 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2294 * Remove the 'dist' target and add a tarball building script. The
2295 'dist' target has fallen out of use, and it shouldn't be
2296 necessary to configure just to create a source distribution.
2300 * Recreate the OS390-Unix config target. It no longer relies on a
2301 special script like it did for OpenSSL pre-1.1.0.
2305 * Instead of having the source directories listed in Configure, add
2306 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2311 * Add GMAC to EVP_MAC.
2315 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2319 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2320 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2321 to facilitate the continued use of MACs through raw private keys in
2322 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2326 * Deprecate ECDH_KDF_X9_62().
2330 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2331 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2332 are retained for backwards compatibility.
2336 * AES-XTS mode now enforces that its two keys are different to mitigate
2337 the attacked described in "Efficient Instantiations of Tweakable
2338 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2339 Details of this attack can be obtained from:
2340 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2344 * Rename the object files, i.e. give them other names than in previous
2345 versions. Their names now include the name of the final product, as
2346 well as its type mnemonic (bin, lib, shlib).
2350 * Added new option for 'openssl list', '-objects', which will display the
2351 list of built in objects, i.e. OIDs with names.
2355 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2356 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2361 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2362 improves application performance by removing data copies and providing
2363 applications with zero-copy system calls such as sendfile and splice.
2367 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2371 * `PKCS12_parse` now maintains the order of the parsed certificates
2372 when outputting them via `*ca` (rather than reversing it).
2376 * Deprecated pthread fork support methods.
2380 * Added support for FFDHE key exchange in TLS 1.3.
2384 * Added a new concept for OpenSSL plugability: providers. This
2385 functionality is designed to replace the ENGINE API and ENGINE
2386 implementations, and to be much more dynamic, allowing provider
2387 authors to introduce new algorithms among other things, as long as
2388 there's an API that supports the algorithm type.
2390 With this concept comes a new core API for interaction between
2391 libcrypto and provider implementations. Public libcrypto functions
2392 that want to use providers do so through this core API.
2394 The main documentation for this core API is found in
2395 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2396 refer to other manuals describing the API specific for supported
2397 algorithm types (also called operations).
2404 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2406 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2408 * Avoid loading of a dynamic engine twice.
2412 * Prioritise DANE TLSA issuer certs over peer certs
2416 * Fixed random API for MacOS prior to 10.12
2418 These MacOS versions don't support the CommonCrypto APIs
2422 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2424 * Fixed an SM2 Decryption Buffer Overflow.
2426 In order to decrypt SM2 encrypted data an application is expected to
2427 call the API function EVP_PKEY_decrypt(). Typically an application will
2428 call this function twice. The first time, on entry, the "out" parameter
2429 can be NULL and, on exit, the "outlen" parameter is populated with the
2430 buffer size required to hold the decrypted plaintext. The application
2431 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2432 again, but this time passing a non-NULL value for the "out" parameter.
2434 A bug in the implementation of the SM2 decryption code means that the
2435 calculation of the buffer size required to hold the plaintext returned
2436 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2437 size required by the second call. This can lead to a buffer overflow
2438 when EVP_PKEY_decrypt() is called by the application a second time with
2439 a buffer that is too small.
2441 A malicious attacker who is able present SM2 content for decryption to
2442 an application could cause attacker chosen data to overflow the buffer
2443 by up to a maximum of 62 bytes altering the contents of other data held
2444 after the buffer, possibly changing application behaviour or causing
2445 the application to crash. The location of the buffer is application
2446 dependent but is typically heap allocated.
2451 * Fixed various read buffer overruns processing ASN.1 strings
2453 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2454 structure which contains a buffer holding the string data and a field
2455 holding the buffer length. This contrasts with normal C strings which
2456 are represented as a buffer for the string data which is terminated
2457 with a NUL (0) byte.
2459 Although not a strict requirement, ASN.1 strings that are parsed using
2460 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2461 well as any string whose value has been set with the ASN1_STRING_set()
2462 function will additionally NUL terminate the byte array in the
2463 ASN1_STRING structure.
2465 However, it is possible for applications to directly construct valid
2466 ASN1_STRING structures which do not NUL terminate the byte array by
2467 directly setting the "data" and "length" fields in the ASN1_STRING
2468 array. This can also happen by using the ASN1_STRING_set0() function.
2470 Numerous OpenSSL functions that print ASN.1 data have been found to
2471 assume that the ASN1_STRING byte array will be NUL terminated, even
2472 though this is not guaranteed for strings that have been directly
2473 constructed. Where an application requests an ASN.1 structure to be
2474 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2475 been directly constructed by the application without NUL terminating
2476 the "data" field, then a read buffer overrun can occur.
2478 The same thing can also occur during name constraints processing
2479 of certificates (for example if a certificate has been directly
2480 constructed by the application instead of loading it via the OpenSSL
2481 parsing functions, and the certificate contains non NUL terminated
2482 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2483 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2485 If a malicious actor can cause an application to directly construct an
2486 ASN1_STRING and then process it through one of the affected OpenSSL
2487 functions then this issue could be hit. This might result in a crash
2488 (causing a Denial of Service attack). It could also result in the
2489 disclosure of private memory contents (such as private keys, or
2490 sensitive plaintext).
2495 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2497 * Fixed a problem with verifying a certificate chain when using the
2498 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2499 the certificates present in a certificate chain. It is not set by default.
2501 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2502 the chain that have explicitly encoded elliptic curve parameters was added
2503 as an additional strict check.
2505 An error in the implementation of this check meant that the result of a
2506 previous check to confirm that certificates in the chain are valid CA
2507 certificates was overwritten. This effectively bypasses the check
2508 that non-CA certificates must not be able to issue other certificates.
2510 If a "purpose" has been configured then there is a subsequent opportunity
2511 for checks that the certificate is a valid CA. All of the named "purpose"
2512 values implemented in libcrypto perform this check. Therefore, where
2513 a purpose is set the certificate chain will still be rejected even when the
2514 strict flag has been used. A purpose is set by default in libssl client and
2515 server certificate verification routines, but it can be overridden or
2516 removed by an application.
2518 In order to be affected, an application must explicitly set the
2519 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2520 for the certificate verification or, in the case of TLS client or server
2521 applications, override the default purpose.
2526 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2527 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2528 renegotiation ClientHello omits the signature_algorithms extension (where it
2529 was present in the initial ClientHello), but includes a
2530 signature_algorithms_cert extension then a NULL pointer dereference will
2531 result, leading to a crash and a denial of service attack.
2533 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2534 (which is the default configuration). OpenSSL TLS clients are not impacted by
2538 *Peter Kästle and Samuel Sapalski*
2540 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2542 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2543 create a unique hash value based on the issuer and serial number data
2544 contained within an X509 certificate. However, it was failing to correctly
2545 handle any errors that may occur while parsing the issuer field (which might
2546 occur if the issuer field is maliciously constructed). This may subsequently
2547 result in a NULL pointer deref and a crash leading to a potential denial of
2553 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2554 padding mode to correctly check for rollback attacks. This is considered a
2555 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2560 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2561 functions. Previously they could overflow the output length argument in some
2562 cases where the input length is close to the maximum permissible length for
2563 an integer on the platform. In such cases the return value from the function
2564 call would be 1 (indicating success), but the output length value would be
2565 negative. This could cause applications to behave incorrectly or crash.
2570 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2571 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2572 could be exploited in a side channel attack to recover the password. Since
2573 the attack is local host only this is outside of the current OpenSSL
2574 threat model and therefore no CVE is assigned.
2576 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2581 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2583 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2584 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2585 If an attacker can control both items being compared then this could lead
2586 to a possible denial of service attack. OpenSSL itself uses the
2587 GENERAL_NAME_cmp function for two purposes:
2588 1) Comparing CRL distribution point names between an available CRL and a
2589 CRL distribution point embedded in an X509 certificate
2590 2) When verifying that a timestamp response token signer matches the
2591 timestamp authority name (exposed via the API functions
2592 TS_RESP_verify_response and TS_RESP_verify_token)
2597 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2599 * Certificates with explicit curve parameters are now disallowed in
2600 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2604 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2605 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2606 conversely, silently ignore DTLS protocol version bounds when configuring
2607 TLS-based contexts. The commands can be repeated to set bounds of both
2608 types. The same applies with the corresponding "min_protocol" and
2609 "max_protocol" command-line switches, in case some application uses both TLS
2612 SSL_CTX instances that are created for a fixed protocol version (e.g.
2613 TLSv1_server_method()) also silently ignore version bounds. Previously
2614 attempts to apply bounds to these protocol versions would result in an
2615 error. Now only the "version-flexible" SSL_CTX instances are subject to
2616 limits in configuration files in command-line options.
2620 * Handshake now fails if Extended Master Secret extension is dropped
2625 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2627 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2629 * Fixed segmentation fault in SSL_check_chain()
2630 Server or client applications that call the SSL_check_chain() function
2631 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2632 dereference as a result of incorrect handling of the
2633 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2634 or unrecognised signature algorithm is received from the peer. This could
2635 be exploited by a malicious peer in a Denial of Service attack.
2640 * Added AES consttime code for no-asm configurations
2641 an optional constant time support for AES was added
2642 when building openssl for no-asm.
2643 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2644 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2645 At this time this feature is by default disabled.
2646 It will be enabled by default in 3.0.
2650 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2652 * Revert the change of EOF detection while reading in libssl to avoid
2653 regressions in applications depending on the current way of reporting
2654 the EOF. As the existing method is not fully accurate the change to
2655 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2656 branch and will be present in the 3.0 release.
2660 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2661 when primes for RSA keys are computed.
2662 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2663 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2664 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2665 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2666 This avoids possible fingerprinting of newly generated RSA modules.
2670 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2672 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2673 while reading in libssl then we would report an error back to the
2674 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2675 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2676 therefore give a hint as to what went wrong.
2680 * Check that ed25519 and ed448 are allowed by the security level. Previously
2681 signature algorithms not using an MD were not being checked that they were
2682 allowed by the security level.
2686 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2687 was not quite right. The behaviour was not consistent between resumption
2688 and normal handshakes, and also not quite consistent with historical
2689 behaviour. The behaviour in various scenarios has been clarified and
2690 it has been updated to make it match historical behaviour as closely as
2695 * *[VMS only]* The header files that the VMS compilers include automatically,
2696 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2697 that the C++ compiler doesn't understand. This is a shortcoming in the
2698 compiler, but can be worked around with `__cplusplus` guards.
2700 C++ applications that use OpenSSL libraries must be compiled using the
2701 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2702 functions. Otherwise, only functions with symbols of less than 31
2703 characters can be used, as the linker will not be able to successfully
2704 resolve symbols with longer names.
2708 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2709 The presence of this system service is determined at run-time.
2713 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2718 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2720 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2721 number generator (RNG). This was intended to include protection in the
2722 event of a fork() system call in order to ensure that the parent and child
2723 processes did not share the same RNG state. However, this protection was not
2724 being used in the default case.
2726 A partial mitigation for this issue is that the output from a high
2727 precision timer is mixed into the RNG state so the likelihood of a parent
2728 and child process sharing state is significantly reduced.
2730 If an application already calls OPENSSL_init_crypto() explicitly using
2731 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2734 *Matthias St. Pierre*
2736 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2737 used even when parsing explicit parameters, when loading a encoded key
2738 or calling `EC_GROUP_new_from_ecpkparameters()`/
2739 `EC_GROUP_new_from_ecparameters()`.
2740 This prevents bypass of security hardening and performance gains,
2741 especially for curves with specialized EC_METHODs.
2742 By default, if a key encoded with explicit parameters is loaded and later
2743 encoded, the output is still encoded with explicit parameters, even if
2744 internally a "named" EC_GROUP is used for computation.
2748 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2749 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2750 NULL. After this change, only the cofactor parameter can be NULL. It also
2751 does some minimal sanity checks on the passed order.
2756 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2757 An attack is simple, if the first CMS_recipientInfo is valid but the
2758 second CMS_recipientInfo is chosen ciphertext. If the second
2759 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2760 encryption key will be replaced by garbage, and the message cannot be
2761 decoded, but if the RSA decryption fails, the correct encryption key is
2762 used and the recipient will not notice the attack.
2763 As a work around for this potential attack the length of the decrypted
2764 key must be equal to the cipher default key length, in case the
2765 certificate is not given and all recipientInfo are tried out.
2766 The old behaviour can be re-enabled in the CMS code by setting the
2767 CMS_DEBUG_DECRYPT flag.
2772 * Early start up entropy quality from the DEVRANDOM seed source has been
2773 improved for older Linux systems. The RAND subsystem will wait for
2774 /dev/random to be producing output before seeding from /dev/urandom.
2775 The seeded state is stored for future library initialisations using
2776 a system global shared memory segment. The shared memory identifier
2777 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2778 the desired value. The default identifier is 114.
2782 * Correct the extended master secret constant on EBCDIC systems. Without this
2783 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2784 negotiate EMS will fail. Unfortunately this also means that TLS connections
2785 between EBCDIC systems with this fix, and EBCDIC systems without this
2786 fix will fail if they negotiate EMS.
2790 * Use Windows installation paths in the mingw builds
2792 Mingw isn't a POSIX environment per se, which means that Windows
2793 paths should be used for installation.
2798 * Changed DH_check to accept parameters with order q and 2q subgroups.
2799 With order 2q subgroups the bit 0 of the private key is not secret
2800 but DH_generate_key works around that by clearing bit 0 of the
2801 private key for those. This avoids leaking bit 0 of the private key.
2805 * Significantly reduce secure memory usage by the randomness pools.
2809 * Revert the DEVRANDOM_WAIT feature for Linux systems
2811 The DEVRANDOM_WAIT feature added a select() call to wait for the
2812 /dev/random device to become readable before reading from the
2813 /dev/urandom device.
2815 It turned out that this change had negative side effects on
2816 performance which were not acceptable. After some discussion it
2817 was decided to revert this feature and leave it up to the OS
2818 resp. the platform maintainer to ensure a proper initialization
2819 during early boot time.
2821 *Matthias St. Pierre*
2823 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2825 * Add build tests for C++. These are generated files that only do one
2826 thing, to include one public OpenSSL head file each. This tests that
2827 the public header files can be usefully included in a C++ application.
2829 This test isn't enabled by default. It can be enabled with the option
2830 'enable-buildtest-c++'.
2834 * Enable SHA3 pre-hashing for ECDSA and DSA.
2838 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2839 This changes the size when using the `genpkey` command when no size is given.
2840 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2841 generation commands to use 2048 bits by default.
2845 * Reorganize the manual pages to consistently have RETURN VALUES,
2846 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2847 util/fix-doc-nits accordingly.
2849 *Paul Yang, Joshua Lock*
2851 * Add the missing accessor EVP_PKEY_get0_engine()
2855 * Have commands like `s_client` and `s_server` output the signature scheme
2856 along with other cipher suite parameters when debugging.
2860 * Make OPENSSL_config() error agnostic again.
2864 * Do the error handling in RSA decryption constant time.
2868 * Prevent over long nonces in ChaCha20-Poly1305.
2870 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2871 for every encryption operation. RFC 7539 specifies that the nonce value
2872 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2873 and front pads the nonce with 0 bytes if it is less than 12
2874 bytes. However it also incorrectly allows a nonce to be set of up to 16
2875 bytes. In this case only the last 12 bytes are significant and any
2876 additional leading bytes are ignored.
2878 It is a requirement of using this cipher that nonce values are
2879 unique. Messages encrypted using a reused nonce value are susceptible to
2880 serious confidentiality and integrity attacks. If an application changes
2881 the default nonce length to be longer than 12 bytes and then makes a
2882 change to the leading bytes of the nonce expecting the new value to be a
2883 new unique nonce then such an application could inadvertently encrypt
2884 messages with a reused nonce.
2886 Additionally the ignored bytes in a long nonce are not covered by the
2887 integrity guarantee of this cipher. Any application that relies on the
2888 integrity of these ignored leading bytes of a long nonce may be further
2889 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2890 is safe because no such use sets such a long nonce value. However user
2891 applications that use this cipher directly and set a non-default nonce
2892 length to be longer than 12 bytes may be vulnerable.
2894 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2900 * Add DEVRANDOM_WAIT feature for Linux systems
2902 On older Linux systems where the getrandom() system call is not available,
2903 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2904 Contrary to getrandom(), the /dev/urandom device will not block during
2905 early boot when the kernel CSPRNG has not been seeded yet.
2907 To mitigate this known weakness, use select() to wait for /dev/random to
2908 become readable before reading from /dev/urandom.
2910 * Ensure that SM2 only uses SM3 as digest algorithm
2914 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2916 * Change the info callback signals for the start and end of a post-handshake
2917 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2918 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2919 confused by this and assume that a TLSv1.2 renegotiation has started. This
2920 can break KeyUpdate handling. Instead we no longer signal the start and end
2921 of a post handshake message exchange (although the messages themselves are
2922 still signalled). This could break some applications that were expecting
2923 the old signals. However without this KeyUpdate is not usable for many
2928 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2930 * Timing vulnerability in DSA signature generation
2932 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2933 timing side channel attack. An attacker could use variations in the signing
2934 algorithm to recover the private key.
2936 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2941 * Timing vulnerability in ECDSA signature generation
2943 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2944 timing side channel attack. An attacker could use variations in the signing
2945 algorithm to recover the private key.
2947 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2952 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2953 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2954 of two gigabytes and the error handling improved.
2956 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2957 categorized as a normal bug, not a security issue, because the DRBG reseeds
2958 automatically and is fully functional even without additional randomness
2959 provided by the application.
2961 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2963 * Add a new ClientHello callback. Provides a callback interface that gives
2964 the application the ability to adjust the nascent SSL object at the
2965 earliest stage of ClientHello processing, immediately after extensions have
2966 been collected but before they have been processed. In particular, this
2967 callback can adjust the supported TLS versions in response to the contents
2972 * Add SM2 base algorithm support.
2976 * s390x assembly pack: add (improved) hardware-support for the following
2977 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2978 aes-cfb/cfb8, aes-ecb.
2982 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2983 parameter is no longer accepted, as it leads to a corrupt table. NULL
2984 pem_str is reserved for alias entries only.
2988 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2989 step for prime curves. The new implementation is based on formulae from
2990 differential addition-and-doubling in homogeneous projective coordinates
2991 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2992 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2993 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2994 to work in projective coordinates.
2996 *Billy Bob Brumley, Nicola Tuveri*
2998 * Change generating and checking of primes so that the error rate of not
2999 being prime depends on the intended use based on the size of the input.
3000 For larger primes this will result in more rounds of Miller-Rabin.
3001 The maximal error rate for primes with more than 1080 bits is lowered
3004 *Kurt Roeckx, Annie Yousar*
3006 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3010 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
3011 moving between systems, and to avoid confusion when a Windows build is
3012 done with mingw vs with MSVC. For POSIX installs, there's still a
3013 symlink or copy named 'tsget' to avoid that confusion as well.
3017 * Revert blinding in ECDSA sign and instead make problematic addition
3018 length-invariant. Switch even to fixed-length Montgomery multiplication.
3022 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3023 step for binary curves. The new implementation is based on formulae from
3024 differential addition-and-doubling in mixed Lopez-Dahab projective
3025 coordinates, modified to independently blind the operands.
3027 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3029 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3030 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3031 EC_METHODs to implement their own specialized "ladder step", to take
3032 advantage of more favorable coordinate systems or more efficient
3033 differential addition-and-doubling algorithms.
3035 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3037 * Modified the random device based seed sources to keep the relevant
3038 file descriptors open rather than reopening them on each access.
3039 This allows such sources to operate in a chroot() jail without
3040 the associated device nodes being available. This behaviour can be
3041 controlled using RAND_keep_random_devices_open().
3045 * Numerous side-channel attack mitigations have been applied. This may have
3046 performance impacts for some algorithms for the benefit of improved
3047 security. Specific changes are noted in this change log by their respective
3052 * AIX shared library support overhaul. Switch to AIX "natural" way of
3053 handling shared libraries, which means collecting shared objects of
3054 different versions and bitnesses in one common archive. This allows to
3055 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3056 doesn't affect the way 3rd party applications are linked, only how
3057 multi-version installation is managed.
3061 * Make ec_group_do_inverse_ord() more robust and available to other
3062 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3063 mitigations are applied to the fallback BN_mod_inverse().
3064 When using this function rather than BN_mod_inverse() directly, new
3065 EC cryptosystem implementations are then safer-by-default.
3069 * Add coordinate blinding for EC_POINT and implement projective
3070 coordinate blinding for generic prime curves as a countermeasure to
3071 chosen point SCA attacks.
3073 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3075 * Add blinding to ECDSA and DSA signatures to protect against side channel
3076 attacks discovered by Keegan Ryan (NCC Group).
3080 * Enforce checking in the `pkeyutl` command to ensure that the input
3081 length does not exceed the maximum supported digest length when performing
3082 a sign, verify or verifyrecover operation.
3086 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3087 I/O in combination with something like select() or poll() will hang. This
3088 can be turned off again using SSL_CTX_clear_mode().
3089 Many applications do not properly handle non-application data records, and
3090 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3091 around the problems in those applications, but can also break some.
3092 It's recommended to read the manpages about SSL_read(), SSL_write(),
3093 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3094 SSL_CTX_set_read_ahead() again.
3098 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3099 now allow empty (zero character) pass phrases.
3103 * Apply blinding to binary field modular inversion and remove patent
3104 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3108 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3109 binary and prime elliptic curves.
3113 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3114 constant time fixed point multiplication.
3118 * Revise elliptic curve scalar multiplication with timing attack
3119 defenses: ec_wNAF_mul redirects to a constant time implementation
3120 when computing fixed point and variable point multiplication (which
3121 in OpenSSL are mostly used with secret scalars in keygen, sign,
3122 ECDH derive operations).
3123 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3126 * Updated CONTRIBUTING
3130 * Updated DRBG / RAND to request nonce and additional low entropy
3131 randomness from the system.
3133 *Matthias St. Pierre*
3135 * Updated 'openssl rehash' to use OpenSSL consistent default.
3139 * Moved the load of the ssl_conf module to libcrypto, which helps
3140 loading engines that libssl uses before libssl is initialised.
3144 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3148 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3150 *Ingo Schwarze, Rich Salz*
3152 * Added output of accepting IP address and port for 'openssl s_server'
3156 * Added a new API for TLSv1.3 ciphersuites:
3157 SSL_CTX_set_ciphersuites()
3158 SSL_set_ciphersuites()
3162 * Memory allocation failures consistently add an error to the error
3167 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3168 in libcrypto when run as setuid/setgid.
3172 * Load any config file by default when libssl is used.
3176 * Added new public header file <openssl/rand_drbg.h> and documentation
3177 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3179 *Matthias St. Pierre*
3181 * QNX support removed (cannot find contributors to get their approval
3182 for the license change).
3186 * TLSv1.3 replay protection for early data has been implemented. See the
3187 SSL_read_early_data() man page for further details.
3191 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3192 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3193 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3194 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3195 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3196 configuration has been separated out. See the ciphers man page or the
3197 SSL_CTX_set_ciphersuites() man page for more information.
3201 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3202 in responder mode now supports the new "-multi" option, which
3203 spawns the specified number of child processes to handle OCSP
3204 requests. The "-timeout" option now also limits the OCSP
3205 responder's patience to wait to receive the full client request
3206 on a newly accepted connection. Child processes are respawned
3207 as needed, and the CA index file is automatically reloaded
3208 when changed. This makes it possible to run the "ocsp" responder
3209 as a long-running service, making the OpenSSL CA somewhat more
3210 feature-complete. In this mode, most diagnostic messages logged
3211 after entering the event loop are logged via syslog(3) rather than
3216 * Added support for X448 and Ed448. Heavily based on original work by
3221 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3222 objects loaded. This adds the functions OSSL_STORE_expect() and
3223 OSSL_STORE_find() as well as needed tools to construct searches and
3224 get the search data out of them.
3228 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3229 version of OpenSSL should review their configuration settings to ensure
3230 that they are still appropriate for TLSv1.3. For further information see:
3231 <https://wiki.openssl.org/index.php/TLS1.3>
3235 * Grand redesign of the OpenSSL random generator
3237 The default RAND method now utilizes an AES-CTR DRBG according to
3238 NIST standard SP 800-90Ar1. The new random generator is essentially
3239 a port of the default random generator from the OpenSSL FIPS 2.0
3240 object module. It is a hybrid deterministic random bit generator
3241 using an AES-CTR bit stream and which seeds and reseeds itself
3242 automatically using trusted system entropy sources.
3244 Some of its new features are:
3245 - Support for multiple DRBG instances with seed chaining.
3246 - The default RAND method makes use of a DRBG.
3247 - There is a public and private DRBG instance.
3248 - The DRBG instances are fork-safe.
3249 - Keep all global DRBG instances on the secure heap if it is enabled.
3250 - The public and private DRBG instance are per thread for lock free
3253 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3255 * Changed Configure so it only says what it does and doesn't dump
3256 so much data. Instead, ./configdata.pm should be used as a script
3257 to display all sorts of configuration data.
3261 * Added processing of "make variables" to Configure.
3265 * Added SHA512/224 and SHA512/256 algorithm support.
3269 * The last traces of Netware support, first removed in 1.1.0, have
3274 * Get rid of Makefile.shared, and in the process, make the processing
3275 of certain files (rc.obj, or the .def/.map/.opt files produced from
3276 the ordinal files) more visible and hopefully easier to trace and
3277 debug (or make silent).
3281 * Make it possible to have environment variable assignments as
3282 arguments to config / Configure.
3286 * Add multi-prime RSA (RFC 8017) support.
3290 * Add SM3 implemented according to GB/T 32905-2016
3291 *Jack Lloyd <jack.lloyd@ribose.com>,*
3292 *Ronald Tse <ronald.tse@ribose.com>,*
3293 *Erick Borsboom <erick.borsboom@ribose.com>*
3295 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3296 as documented in RFC6066.
3297 Based on a patch from Tomasz Moń
3299 *Filipe Raimundo da Silva*
3301 * Add SM4 implemented according to GB/T 32907-2016.
3302 *Jack Lloyd <jack.lloyd@ribose.com>,*
3303 *Ronald Tse <ronald.tse@ribose.com>,*
3304 *Erick Borsboom <erick.borsboom@ribose.com>*
3306 * Reimplement -newreq-nodes and ERR_error_string_n; the
3307 original author does not agree with the license change.
3311 * Add ARIA AEAD TLS support.
3315 * Some macro definitions to support VS6 have been removed. Visual
3316 Studio 6 has not worked since 1.1.0
3320 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3321 without clearing the errors.
3325 * Add "atfork" functions. If building on a system that without
3326 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3327 requirements. The RAND facility now uses/requires this.
3335 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3336 not possible to disable entirely. However, it's still possible to
3337 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3340 To disable, configure with 'no-ui-console'. 'no-ui' is still
3341 possible to use as an alias. Check at compile time with the
3342 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3343 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3347 * Add a STORE module, which implements a uniform and URI based reader of
3348 stores that can contain keys, certificates, CRLs and numerous other
3349 objects. The main API is loosely based on a few stdio functions,
3350 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3351 OSSL_STORE_error and OSSL_STORE_close.
3352 The implementation uses backends called "loaders" to implement arbitrary
3353 URI schemes. There is one built in "loader" for the 'file' scheme.
3357 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3358 then adjusted to work on FreeBSD 8.4 as well.
3359 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3360 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3364 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3365 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3366 error code calls like this:
3368 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3370 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3371 that can be encoded in C. For the foreseeable future, this will only
3374 *Richard Levitte and Tim Hudson*
3376 * Removed BSD cryptodev engine.
3380 * Add a build target 'build_all_generated', to build all generated files
3381 and only that. This can be used to prepare everything that requires
3382 things like perl for a system that lacks perl and then move everything
3383 to that system and do the rest of the build there.
3387 * In the UI interface, make it possible to duplicate the user data. This
3388 can be used by engines that need to retain the data for a longer time
3389 than just the call where this user data is passed.
3393 * Ignore the '-named_curve auto' value for compatibility of applications
3396 *Tomáš Mráz <tmraz@fedoraproject.org>*
3398 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3399 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3400 alerts across multiple records (some of which could be empty). In practice
3401 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3402 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3403 support this at all. Supporting it adds significant complexity to the
3404 record layer, and its removal is unlikely to cause interoperability
3409 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3410 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3411 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3416 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3417 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3419 *Richard Levitte, Andy Polyakov*
3421 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3426 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3427 platform rather than 'mingw'.
3431 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3432 success if they are asked to add an object which already exists
3433 in the store. This change cascades to other functions which load
3434 certificates and CRLs.
3438 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3439 facilitate stack unwinding even from assembly subroutines.
3443 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3444 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3448 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3449 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3450 which is the minimum version we support.
3454 * Certificate time validation (X509_cmp_time) enforces stricter
3455 compliance with RFC 5280. Fractional seconds and timezone offsets
3456 are no longer allowed.
3460 * Add support for ARIA
3464 * s_client will now send the Server Name Indication (SNI) extension by
3465 default unless the new "-noservername" option is used. The server name is
3466 based on the host provided to the "-connect" option unless overridden by
3467 using "-servername".
3471 * Add support for SipHash
3475 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3476 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3477 prevent issues where no progress is being made and the peer continually
3478 sends unrecognised record types, using up resources processing them.
3482 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3483 using the algorithm defined in
3484 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3488 * Heartbeat support has been removed; the ABI is changed for now.
3490 *Richard Levitte, Rich Salz*
3492 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3496 * The RSA "null" method, which was partially supported to avoid patent
3497 issues, has been replaced to always returns NULL.
3504 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3506 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3507 used even when parsing explicit parameters, when loading a encoded key
3508 or calling `EC_GROUP_new_from_ecpkparameters()`/
3509 `EC_GROUP_new_from_ecparameters()`.
3510 This prevents bypass of security hardening and performance gains,
3511 especially for curves with specialized EC_METHODs.
3512 By default, if a key encoded with explicit parameters is loaded and later
3513 encoded, the output is still encoded with explicit parameters, even if
3514 internally a "named" EC_GROUP is used for computation.
3518 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3519 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3520 NULL. After this change, only the cofactor parameter can be NULL. It also
3521 does some minimal sanity checks on the passed order.
3526 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3527 An attack is simple, if the first CMS_recipientInfo is valid but the
3528 second CMS_recipientInfo is chosen ciphertext. If the second
3529 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3530 encryption key will be replaced by garbage, and the message cannot be
3531 decoded, but if the RSA decryption fails, the correct encryption key is
3532 used and the recipient will not notice the attack.
3533 As a work around for this potential attack the length of the decrypted
3534 key must be equal to the cipher default key length, in case the
3535 certificate is not given and all recipientInfo are tried out.
3536 The old behaviour can be re-enabled in the CMS code by setting the
3537 CMS_DEBUG_DECRYPT flag.
3542 * Use Windows installation paths in the mingw builds
3544 Mingw isn't a POSIX environment per se, which means that Windows
3545 paths should be used for installation.
3550 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3552 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3553 This changes the size when using the `genpkey` command when no size is given.
3554 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3555 generation commands to use 2048 bits by default.
3559 * Prevent over long nonces in ChaCha20-Poly1305.
3561 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3562 for every encryption operation. RFC 7539 specifies that the nonce value
3563 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3564 and front pads the nonce with 0 bytes if it is less than 12
3565 bytes. However it also incorrectly allows a nonce to be set of up to 16
3566 bytes. In this case only the last 12 bytes are significant and any
3567 additional leading bytes are ignored.
3569 It is a requirement of using this cipher that nonce values are
3570 unique. Messages encrypted using a reused nonce value are susceptible to
3571 serious confidentiality and integrity attacks. If an application changes
3572 the default nonce length to be longer than 12 bytes and then makes a
3573 change to the leading bytes of the nonce expecting the new value to be a
3574 new unique nonce then such an application could inadvertently encrypt
3575 messages with a reused nonce.
3577 Additionally the ignored bytes in a long nonce are not covered by the
3578 integrity guarantee of this cipher. Any application that relies on the
3579 integrity of these ignored leading bytes of a long nonce may be further
3580 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3581 is safe because no such use sets such a long nonce value. However user
3582 applications that use this cipher directly and set a non-default nonce
3583 length to be longer than 12 bytes may be vulnerable.
3585 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3591 * Added SCA hardening for modular field inversion in EC_GROUP through
3592 a new dedicated field_inv() pointer in EC_METHOD.
3593 This also addresses a leakage affecting conversions from projective
3594 to affine coordinates.
3596 *Billy Bob Brumley, Nicola Tuveri*
3598 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3599 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3603 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3607 * Remove the 'dist' target and add a tarball building script. The
3608 'dist' target has fallen out of use, and it shouldn't be
3609 necessary to configure just to create a source distribution.
3613 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3615 * Timing vulnerability in DSA signature generation
3617 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3618 timing side channel attack. An attacker could use variations in the signing
3619 algorithm to recover the private key.
3621 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3626 * Timing vulnerability in ECDSA signature generation
3628 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3629 timing side channel attack. An attacker could use variations in the signing
3630 algorithm to recover the private key.
3632 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3637 * Add coordinate blinding for EC_POINT and implement projective
3638 coordinate blinding for generic prime curves as a countermeasure to
3639 chosen point SCA attacks.
3641 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3643 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3645 * Client DoS due to large DH parameter
3647 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3648 malicious server can send a very large prime value to the client. This will
3649 cause the client to spend an unreasonably long period of time generating a
3650 key for this prime resulting in a hang until the client has finished. This
3651 could be exploited in a Denial Of Service attack.
3653 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3658 * Cache timing vulnerability in RSA Key Generation
3660 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3661 a cache timing side channel attack. An attacker with sufficient access to
3662 mount cache timing attacks during the RSA key generation process could
3663 recover the private key.
3665 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3666 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3671 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3672 parameter is no longer accepted, as it leads to a corrupt table. NULL
3673 pem_str is reserved for alias entries only.
3677 * Revert blinding in ECDSA sign and instead make problematic addition
3678 length-invariant. Switch even to fixed-length Montgomery multiplication.
3682 * Change generating and checking of primes so that the error rate of not
3683 being prime depends on the intended use based on the size of the input.
3684 For larger primes this will result in more rounds of Miller-Rabin.
3685 The maximal error rate for primes with more than 1080 bits is lowered
3688 *Kurt Roeckx, Annie Yousar*
3690 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3694 * Add blinding to ECDSA and DSA signatures to protect against side channel
3695 attacks discovered by Keegan Ryan (NCC Group).
3699 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3700 now allow empty (zero character) pass phrases.
3704 * Certificate time validation (X509_cmp_time) enforces stricter
3705 compliance with RFC 5280. Fractional seconds and timezone offsets
3706 are no longer allowed.
3710 * Fixed a text canonicalisation bug in CMS
3712 Where a CMS detached signature is used with text content the text goes
3713 through a canonicalisation process first prior to signing or verifying a
3714 signature. This process strips trailing space at the end of lines, converts
3715 line terminators to CRLF and removes additional trailing line terminators
3716 at the end of a file. A bug in the canonicalisation process meant that
3717 some characters, such as form-feed, were incorrectly treated as whitespace
3718 and removed. This is contrary to the specification (RFC5485). This fix
3719 could mean that detached text data signed with an earlier version of
3720 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3721 signed with a fixed OpenSSL may fail to verify with an earlier version of
3722 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3723 and use the "-binary" flag (for the "cms" command line application) or set
3724 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3728 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3730 * Constructed ASN.1 types with a recursive definition could exceed the stack
3732 Constructed ASN.1 types with a recursive definition (such as can be found
3733 in PKCS7) could eventually exceed the stack given malicious input with
3734 excessive recursion. This could result in a Denial Of Service attack. There
3735 are no such structures used within SSL/TLS that come from untrusted sources
3736 so this is considered safe.
3738 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3744 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3746 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3747 effectively reduced to only comparing the least significant bit of each
3748 byte. This allows an attacker to forge messages that would be considered as
3749 authenticated in an amount of tries lower than that guaranteed by the
3750 security claims of the scheme. The module can only be compiled by the
3751 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3753 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3759 * Add a build target 'build_all_generated', to build all generated files
3760 and only that. This can be used to prepare everything that requires
3761 things like perl for a system that lacks perl and then move everything
3762 to that system and do the rest of the build there.
3766 * Backport SSL_OP_NO_RENGOTIATION
3768 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3769 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3770 changes this is no longer possible in 1.1.0. Therefore, the new
3771 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3772 1.1.0 to provide equivalent functionality.
3774 Note that if an application built against 1.1.0h headers (or above) is run
3775 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3776 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3780 * Removed the OS390-Unix config target. It relied on a script that doesn't
3785 * rsaz_1024_mul_avx2 overflow bug on x86_64
3787 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3788 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3789 Analysis suggests that attacks against RSA and DSA as a result of this
3790 defect would be very difficult to perform and are not believed likely.
3791 Attacks against DH1024 are considered just feasible, because most of the
3792 work necessary to deduce information about a private key may be performed
3793 offline. The amount of resources required for such an attack would be
3794 significant. However, for an attack on TLS to be meaningful, the server
3795 would have to share the DH1024 private key among multiple clients, which is
3796 no longer an option since CVE-2016-0701.
3798 This only affects processors that support the AVX2 but not ADX extensions
3799 like Intel Haswell (4th generation).
3801 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3802 was originally found via the OSS-Fuzz project.
3807 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3809 * bn_sqrx8x_internal carry bug on x86_64
3811 There is a carry propagating bug in the x86_64 Montgomery squaring
3812 procedure. No EC algorithms are affected. Analysis suggests that attacks
3813 against RSA and DSA as a result of this defect would be very difficult to
3814 perform and are not believed likely. Attacks against DH are considered just
3815 feasible (although very difficult) because most of the work necessary to
3816 deduce information about a private key may be performed offline. The amount
3817 of resources required for such an attack would be very significant and
3818 likely only accessible to a limited number of attackers. An attacker would
3819 additionally need online access to an unpatched system using the target
3820 private key in a scenario with persistent DH parameters and a private
3821 key that is shared between multiple clients.
3823 This only affects processors that support the BMI1, BMI2 and ADX extensions
3824 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3826 This issue was reported to OpenSSL by the OSS-Fuzz project.
3831 * Malformed X.509 IPAddressFamily could cause OOB read
3833 If an X.509 certificate has a malformed IPAddressFamily extension,
3834 OpenSSL could do a one-byte buffer overread. The most likely result
3835 would be an erroneous display of the certificate in text format.
3837 This issue was reported to OpenSSL by the OSS-Fuzz project.
3842 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3844 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3845 platform rather than 'mingw'.
3849 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3850 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3851 which is the minimum version we support.
3855 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3857 * Encrypt-Then-Mac renegotiation crash
3859 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3860 negotiated where it was not in the original handshake (or vice-versa) then
3861 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3862 and servers are affected.
3864 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3869 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3871 * Truncated packet could crash via OOB read
3873 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3874 cipher is being used, then a truncated packet can cause that host to
3875 perform an out-of-bounds read, usually resulting in a crash.
3877 This issue was reported to OpenSSL by Robert Święcki of Google.
3882 * Bad (EC)DHE parameters cause a client crash
3884 If a malicious server supplies bad parameters for a DHE or ECDHE key
3885 exchange then this can result in the client attempting to dereference a
3886 NULL pointer leading to a client crash. This could be exploited in a Denial
3889 This issue was reported to OpenSSL by Guido Vranken.
3894 * BN_mod_exp may produce incorrect results on x86_64
3896 There is a carry propagating bug in the x86_64 Montgomery squaring
3897 procedure. No EC algorithms are affected. Analysis suggests that attacks
3898 against RSA and DSA as a result of this defect would be very difficult to
3899 perform and are not believed likely. Attacks against DH are considered just
3900 feasible (although very difficult) because most of the work necessary to
3901 deduce information about a private key may be performed offline. The amount
3902 of resources required for such an attack would be very significant and
3903 likely only accessible to a limited number of attackers. An attacker would
3904 additionally need online access to an unpatched system using the target
3905 private key in a scenario with persistent DH parameters and a private
3906 key that is shared between multiple clients. For example this can occur by
3907 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3908 similar to CVE-2015-3193 but must be treated as a separate problem.
3910 This issue was reported to OpenSSL by the OSS-Fuzz project.
3915 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3917 * ChaCha20/Poly1305 heap-buffer-overflow
3919 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3920 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3921 crash. This issue is not considered to be exploitable beyond a DoS.
3923 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3928 * CMS Null dereference
3930 Applications parsing invalid CMS structures can crash with a NULL pointer
3931 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3932 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3933 structure callback if an attempt is made to free certain invalid encodings.
3934 Only CHOICE structures using a callback which do not handle NULL value are
3937 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3942 * Montgomery multiplication may produce incorrect results
3944 There is a carry propagating bug in the Broadwell-specific Montgomery
3945 multiplication procedure that handles input lengths divisible by, but
3946 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3947 and DH private keys are impossible. This is because the subroutine in
3948 question is not used in operations with the private key itself and an input
3949 of the attacker's direct choice. Otherwise the bug can manifest itself as
3950 transient authentication and key negotiation failures or reproducible
3951 erroneous outcome of public-key operations with specially crafted input.
3952 Among EC algorithms only Brainpool P-512 curves are affected and one
3953 presumably can attack ECDH key negotiation. Impact was not analyzed in
3954 detail, because pre-requisites for attack are considered unlikely. Namely
3955 multiple clients have to choose the curve in question and the server has to
3956 share the private key among them, neither of which is default behaviour.
3957 Even then only clients that chose the curve will be affected.
3959 This issue was publicly reported as transient failures and was not
3960 initially recognized as a security issue. Thanks to Richard Morgan for
3961 providing reproducible case.
3966 * Removed automatic addition of RPATH in shared libraries and executables,
3967 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3971 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3973 * Fix Use After Free for large message sizes
3975 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3976 message larger than approx 16k is received then the underlying buffer to
3977 store the incoming message is reallocated and moved. Unfortunately a
3978 dangling pointer to the old location is left which results in an attempt to
3979 write to the previously freed location. This is likely to result in a
3980 crash, however it could potentially lead to execution of arbitrary code.
3982 This issue only affects OpenSSL 1.1.0a.
3984 This issue was reported to OpenSSL by Robert Święcki.
3989 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3991 * OCSP Status Request extension unbounded memory growth
3993 A malicious client can send an excessively large OCSP Status Request
3994 extension. If that client continually requests renegotiation, sending a
3995 large OCSP Status Request extension each time, then there will be unbounded
3996 memory growth on the server. This will eventually lead to a Denial Of
3997 Service attack through memory exhaustion. Servers with a default
3998 configuration are vulnerable even if they do not support OCSP. Builds using
3999 the "no-ocsp" build time option are not affected.
4001 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4006 * SSL_peek() hang on empty record
4008 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
4009 sends an empty record. This could be exploited by a malicious peer in a
4010 Denial Of Service attack.
4012 This issue was reported to OpenSSL by Alex Gaynor.
4017 * Excessive allocation of memory in tls_get_message_header() and
4018 dtls1_preprocess_fragment()
4020 A (D)TLS message includes 3 bytes for its length in the header for the
4021 message. This would allow for messages up to 16Mb in length. Messages of
4022 this length are excessive and OpenSSL includes a check to ensure that a
4023 peer is sending reasonably sized messages in order to avoid too much memory
4024 being consumed to service a connection. A flaw in the logic of version
4025 1.1.0 means that memory for the message is allocated too early, prior to
4026 the excessive message length check. Due to way memory is allocated in
4027 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4028 to service a connection. This could lead to a Denial of Service through
4029 memory exhaustion. However, the excessive message length check still takes
4030 place, and this would cause the connection to immediately fail. Assuming
4031 that the application calls SSL_free() on the failed connection in a timely
4032 manner then the 21Mb of allocated memory will then be immediately freed
4033 again. Therefore, the excessive memory allocation will be transitory in
4034 nature. This then means that there is only a security impact if:
4036 1) The application does not call SSL_free() in a timely manner in the event
4037 that the connection fails
4039 2) The application is working in a constrained environment where there is
4040 very little free memory
4042 3) The attacker initiates multiple connection attempts such that there are
4043 multiple connections in a state where memory has been allocated for the
4044 connection; SSL_free() has not yet been called; and there is insufficient
4045 memory to service the multiple requests.
4047 Except in the instance of (1) above any Denial Of Service is likely to be
4048 transitory because as soon as the connection fails the memory is
4049 subsequently freed again in the SSL_free() call. However there is an
4050 increased risk during this period of application crashes due to the lack of
4051 memory - which would then mean a more serious Denial of Service.
4053 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4054 (CVE-2016-6307 and CVE-2016-6308)
4058 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4059 had to be removed. Primary reason is that vendor assembler can't
4060 assemble our modules with -KPIC flag. As result it, assembly
4061 support, was not even available as option. But its lack means
4062 lack of side-channel resistant code, which is incompatible with
4063 security by todays standards. Fortunately gcc is readily available
4064 prepackaged option, which we firmly point at...
4068 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4070 * Windows command-line tool supports UTF-8 opt-in option for arguments
4071 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4072 (to any value) allows Windows user to access PKCS#12 file generated
4073 with Windows CryptoAPI and protected with non-ASCII password, as well
4074 as files generated under UTF-8 locale on Linux also protected with
4079 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4080 have been disabled by default and removed from DEFAULT, just like RC4.
4081 See the RC4 item below to re-enable both.
4085 * The method for finding the storage location for the Windows RAND seed file
4086 has changed. First we check %RANDFILE%. If that is not set then we check
4087 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4088 all else fails we fall back to C:\.
4092 * The EVP_EncryptUpdate() function has had its return type changed from void
4093 to int. A return of 0 indicates and error while a return of 1 indicates
4098 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4099 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4100 off the constant time implementation for RSA, DSA and DH have been made
4101 no-ops and deprecated.
4105 * Windows RAND implementation was simplified to only get entropy by
4106 calling CryptGenRandom(). Various other RAND-related tickets
4109 *Joseph Wylie Yandle, Rich Salz*
4111 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4112 and `OPENSSL_LH_`, respectively. The old names are available
4113 with API compatibility. They new names are now completely documented.
4117 * Unify TYPE_up_ref(obj) methods signature.
4118 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4119 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4120 int (instead of void) like all others TYPE_up_ref() methods.
4121 So now these methods also check the return value of CRYPTO_atomic_add(),
4122 and the validity of object reference counter.
4124 *fdasilvayy@gmail.com*
4126 * With Windows Visual Studio builds, the .pdb files are installed
4127 alongside the installed libraries and executables. For a static
4128 library installation, ossl_static.pdb is the associate compiler
4129 generated .pdb file to be used when linking programs.
4133 * Remove openssl.spec. Packaging files belong with the packagers.
4137 * Automatic Darwin/OSX configuration has had a refresh, it will now
4138 recognise x86_64 architectures automatically. You can still decide
4139 to build for a different bitness with the environment variable
4140 KERNEL_BITS (can be 32 or 64), for example:
4142 KERNEL_BITS=32 ./config
4146 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4147 256 bit AES and HMAC with SHA256.
4151 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4155 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4159 * To enable users to have their own config files and build file templates,
4160 Configure looks in the directory indicated by the environment variable
4161 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4162 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4163 name and is used as is.
4167 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4168 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4169 X509_CERT_FILE_CTX was removed.
4173 * "shared" builds are now the default. To create only static libraries use
4174 the "no-shared" Configure option.
4178 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4179 All of these option have not worked for some while and are fundamental
4184 * Make various cleanup routines no-ops and mark them as deprecated. Most
4185 global cleanup functions are no longer required because they are handled
4186 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4187 Explicitly de-initing can cause problems (e.g. where a library that uses
4188 OpenSSL de-inits, but an application is still using it). The affected
4189 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4190 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4191 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4192 COMP_zlib_cleanup().
4196 * --strict-warnings no longer enables runtime debugging options
4197 such as REF_DEBUG. Instead, debug options are automatically
4198 enabled with '--debug' builds.
4200 *Andy Polyakov, Emilia Käsper*
4202 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4203 have been moved out of the public header files. New functions for managing
4204 these have been added.
4208 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4209 objects have been moved out of the public header files. New
4210 functions for managing these have been added.
4214 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4215 have been moved out of the public header files. New functions for managing
4216 these have been added.
4220 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4221 moved out of the public header files. New functions for managing these
4226 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4230 * Removed the mk1mf build scripts.
4234 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4235 it is always safe to #include a header now.
4239 * Removed the aged BC-32 config and all its supporting scripts
4243 * Removed support for Ultrix, Netware, and OS/2.
4247 * Add support for HKDF.
4249 *Alessandro Ghedini*
4251 * Add support for blake2b and blake2s
4255 * Added support for "pipelining". Ciphers that have the
4256 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4257 encryptions/decryptions simultaneously. There are currently no built-in
4258 ciphers with this property but the expectation is that engines will be able
4259 to offer it to significantly improve throughput. Support has been extended
4260 into libssl so that multiple records for a single connection can be
4261 processed in one go (for >=TLS 1.1).
4265 * Added the AFALG engine. This is an async capable engine which is able to
4266 offload work to the Linux kernel. In this initial version it only supports
4267 AES128-CBC. The kernel must be version 4.1.0 or greater.
4271 * OpenSSL now uses a new threading API. It is no longer necessary to
4272 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4273 are two supported threading models: pthreads and windows threads. It is
4274 also possible to configure OpenSSL at compile time for "no-threads". The
4275 old threading API should no longer be used. The functions have been
4276 replaced with "no-op" compatibility macros.
4278 *Alessandro Ghedini, Matt Caswell*
4280 * Modify behavior of ALPN to invoke callback after SNI/servername
4281 callback, such that updates to the SSL_CTX affect ALPN.
4285 * Add SSL_CIPHER queries for authentication and key-exchange.
4289 * Changes to the DEFAULT cipherlist:
4290 - Prefer (EC)DHE handshakes over plain RSA.
4291 - Prefer AEAD ciphers over legacy ciphers.
4292 - Prefer ECDSA over RSA when both certificates are available.
4293 - Prefer TLSv1.2 ciphers/PRF.
4294 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4299 * Change the ECC default curve list to be this, in order: x25519,
4300 secp256r1, secp521r1, secp384r1.
4304 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4305 disabled by default. They can be re-enabled using the
4306 enable-weak-ssl-ciphers option to Configure.
4310 * If the server has ALPN configured, but supports no protocols that the
4311 client advertises, send a fatal "no_application_protocol" alert.
4312 This behaviour is SHALL in RFC 7301, though it isn't universally
4313 implemented by other servers.
4317 * Add X25519 support.
4318 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4319 for public and private key encoding using the format documented in
4320 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4321 key generation and key derivation.
4323 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4328 * Deprecate SRP_VBASE_get_by_user.
4329 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4330 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4331 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4332 seed, even if the seed is configured.
4334 Users should use SRP_VBASE_get1_by_user instead. Note that in
4335 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4336 also that even though configuring the SRP seed attempts to hide
4337 invalid usernames by continuing the handshake with fake
4338 credentials, this behaviour is not constant time and no strong
4339 guarantees are made that the handshake is indistinguishable from
4340 that of a valid user.
4344 * Configuration change; it's now possible to build dynamic engines
4345 without having to build shared libraries and vice versa. This
4346 only applies to the engines in `engines/`, those in `crypto/engine/`
4347 will always be built into libcrypto (i.e. "static").
4349 Building dynamic engines is enabled by default; to disable, use
4350 the configuration option "disable-dynamic-engine".
4352 The only requirements for building dynamic engines are the
4353 presence of the DSO module and building with position independent
4354 code, so they will also automatically be disabled if configuring
4355 with "disable-dso" or "disable-pic".
4357 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4358 are also taken away from openssl/opensslconf.h, as they are
4363 * Configuration change; if there is a known flag to compile
4364 position independent code, it will always be applied on the
4365 libcrypto and libssl object files, and never on the application
4366 object files. This means other libraries that use routines from
4367 libcrypto / libssl can be made into shared libraries regardless
4368 of how OpenSSL was configured.
4370 If this isn't desirable, the configuration options "disable-pic"
4371 or "no-pic" can be used to disable the use of PIC. This will
4372 also disable building shared libraries and dynamic engines.
4376 * Removed JPAKE code. It was experimental and has no wide use.
4380 * The INSTALL_PREFIX Makefile variable has been renamed to
4381 DESTDIR. That makes for less confusion on what this variable
4382 is for. Also, the configuration option --install_prefix is
4387 * Heartbeat for TLS has been removed and is disabled by default
4388 for DTLS; configure with enable-heartbeats. Code that uses the
4389 old #define's might need to be updated.
4391 *Emilia Käsper, Rich Salz*
4393 * Rename REF_CHECK to REF_DEBUG.
4397 * New "unified" build system
4399 The "unified" build system is aimed to be a common system for all
4400 platforms we support. With it comes new support for VMS.
4402 This system builds supports building in a different directory tree
4403 than the source tree. It produces one Makefile (for unix family
4404 or lookalikes), or one descrip.mms (for VMS).
4406 The source of information to make the Makefile / descrip.mms is
4407 small files called 'build.info', holding the necessary
4408 information for each directory with source to compile, and a
4409 template in Configurations, like unix-Makefile.tmpl or
4412 With this change, the library names were also renamed on Windows
4413 and on VMS. They now have names that are closer to the standard
4414 on Unix, and include the major version number, and in certain
4415 cases, the architecture they are built for. See "Notes on shared
4416 libraries" in INSTALL.
4418 We rely heavily on the perl module Text::Template.
4422 * Added support for auto-initialisation and de-initialisation of the library.
4423 OpenSSL no longer requires explicit init or deinit routines to be called,
4424 except in certain circumstances. See the OPENSSL_init_crypto() and
4425 OPENSSL_init_ssl() man pages for further information.
4429 * The arguments to the DTLSv1_listen function have changed. Specifically the
4430 "peer" argument is now expected to be a BIO_ADDR object.
4432 * Rewrite of BIO networking library. The BIO library lacked consistent
4433 support of IPv6, and adding it required some more extensive
4434 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4435 which hold all types of addresses and chains of address information.
4436 It also introduces a new API, with functions like BIO_socket,
4437 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4438 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4439 have been adapted accordingly.
4443 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4448 * CRIME protection: disable compression by default, even if OpenSSL is
4449 compiled with zlib enabled. Applications can still enable compression
4450 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4451 using the SSL_CONF library to configure compression.
4455 * The signature of the session callback configured with
4456 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4457 was explicitly marked as `const unsigned char*` instead of
4462 * Always DPURIFY. Remove the use of uninitialized memory in the
4463 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4467 * Removed many obsolete configuration items, including
4468 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4469 MD2_CHAR, MD2_INT, MD2_LONG
4471 IDEA_SHORT, IDEA_LONG
4472 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4474 *Rich Salz, with advice from Andy Polyakov*
4476 * Many BN internals have been moved to an internal header file.
4478 *Rich Salz with help from Andy Polyakov*
4480 * Configuration and writing out the results from it has changed.
4481 Files such as Makefile include/openssl/opensslconf.h and are now
4482 produced through general templates, such as Makefile.in and
4483 crypto/opensslconf.h.in and some help from the perl module
4486 Also, the center of configuration information is no longer
4487 Makefile. Instead, Configure produces a perl module in
4488 configdata.pm which holds most of the config data (in the hash
4489 table %config), the target data that comes from the target
4490 configuration in one of the `Configurations/*.conf` files (in
4495 * To clarify their intended purposes, the Configure options
4496 --prefix and --openssldir change their semantics, and become more
4497 straightforward and less interdependent.
4499 --prefix shall be used exclusively to give the location INSTALLTOP
4500 where programs, scripts, libraries, include files and manuals are
4501 going to be installed. The default is now /usr/local.
4503 --openssldir shall be used exclusively to give the default
4504 location OPENSSLDIR where certificates, private keys, CRLs are
4505 managed. This is also where the default openssl.cnf gets
4507 If the directory given with this option is a relative path, the
4508 values of both the --prefix value and the --openssldir value will
4509 be combined to become OPENSSLDIR.
4510 The default for --openssldir is INSTALLTOP/ssl.
4512 Anyone who uses --openssldir to specify where OpenSSL is to be
4513 installed MUST change to use --prefix instead.
4517 * The GOST engine was out of date and therefore it has been removed. An up
4518 to date GOST engine is now being maintained in an external repository.
4519 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4520 support for GOST ciphersuites (these are only activated if a GOST engine
4525 * EGD is no longer supported by default; use enable-egd when
4528 *Ben Kaduk and Rich Salz*
4530 * The distribution now has Makefile.in files, which are used to
4531 create Makefile's when Configure is run. *Configure must be run
4532 before trying to build now.*
4536 * The return value for SSL_CIPHER_description() for error conditions
4541 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4543 Obtaining and performing DNSSEC validation of TLSA records is
4544 the application's responsibility. The application provides
4545 the TLSA records of its choice to OpenSSL, and these are then
4546 used to authenticate the peer.
4548 The TLSA records need not even come from DNS. They can, for
4549 example, be used to implement local end-entity certificate or
4550 trust-anchor "pinning", where the "pin" data takes the form
4551 of TLSA records, which can augment or replace verification
4552 based on the usual WebPKI public certification authorities.
4556 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4557 continues to support deprecated interfaces in default builds.
4558 However, applications are strongly advised to compile their
4559 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4560 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4561 or the 1.1.0 releases.
4563 In environments in which all applications have been ported to
4564 not use any deprecated interfaces OpenSSL's Configure script
4565 should be used with the --api=1.1.0 option to entirely remove
4566 support for the deprecated features from the library and
4567 unconditionally disable them in the installed headers.
4568 Essentially the same effect can be achieved with the "no-deprecated"
4569 argument to Configure, except that this will always restrict
4570 the build to just the latest API, rather than a fixed API
4573 As applications are ported to future revisions of the API,
4574 they should update their compile-time OPENSSL_API_COMPAT define
4575 accordingly, but in most cases should be able to continue to
4576 compile with later releases.
4578 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4579 0x10000000L and 0x00908000L, respectively. However those
4580 versions did not support the OPENSSL_API_COMPAT feature, and
4581 so applications are not typically tested for explicit support
4582 of just the undeprecated features of either release.
4586 * Add support for setting the minimum and maximum supported protocol.
4587 It can bet set via the SSL_set_min_proto_version() and
4588 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4589 MaxProtocol. It's recommended to use the new APIs to disable
4590 protocols instead of disabling individual protocols using
4591 SSL_set_options() or SSL_CONF's Protocol. This change also
4592 removes support for disabling TLS 1.2 in the OpenSSL TLS
4593 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4597 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4601 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4602 and integrates ECDSA and ECDH functionality into EC. Implementations can
4603 now redirect key generation and no longer need to convert to or from
4606 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4607 include the ec.h header file instead.
4611 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4612 ciphers who are no longer supported and drops support the ephemeral RSA key
4613 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4617 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4618 opaque. For HMAC_CTX, the following constructors and destructors
4621 HMAC_CTX *HMAC_CTX_new(void);
4622 void HMAC_CTX_free(HMAC_CTX *ctx);
4624 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4625 destroy such methods has been added. See EVP_MD_meth_new(3) and
4626 EVP_CIPHER_meth_new(3) for documentation.
4629 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4630 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4631 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4632 an already created structure.
4633 2) For consistency with the majority of our object creators and
4634 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4635 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4636 for deprecated builds.
4640 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4641 cryptographic operations to be performed asynchronously as long as an
4642 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4643 further details. Libssl has also had this capability integrated with the
4644 introduction of the new mode SSL_MODE_ASYNC and associated error
4645 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4646 pages. This work was developed in partnership with Intel Corp.
4650 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4651 always enabled now. If you want to disable the support you should
4652 exclude it using the list of supported ciphers. This also means that the
4653 "-no_ecdhe" option has been removed from s_server.
4657 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4658 SSL_{CTX_}set1_curves() which can set a list.
4662 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4663 curve you want to support using SSL_{CTX_}set1_curves().
4667 * State machine rewrite. The state machine code has been significantly
4668 refactored in order to remove much duplication of code and solve issues
4669 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4670 further details). This change does have some associated API changes.
4671 Notably the SSL_state() function has been removed and replaced by
4672 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4673 SSL_set_state() has been removed altogether. The previous handshake states
4674 defined in ssl.h and ssl3.h have also been removed.
4678 * All instances of the string "ssleay" in the public API were replaced
4679 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4680 Some error codes related to internal RSA_eay API's were renamed.
4684 * The demo files in crypto/threads were moved to demo/threads.
4688 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4691 *Matt Caswell, Rich Salz*
4693 * New ASN.1 embed macro.
4695 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4696 structure is not allocated: it is part of the parent. That is instead of
4704 This reduces memory fragmentation and make it impossible to accidentally
4705 set a mandatory field to NULL.
4707 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4708 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4709 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4714 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4718 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4719 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4720 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4721 DES and RC4 ciphersuites.
4725 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4726 This changes the decoding behaviour for some invalid messages,
4727 though the change is mostly in the more lenient direction, and
4728 legacy behaviour is preserved as much as possible.
4732 * Fix no-stdio build.
4733 *David Woodhouse <David.Woodhouse@intel.com> and also*
4734 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4736 * New testing framework
4737 The testing framework has been largely rewritten and is now using
4738 perl and the perl modules Test::Harness and an extended variant of
4739 Test::More called OpenSSL::Test to do its work. All test scripts in
4740 test/ have been rewritten into test recipes, and all direct calls to
4741 executables in test/Makefile have become individual recipes using the
4742 simplified testing OpenSSL::Test::Simple.
4744 For documentation on our testing modules, do:
4746 perldoc test/testlib/OpenSSL/Test/Simple.pm
4747 perldoc test/testlib/OpenSSL/Test.pm
4751 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4752 are used; the latter aborts on memory leaks (usually checked on exit).
4753 Some undocumented "set malloc, etc., hooks" functions were removed
4754 and others were changed. All are now documented.
4758 * In DSA_generate_parameters_ex, if the provided seed is too short,
4761 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4763 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4764 from RFC4279, RFC4785, RFC5487, RFC5489.
4766 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4767 original RSA_PSK patch.
4771 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4772 era flag was never set throughout the codebase (only read). Also removed
4773 SSL3_FLAGS_POP_BUFFER which was only used if
4774 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4778 * Changed the default name options in the "ca", "crl", "req" and "x509"
4779 to be "oneline" instead of "compat".
4783 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4784 not aware of clients that still exhibit this bug, and the workaround
4785 hasn't been working properly for a while.
4789 * The return type of BIO_number_read() and BIO_number_written() as well as
4790 the corresponding num_read and num_write members in the BIO structure has
4791 changed from unsigned long to uint64_t. On platforms where an unsigned
4792 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4797 * Given the pervasive nature of TLS extensions it is inadvisable to run
4798 OpenSSL without support for them. It also means that maintaining
4799 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4800 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4804 * Removed support for the two export grade static DH ciphersuites
4805 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4806 were newly added (along with a number of other static DH ciphersuites) to
4807 1.0.2. However the two export ones have *never* worked since they were
4808 introduced. It seems strange in any case to be adding new export
4809 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4813 * Version negotiation has been rewritten. In particular SSLv23_method(),
4814 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4815 and turned into macros which simply call the new preferred function names
4816 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4817 should use the new names instead. Also as part of this change the ssl23.h
4818 header file has been removed.
4822 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4823 code and the associated standard is no longer considered fit-for-purpose.
4827 * RT2547 was closed. When generating a private key, try to make the
4828 output file readable only by the owner. This behavior change might
4829 be noticeable when interacting with other software.
4831 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4836 * Added HTTP GET support to the ocsp command.
4840 * Changed default digest for the dgst and enc commands from MD5 to
4845 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4849 * Added support for TLS extended master secret from
4850 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4851 initial patch which was a great help during development.
4855 * All libssl internal structures have been removed from the public header
4856 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4857 now redundant). Users should not attempt to access internal structures
4858 directly. Instead they should use the provided API functions.
4862 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4863 Access to deprecated functions can be re-enabled by running config with
4864 "enable-deprecated". In addition applications wishing to use deprecated
4865 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4866 will, by default, disable some transitive includes that previously existed
4867 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4871 * Added support for OCB mode. OpenSSL has been granted a patent license
4872 compatible with the OpenSSL license for use of OCB. Details are available
4873 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4874 for OCB can be removed by calling config with no-ocb.
4878 * SSLv2 support has been removed. It still supports receiving an SSLv2
4879 compatible client hello.
4883 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4884 done while fixing the error code for the key-too-small case.
4886 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4888 * CA.sh has been removed; use CA.pl instead.
4892 * Removed old DES API.
4896 * Remove various unsupported platforms:
4902 Sinix/ReliantUNIX RM400
4907 16-bit platforms such as WIN16
4911 * Clean up OPENSSL_NO_xxx #define's
4912 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4913 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4914 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4915 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4916 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4917 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4918 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4919 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4920 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4921 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4925 * Cleaned up dead code
4926 Remove all but one '#ifdef undef' which is to be looked at.
4930 * Clean up calling of xxx_free routines.
4931 Just like free(), fix most of the xxx_free routines to accept
4932 NULL. Remove the non-null checks from callers. Save much code.
4936 * Add secure heap for storage of private keys (when possible).
4937 Add BIO_s_secmem(), CBIGNUM, etc.
4938 Contributed by Akamai Technologies under our Corporate CLA.
4942 * Experimental support for a new, fast, unbiased prime candidate generator,
4943 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4945 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4947 * New output format NSS in the sess_id command line tool. This allows
4948 exporting the session id and the master key in NSS keylog format.
4950 *Martin Kaiser <martin@kaiser.cx>*
4952 * Harmonize version and its documentation. -f flag is used to display
4955 *mancha <mancha1@zoho.com>*
4957 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4958 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4960 *mancha <mancha1@zoho.com>*
4962 * Fix some double frees. These are not thought to be exploitable.
4964 *mancha <mancha1@zoho.com>*
4966 * A missing bounds check in the handling of the TLS heartbeat extension
4967 can be used to reveal up to 64k of memory to a connected client or
4970 Thanks for Neel Mehta of Google Security for discovering this bug and to
4971 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4972 preparing the fix ([CVE-2014-0160])
4974 *Adam Langley, Bodo Moeller*
4976 * Fix for the attack described in the paper "Recovering OpenSSL
4977 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4978 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4979 <http://eprint.iacr.org/2014/140>
4981 Thanks to Yuval Yarom and Naomi Benger for discovering this
4982 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4984 *Yuval Yarom and Naomi Benger*
4986 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4987 this fixes a limitation in previous versions of OpenSSL.
4991 * Experimental encrypt-then-mac support.
4993 Experimental support for encrypt then mac from
4994 draft-gutmann-tls-encrypt-then-mac-02.txt
4996 To enable it set the appropriate extension number (0x42 for the test
4997 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4999 For non-compliant peers (i.e. just about everything) this should have no
5002 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
5006 * Add EVP support for key wrapping algorithms, to avoid problems with
5007 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5008 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5009 algorithms and include tests cases.
5013 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
5018 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5019 MGF1 digest and OAEP label.
5023 * Make openssl verify return errors.
5025 *Chris Palmer <palmer@google.com> and Ben Laurie*
5027 * New function ASN1_TIME_diff to calculate the difference between two
5028 ASN1_TIME structures or one structure and the current time.
5032 * Update fips_test_suite to support multiple command line options. New
5033 test to induce all self test errors in sequence and check expected
5038 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5039 sign or verify all in one operation.
5043 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5044 test programs and fips_test_suite. Includes functionality to parse
5045 the minimal script output of fipsalgest.pl directly.
5049 * Add authorisation parameter to FIPS_module_mode_set().
5053 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5057 * Use separate DRBG fields for internal and external flags. New function
5058 FIPS_drbg_health_check() to perform on demand health checking. Add
5059 generation tests to fips_test_suite with reduced health check interval to
5060 demonstrate periodic health checking. Add "nodh" option to
5061 fips_test_suite to skip very slow DH test.
5065 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5070 * More extensive health check for DRBG checking many more failure modes.
5071 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5072 combination: call this in fips_test_suite.
5076 * Add support for canonical generation of DSA parameter 'g'. See
5079 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5080 POST to handle HMAC cases.
5084 * Add functions FIPS_module_version() and FIPS_module_version_text()
5085 to return numerical and string versions of the FIPS module number.
5089 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5090 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5091 outside the validated module in the FIPS capable OpenSSL.
5095 * Minor change to DRBG entropy callback semantics. In some cases
5096 there is no multiple of the block length between min_len and
5097 max_len. Allow the callback to return more than max_len bytes
5098 of entropy but discard any extra: it is the callback's responsibility
5099 to ensure that the extra data discarded does not impact the
5100 requested amount of entropy.
5104 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5105 information in FIPS186-3, SP800-57 and SP800-131A.
5109 * CCM support via EVP. Interface is very similar to GCM case except we
5110 must supply all data in one chunk (i.e. no update, final) and the
5111 message length must be supplied if AAD is used. Add algorithm test
5116 * Initial version of POST overhaul. Add POST callback to allow the status
5117 of POST to be monitored and/or failures induced. Modify fips_test_suite
5118 to use callback. Always run all selftests even if one fails.
5122 * XTS support including algorithm test driver in the fips_gcmtest program.
5123 Note: this does increase the maximum key length from 32 to 64 bytes but
5124 there should be no binary compatibility issues as existing applications
5125 will never use XTS mode.
5129 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5130 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5131 performs algorithm blocking for unapproved PRNG types. Also do not
5132 set PRNG type in FIPS_mode_set(): leave this to the application.
5133 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5134 the standard OpenSSL PRNG: set additional data to a date time vector.
5138 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5139 This shouldn't present any incompatibility problems because applications
5140 shouldn't be using these directly and any that are will need to rethink
5141 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5145 * Extensive self tests and health checking required by SP800-90 DRBG.
5146 Remove strength parameter from FIPS_drbg_instantiate and always
5147 instantiate at maximum supported strength.
5151 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5155 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5159 * New function DH_compute_key_padded() to compute a DH key and pad with
5160 leading zeroes if needed: this complies with SP800-56A et al.
5164 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5165 anything, incomplete, subject to change and largely untested at present.
5169 * Modify fipscanisteronly build option to only build the necessary object
5170 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5174 * Add experimental option FIPSSYMS to give all symbols in
5175 fipscanister.o and FIPS or fips prefix. This will avoid
5176 conflicts with future versions of OpenSSL. Add perl script
5177 util/fipsas.pl to preprocess assembly language source files
5178 and rename any affected symbols.
5182 * Add selftest checks and algorithm block of non-fips algorithms in
5183 FIPS mode. Remove DES2 from selftests.
5187 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5188 return internal method without any ENGINE dependencies. Add new
5189 tiny fips sign and verify functions.
5193 * New build option no-ec2m to disable characteristic 2 code.
5197 * New build option "fipscanisteronly". This only builds fipscanister.o
5198 and (currently) associated fips utilities. Uses the file Makefile.fips
5199 instead of Makefile.org as the prototype.
5203 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5204 Update fips_gcmtest to use IV generator.
5208 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5209 setting output buffer to NULL. The `*Final` function must be
5210 called although it will not retrieve any additional data. The tag
5211 can be set or retrieved with a ctrl. The IV length is by default 12
5212 bytes (96 bits) but can be set to an alternative value. If the IV
5213 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5218 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5219 underlying do_cipher function handles all cipher semantics itself
5220 including padding and finalisation. This is useful if (for example)
5221 an ENGINE cipher handles block padding itself. The behaviour of
5222 do_cipher is subtly changed if this flag is set: the return value
5223 is the number of characters written to the output buffer (zero is
5224 no longer an error code) or a negative error code. Also if the
5225 input buffer is NULL and length 0 finalisation should be performed.
5229 * If a candidate issuer certificate is already part of the constructed
5230 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5234 * Improve forward-security support: add functions
5236 void SSL_CTX_set_not_resumable_session_callback(
5237 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5238 void SSL_set_not_resumable_session_callback(
5239 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5241 for use by SSL/TLS servers; the callback function will be called whenever a
5242 new session is created, and gets to decide whether the session may be
5243 cached to make it resumable (return 0) or not (return 1). (As by the
5244 SSL/TLS protocol specifications, the session_id sent by the server will be
5245 empty to indicate that the session is not resumable; also, the server will
5246 not generate RFC 4507 (RFC 5077) session tickets.)
5248 A simple reasonable callback implementation is to return is_forward_secure.
5249 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5250 by the SSL/TLS server library, indicating whether it can provide forward
5253 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5255 * New -verify_name option in command line utilities to set verification
5260 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5261 Add CMAC pkey methods.
5265 * Experimental renegotiation in s_server -www mode. If the client
5266 browses /reneg connection is renegotiated. If /renegcert it is
5267 renegotiated requesting a certificate.
5271 * Add an "external" session cache for debugging purposes to s_server. This
5272 should help trace issues which normally are only apparent in deployed
5273 multi-process servers.
5277 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5278 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5279 BIO_set_cipher() and some obscure PEM functions were changed so they
5280 can now return an error. The RAND changes required a change to the
5281 RAND_METHOD structure.
5285 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5286 a gcc attribute to warn if the result of a function is ignored. This
5287 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5288 whose return value is often ignored.
5292 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5293 These allow SCTs (signed certificate timestamps) to be requested and
5294 validated when establishing a connection.
5296 *Rob Percival <robpercival@google.com>*
5301 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5303 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5304 used even when parsing explicit parameters, when loading a encoded key
5305 or calling `EC_GROUP_new_from_ecpkparameters()`/
5306 `EC_GROUP_new_from_ecparameters()`.
5307 This prevents bypass of security hardening and performance gains,
5308 especially for curves with specialized EC_METHODs.
5309 By default, if a key encoded with explicit parameters is loaded and later
5310 encoded, the output is still encoded with explicit parameters, even if
5311 internally a "named" EC_GROUP is used for computation.
5315 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5316 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5317 NULL. After this change, only the cofactor parameter can be NULL. It also
5318 does some minimal sanity checks on the passed order.
5323 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5324 An attack is simple, if the first CMS_recipientInfo is valid but the
5325 second CMS_recipientInfo is chosen ciphertext. If the second
5326 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5327 encryption key will be replaced by garbage, and the message cannot be
5328 decoded, but if the RSA decryption fails, the correct encryption key is
5329 used and the recipient will not notice the attack.
5330 As a work around for this potential attack the length of the decrypted
5331 key must be equal to the cipher default key length, in case the
5332 certificate is not given and all recipientInfo are tried out.
5333 The old behaviour can be re-enabled in the CMS code by setting the
5334 CMS_DEBUG_DECRYPT flag.
5339 * Document issue with installation paths in diverse Windows builds
5341 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5342 binaries and run-time config file.
5347 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5349 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5350 This changes the size when using the `genpkey` command when no size is given.
5351 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5352 generation commands to use 2048 bits by default.
5356 * Add FIPS support for Android Arm 64-bit
5358 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5359 Module in Version 2.0.10. For some reason, the corresponding target
5360 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5361 built with FIPS support on Android Arm 64-bit. This omission has been
5364 *Matthias St. Pierre*
5366 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5368 * 0-byte record padding oracle
5370 If an application encounters a fatal protocol error and then calls
5371 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5372 then OpenSSL can respond differently to the calling application if a 0 byte
5373 record is received with invalid padding compared to if a 0 byte record is
5374 received with an invalid MAC. If the application then behaves differently
5375 based on that in a way that is detectable to the remote peer, then this
5376 amounts to a padding oracle that could be used to decrypt data.
5378 In order for this to be exploitable "non-stitched" ciphersuites must be in
5379 use. Stitched ciphersuites are optimised implementations of certain
5380 commonly used ciphersuites. Also the application must call SSL_shutdown()
5381 twice even if a protocol error has occurred (applications should not do
5382 this but some do anyway).
5384 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5385 Aviram, with additional investigation by Steven Collison and Andrew
5386 Hourselt. It was reported to OpenSSL on 10th December 2018.
5391 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5395 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5397 * Microarchitecture timing vulnerability in ECC scalar multiplication
5399 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5400 shown to be vulnerable to a microarchitecture timing side channel attack.
5401 An attacker with sufficient access to mount local timing attacks during
5402 ECDSA signature generation could recover the private key.
5404 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5405 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5411 * Timing vulnerability in DSA signature generation
5413 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5414 timing side channel attack. An attacker could use variations in the signing
5415 algorithm to recover the private key.
5417 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5422 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5423 Module, accidentally introduced while backporting security fixes from the
5424 development branch and hindering the use of ECC in FIPS mode.
5428 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5430 * Client DoS due to large DH parameter
5432 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5433 malicious server can send a very large prime value to the client. This will
5434 cause the client to spend an unreasonably long period of time generating a
5435 key for this prime resulting in a hang until the client has finished. This
5436 could be exploited in a Denial Of Service attack.
5438 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5443 * Cache timing vulnerability in RSA Key Generation
5445 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5446 a cache timing side channel attack. An attacker with sufficient access to
5447 mount cache timing attacks during the RSA key generation process could
5448 recover the private key.
5450 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5451 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5456 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5457 parameter is no longer accepted, as it leads to a corrupt table. NULL
5458 pem_str is reserved for alias entries only.
5462 * Revert blinding in ECDSA sign and instead make problematic addition
5463 length-invariant. Switch even to fixed-length Montgomery multiplication.
5467 * Change generating and checking of primes so that the error rate of not
5468 being prime depends on the intended use based on the size of the input.
5469 For larger primes this will result in more rounds of Miller-Rabin.
5470 The maximal error rate for primes with more than 1080 bits is lowered
5473 *Kurt Roeckx, Annie Yousar*
5475 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5479 * Add blinding to ECDSA and DSA signatures to protect against side channel
5480 attacks discovered by Keegan Ryan (NCC Group).
5484 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5485 now allow empty (zero character) pass phrases.
5489 * Certificate time validation (X509_cmp_time) enforces stricter
5490 compliance with RFC 5280. Fractional seconds and timezone offsets
5491 are no longer allowed.
5495 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5497 * Constructed ASN.1 types with a recursive definition could exceed the stack
5499 Constructed ASN.1 types with a recursive definition (such as can be found
5500 in PKCS7) could eventually exceed the stack given malicious input with
5501 excessive recursion. This could result in a Denial Of Service attack. There
5502 are no such structures used within SSL/TLS that come from untrusted sources
5503 so this is considered safe.
5505 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5511 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5513 * Read/write after SSL object in error state
5515 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5516 mechanism. The intent was that if a fatal error occurred during a handshake
5517 then OpenSSL would move into the error state and would immediately fail if
5518 you attempted to continue the handshake. This works as designed for the
5519 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5520 SSL_connect()), however due to a bug it does not work correctly if
5521 SSL_read() or SSL_write() is called directly. In that scenario, if the
5522 handshake fails then a fatal error will be returned in the initial function
5523 call. If SSL_read()/SSL_write() is subsequently called by the application
5524 for the same SSL object then it will succeed and the data is passed without
5525 being decrypted/encrypted directly from the SSL/TLS record layer.
5527 In order to exploit this issue an application bug would have to be present
5528 that resulted in a call to SSL_read()/SSL_write() being issued after having
5529 already received a fatal error.
5531 This issue was reported to OpenSSL by David Benjamin (Google).
5536 * rsaz_1024_mul_avx2 overflow bug on x86_64
5538 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5539 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5540 Analysis suggests that attacks against RSA and DSA as a result of this
5541 defect would be very difficult to perform and are not believed likely.
5542 Attacks against DH1024 are considered just feasible, because most of the
5543 work necessary to deduce information about a private key may be performed
5544 offline. The amount of resources required for such an attack would be
5545 significant. However, for an attack on TLS to be meaningful, the server
5546 would have to share the DH1024 private key among multiple clients, which is
5547 no longer an option since CVE-2016-0701.
5549 This only affects processors that support the AVX2 but not ADX extensions
5550 like Intel Haswell (4th generation).
5552 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5553 was originally found via the OSS-Fuzz project.
5558 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5560 * bn_sqrx8x_internal carry bug on x86_64
5562 There is a carry propagating bug in the x86_64 Montgomery squaring
5563 procedure. No EC algorithms are affected. Analysis suggests that attacks
5564 against RSA and DSA as a result of this defect would be very difficult to
5565 perform and are not believed likely. Attacks against DH are considered just
5566 feasible (although very difficult) because most of the work necessary to
5567 deduce information about a private key may be performed offline. The amount
5568 of resources required for such an attack would be very significant and
5569 likely only accessible to a limited number of attackers. An attacker would
5570 additionally need online access to an unpatched system using the target
5571 private key in a scenario with persistent DH parameters and a private
5572 key that is shared between multiple clients.
5574 This only affects processors that support the BMI1, BMI2 and ADX extensions
5575 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5577 This issue was reported to OpenSSL by the OSS-Fuzz project.
5582 * Malformed X.509 IPAddressFamily could cause OOB read
5584 If an X.509 certificate has a malformed IPAddressFamily extension,
5585 OpenSSL could do a one-byte buffer overread. The most likely result
5586 would be an erroneous display of the certificate in text format.
5588 This issue was reported to OpenSSL by the OSS-Fuzz project.
5592 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5594 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5595 platform rather than 'mingw'.
5599 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5601 * Truncated packet could crash via OOB read
5603 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5604 cipher is being used, then a truncated packet can cause that host to
5605 perform an out-of-bounds read, usually resulting in a crash.
5607 This issue was reported to OpenSSL by Robert Święcki of Google.
5612 * BN_mod_exp may produce incorrect results on x86_64
5614 There is a carry propagating bug in the x86_64 Montgomery squaring
5615 procedure. No EC algorithms are affected. Analysis suggests that attacks
5616 against RSA and DSA as a result of this defect would be very difficult to
5617 perform and are not believed likely. Attacks against DH are considered just
5618 feasible (although very difficult) because most of the work necessary to
5619 deduce information about a private key may be performed offline. The amount
5620 of resources required for such an attack would be very significant and
5621 likely only accessible to a limited number of attackers. An attacker would
5622 additionally need online access to an unpatched system using the target
5623 private key in a scenario with persistent DH parameters and a private
5624 key that is shared between multiple clients. For example this can occur by
5625 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5626 similar to CVE-2015-3193 but must be treated as a separate problem.
5628 This issue was reported to OpenSSL by the OSS-Fuzz project.
5633 * Montgomery multiplication may produce incorrect results
5635 There is a carry propagating bug in the Broadwell-specific Montgomery
5636 multiplication procedure that handles input lengths divisible by, but
5637 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5638 and DH private keys are impossible. This is because the subroutine in
5639 question is not used in operations with the private key itself and an input
5640 of the attacker's direct choice. Otherwise the bug can manifest itself as
5641 transient authentication and key negotiation failures or reproducible
5642 erroneous outcome of public-key operations with specially crafted input.
5643 Among EC algorithms only Brainpool P-512 curves are affected and one
5644 presumably can attack ECDH key negotiation. Impact was not analyzed in
5645 detail, because pre-requisites for attack are considered unlikely. Namely
5646 multiple clients have to choose the curve in question and the server has to
5647 share the private key among them, neither of which is default behaviour.
5648 Even then only clients that chose the curve will be affected.
5650 This issue was publicly reported as transient failures and was not
5651 initially recognized as a security issue. Thanks to Richard Morgan for
5652 providing reproducible case.
5657 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5658 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5659 prevent issues where no progress is being made and the peer continually
5660 sends unrecognised record types, using up resources processing them.
5664 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5666 * Missing CRL sanity check
5668 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5669 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5670 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5672 This issue only affects the OpenSSL 1.0.2i
5677 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5679 * OCSP Status Request extension unbounded memory growth
5681 A malicious client can send an excessively large OCSP Status Request
5682 extension. If that client continually requests renegotiation, sending a
5683 large OCSP Status Request extension each time, then there will be unbounded
5684 memory growth on the server. This will eventually lead to a Denial Of
5685 Service attack through memory exhaustion. Servers with a default
5686 configuration are vulnerable even if they do not support OCSP. Builds using
5687 the "no-ocsp" build time option are not affected.
5689 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5694 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5697 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5703 * OOB write in MDC2_Update()
5705 An overflow can occur in MDC2_Update() either if called directly or
5706 through the EVP_DigestUpdate() function using MDC2. If an attacker
5707 is able to supply very large amounts of input data after a previous
5708 call to EVP_EncryptUpdate() with a partial block then a length check
5709 can overflow resulting in a heap corruption.
5711 The amount of data needed is comparable to SIZE_MAX which is impractical
5714 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5719 * Malformed SHA512 ticket DoS
5721 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5722 DoS attack where a malformed ticket will result in an OOB read which will
5725 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5726 a custom server callback and ticket lookup mechanism.
5728 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5733 * OOB write in BN_bn2dec()
5735 The function BN_bn2dec() does not check the return value of BN_div_word().
5736 This can cause an OOB write if an application uses this function with an
5737 overly large BIGNUM. This could be a problem if an overly large certificate
5738 or CRL is printed out from an untrusted source. TLS is not affected because
5739 record limits will reject an oversized certificate before it is parsed.
5741 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5746 * OOB read in TS_OBJ_print_bio()
5748 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5749 the total length the OID text representation would use and not the amount
5750 of data written. This will result in OOB reads when large OIDs are
5753 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5758 * Pointer arithmetic undefined behaviour
5760 Avoid some undefined pointer arithmetic
5762 A common idiom in the codebase is to check limits in the following manner:
5765 Where "p" points to some malloc'd data of SIZE bytes and
5768 "len" here could be from some externally supplied data (e.g. from a TLS
5771 The rules of C pointer arithmetic are such that "p + len" is only well
5772 defined where len <= SIZE. Therefore the above idiom is actually
5773 undefined behaviour.
5775 For example this could cause problems if some malloc implementation
5776 provides an address for "p" such that "p + len" actually overflows for
5777 values of len that are too big and therefore p + len < limit.
5779 This issue was reported to OpenSSL by Guido Vranken
5784 * Constant time flag not preserved in DSA signing
5786 Operations in the DSA signing algorithm should run in constant time in
5787 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5788 implementation means that a non-constant time codepath is followed for
5789 certain operations. This has been demonstrated through a cache-timing
5790 attack to be sufficient for an attacker to recover the private DSA key.
5792 This issue was reported by César Pereida (Aalto University), Billy Brumley
5793 (Tampere University of Technology), and Yuval Yarom (The University of
5794 Adelaide and NICTA).
5799 * DTLS buffered message DoS
5801 In a DTLS connection where handshake messages are delivered out-of-order
5802 those messages that OpenSSL is not yet ready to process will be buffered
5803 for later use. Under certain circumstances, a flaw in the logic means that
5804 those messages do not get removed from the buffer even though the handshake
5805 has been completed. An attacker could force up to approx. 15 messages to
5806 remain in the buffer when they are no longer required. These messages will
5807 be cleared when the DTLS connection is closed. The default maximum size for
5808 a message is 100k. Therefore, the attacker could force an additional 1500k
5809 to be consumed per connection. By opening many simultaneous connections an
5810 attacker could cause a DoS attack through memory exhaustion.
5812 This issue was reported to OpenSSL by Quan Luo.
5817 * DTLS replay protection DoS
5819 A flaw in the DTLS replay attack protection mechanism means that records
5820 that arrive for future epochs update the replay protection "window" before
5821 the MAC for the record has been validated. This could be exploited by an
5822 attacker by sending a record for the next epoch (which does not have to
5823 decrypt or have a valid MAC), with a very large sequence number. This means
5824 that all subsequent legitimate packets are dropped causing a denial of
5825 service for a specific DTLS connection.
5827 This issue was reported to OpenSSL by the OCAP audit team.
5832 * Certificate message OOB reads
5834 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5835 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5836 theoretical DoS risk but this has not been observed in practice on common
5839 The messages affected are client certificate, client certificate request
5840 and server certificate. As a result the attack can only be performed
5841 against a client or a server which enables client authentication.
5843 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5848 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5850 * Prevent padding oracle in AES-NI CBC MAC check
5852 A MITM attacker can use a padding oracle attack to decrypt traffic
5853 when the connection uses an AES CBC cipher and the server support
5856 This issue was introduced as part of the fix for Lucky 13 padding
5857 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5858 constant time by making sure that always the same bytes are read and
5859 compared against either the MAC or padding bytes. But it no longer
5860 checked that there was enough data to have both the MAC and padding
5863 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5867 * Fix EVP_EncodeUpdate overflow
5869 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5870 Base64 encoding of binary data. If an attacker is able to supply very large
5871 amounts of input data then a length check can overflow resulting in a heap
5874 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5875 the `PEM_write_bio*` family of functions. These are mainly used within the
5876 OpenSSL command line applications, so any application which processes data
5877 from an untrusted source and outputs it as a PEM file should be considered
5878 vulnerable to this issue. User applications that call these APIs directly
5879 with large amounts of untrusted data may also be vulnerable.
5881 This issue was reported by Guido Vranken.
5886 * Fix EVP_EncryptUpdate overflow
5888 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5889 is able to supply very large amounts of input data after a previous call to
5890 EVP_EncryptUpdate() with a partial block then a length check can overflow
5891 resulting in a heap corruption. Following an analysis of all OpenSSL
5892 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5893 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5894 the first called function after an EVP_EncryptInit(), and therefore that
5895 specific call must be safe. The second form is where the length passed to
5896 EVP_EncryptUpdate() can be seen from the code to be some small value and
5897 therefore there is no possibility of an overflow. Since all instances are
5898 one of these two forms, it is believed that there can be no overflows in
5899 internal code due to this problem. It should be noted that
5900 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5901 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5902 of these calls have also been analysed too and it is believed there are no
5903 instances in internal usage where an overflow could occur.
5905 This issue was reported by Guido Vranken.
5910 * Prevent ASN.1 BIO excessive memory allocation
5912 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5913 a short invalid encoding can cause allocation of large amounts of memory
5914 potentially consuming excessive resources or exhausting memory.
5916 Any application parsing untrusted data through d2i BIO functions is
5917 affected. The memory based functions such as d2i_X509() are *not* affected.
5918 Since the memory based functions are used by the TLS library, TLS
5919 applications are not affected.
5921 This issue was reported by Brian Carpenter.
5928 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5929 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5930 in arbitrary stack data being returned in the buffer.
5932 This issue was reported by Guido Vranken.
5937 * Modify behavior of ALPN to invoke callback after SNI/servername
5938 callback, such that updates to the SSL_CTX affect ALPN.
5942 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5947 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5948 methods are enabled and ssl2 is disabled the methods return NULL.
5952 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5954 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5955 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5956 provide any "EXPORT" or "LOW" strength ciphers.
5960 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5961 is by default disabled at build-time. Builds that are not configured with
5962 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5963 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5964 will need to explicitly call either of:
5966 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5968 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5970 as appropriate. Even if either of those is used, or the application
5971 explicitly uses the version-specific SSLv2_method() or its client and
5972 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5973 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5974 ciphers, and SSLv2 56-bit DES are no longer available.
5979 * Fix a double-free in DSA code
5981 A double free bug was discovered when OpenSSL parses malformed DSA private
5982 keys and could lead to a DoS attack or memory corruption for applications
5983 that receive DSA private keys from untrusted sources. This scenario is
5986 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5992 * Disable SRP fake user seed to address a server memory leak.
5994 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5996 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5997 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5998 was changed to ignore the "fake user" SRP seed, even if the seed
6001 Users should use SRP_VBASE_get1_by_user instead. Note that in
6002 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6003 also that even though configuring the SRP seed attempts to hide
6004 invalid usernames by continuing the handshake with fake
6005 credentials, this behaviour is not constant time and no strong
6006 guarantees are made that the handshake is indistinguishable from
6007 that of a valid user.
6012 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6014 In the BN_hex2bn function the number of hex digits is calculated using an
6015 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6016 large values of `i` this can result in `bn_expand` not allocating any
6017 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6018 field as NULL leading to a subsequent NULL ptr deref. For very large values
6019 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6020 In this case memory is allocated to the internal BIGNUM data field, but it
6021 is insufficiently sized leading to heap corruption. A similar issue exists
6022 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6023 is ever called by user applications with very large untrusted hex/dec data.
6024 This is anticipated to be a rare occurrence.
6026 All OpenSSL internal usage of these functions use data that is not expected
6027 to be untrusted, e.g. config file data or application command line
6028 arguments. If user developed applications generate config file data based
6029 on untrusted data then it is possible that this could also lead to security
6030 consequences. This is also anticipated to be rare.
6032 This issue was reported to OpenSSL by Guido Vranken.
6037 * Fix memory issues in `BIO_*printf` functions
6039 The internal `fmtstr` function used in processing a "%s" format string in
6040 the `BIO_*printf` functions could overflow while calculating the length of a
6041 string and cause an OOB read when printing very long strings.
6043 Additionally the internal `doapr_outch` function can attempt to write to an
6044 OOB memory location (at an offset from the NULL pointer) in the event of a
6045 memory allocation failure. In 1.0.2 and below this could be caused where
6046 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6047 could be in processing a very long "%s" format string. Memory leaks can
6050 The first issue may mask the second issue dependent on compiler behaviour.
6051 These problems could enable attacks where large amounts of untrusted data
6052 is passed to the `BIO_*printf` functions. If applications use these functions
6053 in this way then they could be vulnerable. OpenSSL itself uses these
6054 functions when printing out human-readable dumps of ASN.1 data. Therefore
6055 applications that print this data could be vulnerable if the data is from
6056 untrusted sources. OpenSSL command line applications could also be
6057 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6058 as command line arguments.
6060 Libssl is not considered directly vulnerable. Additionally certificates etc
6061 received via remote connections via libssl are also unlikely to be able to
6062 trigger these issues because of message size limits enforced within libssl.
6064 This issue was reported to OpenSSL Guido Vranken.
6069 * Side channel attack on modular exponentiation
6071 A side-channel attack was found which makes use of cache-bank conflicts on
6072 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6073 of RSA keys. The ability to exploit this issue is limited as it relies on
6074 an attacker who has control of code in a thread running on the same
6075 hyper-threaded core as the victim thread which is performing decryptions.
6077 This issue was reported to OpenSSL by Yuval Yarom, The University of
6078 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6079 Nadia Heninger, University of Pennsylvania with more information at
6080 <http://cachebleed.info>.
6085 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6086 if no keysize is specified with default_bits. This fixes an
6087 omission in an earlier change that changed all RSA/DSA key generation
6088 commands to use 2048 bits by default.
6092 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6094 * DH small subgroups
6096 Historically OpenSSL only ever generated DH parameters based on "safe"
6097 primes. More recently (in version 1.0.2) support was provided for
6098 generating X9.42 style parameter files such as those required for RFC 5114
6099 support. The primes used in such files may not be "safe". Where an
6100 application is using DH configured with parameters based on primes that are
6101 not "safe" then an attacker could use this fact to find a peer's private
6102 DH exponent. This attack requires that the attacker complete multiple
6103 handshakes in which the peer uses the same private DH exponent. For example
6104 this could be used to discover a TLS server's private DH exponent if it's
6105 reusing the private DH exponent or it's using a static DH ciphersuite.
6107 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6108 TLS. It is not on by default. If the option is not set then the server
6109 reuses the same private DH exponent for the life of the server process and
6110 would be vulnerable to this attack. It is believed that many popular
6111 applications do set this option and would therefore not be at risk.
6113 The fix for this issue adds an additional check where a "q" parameter is
6114 available (as is the case in X9.42 based parameters). This detects the
6115 only known attack, and is the only possible defense for static DH
6116 ciphersuites. This could have some performance impact.
6118 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6119 default and cannot be disabled. This could have some performance impact.
6121 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6126 * SSLv2 doesn't block disabled ciphers
6128 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6129 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6130 been disabled, provided that the SSLv2 protocol was not also disabled via
6133 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6134 and Sebastian Schinzel.
6139 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6141 * BN_mod_exp may produce incorrect results on x86_64
6143 There is a carry propagating bug in the x86_64 Montgomery squaring
6144 procedure. No EC algorithms are affected. Analysis suggests that attacks
6145 against RSA and DSA as a result of this defect would be very difficult to
6146 perform and are not believed likely. Attacks against DH are considered just
6147 feasible (although very difficult) because most of the work necessary to
6148 deduce information about a private key may be performed offline. The amount
6149 of resources required for such an attack would be very significant and
6150 likely only accessible to a limited number of attackers. An attacker would
6151 additionally need online access to an unpatched system using the target
6152 private key in a scenario with persistent DH parameters and a private
6153 key that is shared between multiple clients. For example this can occur by
6154 default in OpenSSL DHE based SSL/TLS ciphersuites.
6156 This issue was reported to OpenSSL by Hanno Böck.
6161 * Certificate verify crash with missing PSS parameter
6163 The signature verification routines will crash with a NULL pointer
6164 dereference if presented with an ASN.1 signature using the RSA PSS
6165 algorithm and absent mask generation function parameter. Since these
6166 routines are used to verify certificate signature algorithms this can be
6167 used to crash any certificate verification operation and exploited in a
6168 DoS attack. Any application which performs certificate verification is
6169 vulnerable including OpenSSL clients and servers which enable client
6172 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6177 * X509_ATTRIBUTE memory leak
6179 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6180 memory. This structure is used by the PKCS#7 and CMS routines so any
6181 application which reads PKCS#7 or CMS data from untrusted sources is
6182 affected. SSL/TLS is not affected.
6184 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6190 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6191 This changes the decoding behaviour for some invalid messages,
6192 though the change is mostly in the more lenient direction, and
6193 legacy behaviour is preserved as much as possible.
6197 * In DSA_generate_parameters_ex, if the provided seed is too short,
6200 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6202 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6204 * Alternate chains certificate forgery
6206 During certificate verification, OpenSSL will attempt to find an
6207 alternative certificate chain if the first attempt to build such a chain
6208 fails. An error in the implementation of this logic can mean that an
6209 attacker could cause certain checks on untrusted certificates to be
6210 bypassed, such as the CA flag, enabling them to use a valid leaf
6211 certificate to act as a CA and "issue" an invalid certificate.
6213 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6218 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6220 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6221 incompatibility in the handling of HMAC. The previous ABI has now been
6226 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6228 * Malformed ECParameters causes infinite loop
6230 When processing an ECParameters structure OpenSSL enters an infinite loop
6231 if the curve specified is over a specially malformed binary polynomial
6234 This can be used to perform denial of service against any
6235 system which processes public keys, certificate requests or
6236 certificates. This includes TLS clients and TLS servers with
6237 client authentication enabled.
6239 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6244 * Exploitable out-of-bounds read in X509_cmp_time
6246 X509_cmp_time does not properly check the length of the ASN1_TIME
6247 string and can read a few bytes out of bounds. In addition,
6248 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6251 An attacker can use this to craft malformed certificates and CRLs of
6252 various sizes and potentially cause a segmentation fault, resulting in
6253 a DoS on applications that verify certificates or CRLs. TLS clients
6254 that verify CRLs are affected. TLS clients and servers with client
6255 authentication enabled may be affected if they use custom verification
6258 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6259 independently by Hanno Böck.
6264 * PKCS7 crash with missing EnvelopedContent
6266 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6267 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6268 with missing content and trigger a NULL pointer dereference on parsing.
6270 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6271 structures from untrusted sources are affected. OpenSSL clients and
6272 servers are not affected.
6274 This issue was reported to OpenSSL by Michal Zalewski (Google).
6279 * CMS verify infinite loop with unknown hash function
6281 When verifying a signedData message the CMS code can enter an infinite loop
6282 if presented with an unknown hash function OID. This can be used to perform
6283 denial of service against any system which verifies signedData messages using
6285 This issue was reported to OpenSSL by Johannes Bauer.
6290 * Race condition handling NewSessionTicket
6292 If a NewSessionTicket is received by a multi-threaded client when attempting to
6293 reuse a previous ticket then a race condition can occur potentially leading to
6294 a double free of the ticket data.
6299 * Only support 256-bit or stronger elliptic curves with the
6300 'ecdh_auto' setting (server) or by default (client). Of supported
6301 curves, prefer P-256 (both).
6305 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6307 * ClientHello sigalgs DoS fix
6309 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6310 invalid signature algorithms extension a NULL pointer dereference will
6311 occur. This can be exploited in a DoS attack against the server.
6313 This issue was was reported to OpenSSL by David Ramos of Stanford
6317 *Stephen Henson and Matt Caswell*
6319 * Multiblock corrupted pointer fix
6321 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6322 feature only applies on 64 bit x86 architecture platforms that support AES
6323 NI instructions. A defect in the implementation of "multiblock" can cause
6324 OpenSSL's internal write buffer to become incorrectly set to NULL when
6325 using non-blocking IO. Typically, when the user application is using a
6326 socket BIO for writing, this will only result in a failed connection.
6327 However if some other BIO is used then it is likely that a segmentation
6328 fault will be triggered, thus enabling a potential DoS attack.
6330 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6335 * Segmentation fault in DTLSv1_listen fix
6337 The DTLSv1_listen function is intended to be stateless and processes the
6338 initial ClientHello from many peers. It is common for user code to loop
6339 over the call to DTLSv1_listen until a valid ClientHello is received with
6340 an associated cookie. A defect in the implementation of DTLSv1_listen means
6341 that state is preserved in the SSL object from one invocation to the next
6342 that can lead to a segmentation fault. Errors processing the initial
6343 ClientHello can trigger this scenario. An example of such an error could be
6344 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6347 This issue was reported to OpenSSL by Per Allansson.
6352 * Segmentation fault in ASN1_TYPE_cmp fix
6354 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6355 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6356 certificate signature algorithm consistency this can be used to crash any
6357 certificate verification operation and exploited in a DoS attack. Any
6358 application which performs certificate verification is vulnerable including
6359 OpenSSL clients and servers which enable client authentication.
6364 * Segmentation fault for invalid PSS parameters fix
6366 The signature verification routines will crash with a NULL pointer
6367 dereference if presented with an ASN.1 signature using the RSA PSS
6368 algorithm and invalid parameters. Since these routines are used to verify
6369 certificate signature algorithms this can be used to crash any
6370 certificate verification operation and exploited in a DoS attack. Any
6371 application which performs certificate verification is vulnerable including
6372 OpenSSL clients and servers which enable client authentication.
6374 This issue was was reported to OpenSSL by Brian Carpenter.
6379 * ASN.1 structure reuse memory corruption fix
6381 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6382 memory corruption via an invalid write. Such reuse is and has been
6383 strongly discouraged and is believed to be rare.
6385 Applications that parse structures containing CHOICE or ANY DEFINED BY
6386 components may be affected. Certificate parsing (d2i_X509 and related
6387 functions) are however not affected. OpenSSL clients and servers are
6393 * PKCS7 NULL pointer dereferences fix
6395 The PKCS#7 parsing code does not handle missing outer ContentInfo
6396 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6397 missing content and trigger a NULL pointer dereference on parsing.
6399 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6400 otherwise parse PKCS#7 structures from untrusted sources are
6401 affected. OpenSSL clients and servers are not affected.
6403 This issue was reported to OpenSSL by Michal Zalewski (Google).
6408 * DoS via reachable assert in SSLv2 servers fix
6410 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6411 servers that both support SSLv2 and enable export cipher suites by sending
6412 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6414 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6415 (OpenSSL development team).
6420 * Empty CKE with client auth and DHE fix
6422 If client auth is used then a server can seg fault in the event of a DHE
6423 ciphersuite being selected and a zero length ClientKeyExchange message
6424 being sent by the client. This could be exploited in a DoS attack.
6429 * Handshake with unseeded PRNG fix
6431 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6432 with an unseeded PRNG. The conditions are:
6433 - The client is on a platform where the PRNG has not been seeded
6434 automatically, and the user has not seeded manually
6435 - A protocol specific client method version has been used (i.e. not
6436 SSL_client_methodv23)
6437 - A ciphersuite is used that does not require additional random data from
6438 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6440 If the handshake succeeds then the client random that has been used will
6441 have been generated from a PRNG with insufficient entropy and therefore the
6442 output may be predictable.
6444 For example using the following command with an unseeded openssl will
6445 succeed on an unpatched platform:
6447 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6452 * Use After Free following d2i_ECPrivatekey error fix
6454 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6455 could cause a use after free condition. This, in turn, could cause a double
6456 free in several private key parsing functions (such as d2i_PrivateKey
6457 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6458 for applications that receive EC private keys from untrusted
6459 sources. This scenario is considered rare.
6461 This issue was discovered by the BoringSSL project and fixed in their
6467 * X509_to_X509_REQ NULL pointer deref fix
6469 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6470 the certificate key is invalid. This function is rarely used in practice.
6472 This issue was discovered by Brian Carpenter.
6477 * Removed the export ciphers from the DEFAULT ciphers
6481 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6483 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6484 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6485 So far those who have to target multiple platforms would compromise
6486 and argue that binary targeting say ARMv5 would still execute on
6487 ARMv8. "Universal" build resolves this compromise by providing
6488 near-optimal performance even on newer platforms.
6492 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6493 (other platforms pending).
6495 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6497 * Add support for the SignedCertificateTimestampList certificate and
6498 OCSP response extensions from RFC6962.
6502 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6503 for corner cases. (Certain input points at infinity could lead to
6504 bogus results, with non-infinity inputs mapped to infinity too.)
6508 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6509 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6510 common cases are optimized and there still is room for further
6511 improvements. Vector Permutation AES for Altivec is also added.
6515 * Add support for little-endian ppc64 Linux target.
6517 *Marcelo Cerri (IBM)*
6519 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6520 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6521 are optimized and there still is room for further improvements.
6522 Both 32- and 64-bit modes are supported.
6524 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6526 * Improved ARMv7 NEON support.
6530 * Support for SPARC Architecture 2011 crypto extensions, first
6531 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6532 SHA256/512, MD5, GHASH and modular exponentiation.
6534 *Andy Polyakov, David Miller*
6536 * Accelerated modular exponentiation for Intel processors, a.k.a.
6539 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6541 * Support for new and upcoming Intel processors, including AVX2,
6542 BMI and SHA ISA extensions. This includes additional "stitched"
6543 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6546 This work was sponsored by Intel Corp.
6550 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6551 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6552 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6556 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6557 this fixes a limitation in previous versions of OpenSSL.
6561 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6562 MGF1 digest and OAEP label.
6566 * Add EVP support for key wrapping algorithms, to avoid problems with
6567 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6568 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6569 algorithms and include tests cases.
6573 * Add functions to allocate and set the fields of an ECDSA_METHOD
6576 *Douglas E. Engert, Steve Henson*
6578 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6579 difference in days and seconds between two tm or ASN1_TIME structures.
6583 * Add -rev test option to s_server to just reverse order of characters
6584 received by client and send back to server. Also prints an abbreviated
6585 summary of the connection parameters.
6589 * New option -brief for s_client and s_server to print out a brief summary
6590 of connection parameters.
6594 * Add callbacks for arbitrary TLS extensions.
6596 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6598 * New option -crl_download in several openssl utilities to download CRLs
6599 from CRLDP extension in certificates.
6603 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6607 * New function X509_CRL_diff to generate a delta CRL from the difference
6608 of two full CRLs. Add support to "crl" utility.
6612 * New functions to set lookup_crls function and to retrieve
6613 X509_STORE from X509_STORE_CTX.
6617 * Print out deprecated issuer and subject unique ID fields in
6622 * Extend OCSP I/O functions so they can be used for simple general purpose
6623 HTTP as well as OCSP. New wrapper function which can be used to download
6624 CRLs using the OCSP API.
6628 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6632 * `SSL_CONF*` functions. These provide a common framework for application
6633 configuration using configuration files or command lines.
6637 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6638 message callback and prints the results. Needs compile time option
6639 "enable-ssl-trace". New options to s_client and s_server to enable
6644 * New ctrl and macro to retrieve supported points extensions.
6645 Print out extension in s_server and s_client.
6649 * New functions to retrieve certificate signature and signature
6654 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6659 * New Suite B modes for TLS code. These use and enforce the requirements
6660 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6661 only use Suite B curves. The Suite B modes can be set by using the
6662 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6666 * New chain verification flags for Suite B levels of security. Check
6667 algorithms are acceptable when flags are set in X509_verify_cert.
6671 * Make tls1_check_chain return a set of flags indicating checks passed
6672 by a certificate chain. Add additional tests to handle client
6673 certificates: checks for matching certificate type and issuer name
6678 * If an attempt is made to use a signature algorithm not in the peer
6679 preference list abort the handshake. If client has no suitable
6680 signature algorithms in response to a certificate request do not
6681 use the certificate.
6685 * If server EC tmp key is not in client preference list abort handshake.
6689 * Add support for certificate stores in CERT structure. This makes it
6690 possible to have different stores per SSL structure or one store in
6691 the parent SSL_CTX. Include distinct stores for certificate chain
6692 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6693 to build and store a certificate chain in CERT structure: returning
6694 an error if the chain cannot be built: this will allow applications
6695 to test if a chain is correctly configured.
6697 Note: if the CERT based stores are not set then the parent SSL_CTX
6698 store is used to retain compatibility with existing behaviour.
6702 * New function ssl_set_client_disabled to set a ciphersuite disabled
6703 mask based on the current session, check mask when sending client
6704 hello and checking the requested ciphersuite.
6708 * New ctrls to retrieve and set certificate types in a certificate
6709 request message. Print out received values in s_client. If certificate
6710 types is not set with custom values set sensible values based on
6711 supported signature algorithms.
6715 * Support for distinct client and server supported signature algorithms.
6719 * Add certificate callback. If set this is called whenever a certificate
6720 is required by client or server. An application can decide which
6721 certificate chain to present based on arbitrary criteria: for example
6722 supported signature algorithms. Add very simple example to s_server.
6723 This fixes many of the problems and restrictions of the existing client
6724 certificate callback: for example you can now clear an existing
6725 certificate and specify the whole chain.
6729 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6730 the certificate can be used for (if anything). Set valid_flags field
6731 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6732 to have similar checks in it.
6734 Add new "cert_flags" field to CERT structure and include a "strict mode".
6735 This enforces some TLS certificate requirements (such as only permitting
6736 certificate signature algorithms contained in the supported algorithms
6737 extension) which some implementations ignore: this option should be used
6738 with caution as it could cause interoperability issues.
6742 * Update and tidy signature algorithm extension processing. Work out
6743 shared signature algorithms based on preferences and peer algorithms
6744 and print them out in s_client and s_server. Abort handshake if no
6745 shared signature algorithms.
6749 * Add new functions to allow customised supported signature algorithms
6750 for SSL and SSL_CTX structures. Add options to s_client and s_server
6755 * New function SSL_certs_clear() to delete all references to certificates
6756 from an SSL structure. Before this once a certificate had been added
6757 it couldn't be removed.
6761 * Integrate hostname, email address and IP address checking with certificate
6762 verification. New verify options supporting checking in openssl utility.
6766 * Fixes and wildcard matching support to hostname and email checking
6767 functions. Add manual page.
6769 *Florian Weimer (Red Hat Product Security Team)*
6771 * New functions to check a hostname email or IP address against a
6772 certificate. Add options x509 utility to print results of checks against
6777 * Fix OCSP checking.
6779 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6781 * Initial experimental support for explicitly trusted non-root CAs.
6782 OpenSSL still tries to build a complete chain to a root but if an
6783 intermediate CA has a trust setting included that is used. The first
6784 setting is used: whether to trust (e.g., -addtrust option to the x509
6789 * Add -trusted_first option which attempts to find certificates in the
6790 trusted store even if an untrusted chain is also supplied.
6794 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6795 platform support for Linux and Android.
6799 * Support for linux-x32, ILP32 environment in x86_64 framework.
6803 * Experimental multi-implementation support for FIPS capable OpenSSL.
6804 When in FIPS mode the approved implementations are used as normal,
6805 when not in FIPS mode the internal unapproved versions are used instead.
6806 This means that the FIPS capable OpenSSL isn't forced to use the
6807 (often lower performance) FIPS implementations outside FIPS mode.
6811 * Transparently support X9.42 DH parameters when calling
6812 PEM_read_bio_DHparameters. This means existing applications can handle
6813 the new parameter format automatically.
6817 * Initial experimental support for X9.42 DH parameter format: mainly
6818 to support use of 'q' parameter for RFC5114 parameters.
6822 * Add DH parameters from RFC5114 including test data to dhtest.
6826 * Support for automatic EC temporary key parameter selection. If enabled
6827 the most preferred EC parameters are automatically used instead of
6828 hardcoded fixed parameters. Now a server just has to call:
6829 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6830 support ECDH and use the most appropriate parameters.
6834 * Enhance and tidy EC curve and point format TLS extension code. Use
6835 static structures instead of allocation if default values are used.
6836 New ctrls to set curves we wish to support and to retrieve shared curves.
6837 Print out shared curves in s_server. New options to s_server and s_client
6838 to set list of supported curves.
6842 * New ctrls to retrieve supported signature algorithms and
6843 supported curve values as an array of NIDs. Extend openssl utility
6844 to print out received values.
6848 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6849 between NIDs and the more common NIST names such as "P-256". Enhance
6850 ecparam utility and ECC method to recognise the NIST names for curves.
6854 * Enhance SSL/TLS certificate chain handling to support different
6855 chains for each certificate instead of one chain in the parent SSL_CTX.
6859 * Support for fixed DH ciphersuite client authentication: where both
6860 server and client use DH certificates with common parameters.
6864 * Support for fixed DH ciphersuites: those requiring DH server
6869 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6871 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6872 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6873 X509_CINF_get_signature were reverted post internal team review.
6878 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6880 * OCSP Status Request extension unbounded memory growth
6882 A malicious client can send an excessively large OCSP Status Request
6883 extension. If that client continually requests renegotiation, sending a
6884 large OCSP Status Request extension each time, then there will be unbounded
6885 memory growth on the server. This will eventually lead to a Denial Of
6886 Service attack through memory exhaustion. Servers with a default
6887 configuration are vulnerable even if they do not support OCSP. Builds using
6888 the "no-ocsp" build time option are not affected.
6890 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6895 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6898 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6904 * OOB write in MDC2_Update()
6906 An overflow can occur in MDC2_Update() either if called directly or
6907 through the EVP_DigestUpdate() function using MDC2. If an attacker
6908 is able to supply very large amounts of input data after a previous
6909 call to EVP_EncryptUpdate() with a partial block then a length check
6910 can overflow resulting in a heap corruption.
6912 The amount of data needed is comparable to SIZE_MAX which is impractical
6915 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6920 * Malformed SHA512 ticket DoS
6922 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6923 DoS attack where a malformed ticket will result in an OOB read which will
6926 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6927 a custom server callback and ticket lookup mechanism.
6929 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6934 * OOB write in BN_bn2dec()
6936 The function BN_bn2dec() does not check the return value of BN_div_word().
6937 This can cause an OOB write if an application uses this function with an
6938 overly large BIGNUM. This could be a problem if an overly large certificate
6939 or CRL is printed out from an untrusted source. TLS is not affected because
6940 record limits will reject an oversized certificate before it is parsed.
6942 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6947 * OOB read in TS_OBJ_print_bio()
6949 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6950 the total length the OID text representation would use and not the amount
6951 of data written. This will result in OOB reads when large OIDs are
6954 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6959 * Pointer arithmetic undefined behaviour
6961 Avoid some undefined pointer arithmetic
6963 A common idiom in the codebase is to check limits in the following manner:
6966 Where "p" points to some malloc'd data of SIZE bytes and
6969 "len" here could be from some externally supplied data (e.g. from a TLS
6972 The rules of C pointer arithmetic are such that "p + len" is only well
6973 defined where len <= SIZE. Therefore, the above idiom is actually
6974 undefined behaviour.
6976 For example this could cause problems if some malloc implementation
6977 provides an address for "p" such that "p + len" actually overflows for
6978 values of len that are too big and therefore p + len < limit.
6980 This issue was reported to OpenSSL by Guido Vranken
6985 * Constant time flag not preserved in DSA signing
6987 Operations in the DSA signing algorithm should run in constant time in
6988 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6989 implementation means that a non-constant time codepath is followed for
6990 certain operations. This has been demonstrated through a cache-timing
6991 attack to be sufficient for an attacker to recover the private DSA key.
6993 This issue was reported by César Pereida (Aalto University), Billy Brumley
6994 (Tampere University of Technology), and Yuval Yarom (The University of
6995 Adelaide and NICTA).
7000 * DTLS buffered message DoS
7002 In a DTLS connection where handshake messages are delivered out-of-order
7003 those messages that OpenSSL is not yet ready to process will be buffered
7004 for later use. Under certain circumstances, a flaw in the logic means that
7005 those messages do not get removed from the buffer even though the handshake
7006 has been completed. An attacker could force up to approx. 15 messages to
7007 remain in the buffer when they are no longer required. These messages will
7008 be cleared when the DTLS connection is closed. The default maximum size for
7009 a message is 100k. Therefore, the attacker could force an additional 1500k
7010 to be consumed per connection. By opening many simultaneous connections an
7011 attacker could cause a DoS attack through memory exhaustion.
7013 This issue was reported to OpenSSL by Quan Luo.
7018 * DTLS replay protection DoS
7020 A flaw in the DTLS replay attack protection mechanism means that records
7021 that arrive for future epochs update the replay protection "window" before
7022 the MAC for the record has been validated. This could be exploited by an
7023 attacker by sending a record for the next epoch (which does not have to
7024 decrypt or have a valid MAC), with a very large sequence number. This means
7025 that all subsequent legitimate packets are dropped causing a denial of
7026 service for a specific DTLS connection.
7028 This issue was reported to OpenSSL by the OCAP audit team.
7033 * Certificate message OOB reads
7035 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7036 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7037 theoretical DoS risk but this has not been observed in practice on common
7040 The messages affected are client certificate, client certificate request
7041 and server certificate. As a result the attack can only be performed
7042 against a client or a server which enables client authentication.
7044 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7049 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7051 * Prevent padding oracle in AES-NI CBC MAC check
7053 A MITM attacker can use a padding oracle attack to decrypt traffic
7054 when the connection uses an AES CBC cipher and the server support
7057 This issue was introduced as part of the fix for Lucky 13 padding
7058 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7059 constant time by making sure that always the same bytes are read and
7060 compared against either the MAC or padding bytes. But it no longer
7061 checked that there was enough data to have both the MAC and padding
7064 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7069 * Fix EVP_EncodeUpdate overflow
7071 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7072 Base64 encoding of binary data. If an attacker is able to supply very large
7073 amounts of input data then a length check can overflow resulting in a heap
7076 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7077 the `PEM_write_bio*` family of functions. These are mainly used within the
7078 OpenSSL command line applications, so any application which processes data
7079 from an untrusted source and outputs it as a PEM file should be considered
7080 vulnerable to this issue. User applications that call these APIs directly
7081 with large amounts of untrusted data may also be vulnerable.
7083 This issue was reported by Guido Vranken.
7088 * Fix EVP_EncryptUpdate overflow
7090 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7091 is able to supply very large amounts of input data after a previous call to
7092 EVP_EncryptUpdate() with a partial block then a length check can overflow
7093 resulting in a heap corruption. Following an analysis of all OpenSSL
7094 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7095 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7096 the first called function after an EVP_EncryptInit(), and therefore that
7097 specific call must be safe. The second form is where the length passed to
7098 EVP_EncryptUpdate() can be seen from the code to be some small value and
7099 therefore there is no possibility of an overflow. Since all instances are
7100 one of these two forms, it is believed that there can be no overflows in
7101 internal code due to this problem. It should be noted that
7102 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7103 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7104 of these calls have also been analysed too and it is believed there are no
7105 instances in internal usage where an overflow could occur.
7107 This issue was reported by Guido Vranken.
7112 * Prevent ASN.1 BIO excessive memory allocation
7114 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7115 a short invalid encoding can casuse allocation of large amounts of memory
7116 potentially consuming excessive resources or exhausting memory.
7118 Any application parsing untrusted data through d2i BIO functions is
7119 affected. The memory based functions such as d2i_X509() are *not* affected.
7120 Since the memory based functions are used by the TLS library, TLS
7121 applications are not affected.
7123 This issue was reported by Brian Carpenter.
7130 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7131 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7132 in arbitrary stack data being returned in the buffer.
7134 This issue was reported by Guido Vranken.
7139 * Modify behavior of ALPN to invoke callback after SNI/servername
7140 callback, such that updates to the SSL_CTX affect ALPN.
7144 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7149 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7150 methods are enabled and ssl2 is disabled the methods return NULL.
7154 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7156 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7157 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7158 provide any "EXPORT" or "LOW" strength ciphers.
7162 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7163 is by default disabled at build-time. Builds that are not configured with
7164 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7165 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7166 will need to explicitly call either of:
7168 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7170 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7172 as appropriate. Even if either of those is used, or the application
7173 explicitly uses the version-specific SSLv2_method() or its client and
7174 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7175 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7176 ciphers, and SSLv2 56-bit DES are no longer available.
7181 * Fix a double-free in DSA code
7183 A double free bug was discovered when OpenSSL parses malformed DSA private
7184 keys and could lead to a DoS attack or memory corruption for applications
7185 that receive DSA private keys from untrusted sources. This scenario is
7188 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7194 * Disable SRP fake user seed to address a server memory leak.
7196 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7198 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7199 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7200 was changed to ignore the "fake user" SRP seed, even if the seed
7203 Users should use SRP_VBASE_get1_by_user instead. Note that in
7204 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7205 also that even though configuring the SRP seed attempts to hide
7206 invalid usernames by continuing the handshake with fake
7207 credentials, this behaviour is not constant time and no strong
7208 guarantees are made that the handshake is indistinguishable from
7209 that of a valid user.
7214 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7216 In the BN_hex2bn function the number of hex digits is calculated using an
7217 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7218 large values of `i` this can result in `bn_expand` not allocating any
7219 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7220 field as NULL leading to a subsequent NULL ptr deref. For very large values
7221 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7222 In this case memory is allocated to the internal BIGNUM data field, but it
7223 is insufficiently sized leading to heap corruption. A similar issue exists
7224 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7225 is ever called by user applications with very large untrusted hex/dec data.
7226 This is anticipated to be a rare occurrence.
7228 All OpenSSL internal usage of these functions use data that is not expected
7229 to be untrusted, e.g. config file data or application command line
7230 arguments. If user developed applications generate config file data based
7231 on untrusted data then it is possible that this could also lead to security
7232 consequences. This is also anticipated to be rare.
7234 This issue was reported to OpenSSL by Guido Vranken.
7239 * Fix memory issues in `BIO_*printf` functions
7241 The internal `fmtstr` function used in processing a "%s" format string in
7242 the `BIO_*printf` functions could overflow while calculating the length of a
7243 string and cause an OOB read when printing very long strings.
7245 Additionally the internal `doapr_outch` function can attempt to write to an
7246 OOB memory location (at an offset from the NULL pointer) in the event of a
7247 memory allocation failure. In 1.0.2 and below this could be caused where
7248 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7249 could be in processing a very long "%s" format string. Memory leaks can
7252 The first issue may mask the second issue dependent on compiler behaviour.
7253 These problems could enable attacks where large amounts of untrusted data
7254 is passed to the `BIO_*printf` functions. If applications use these functions
7255 in this way then they could be vulnerable. OpenSSL itself uses these
7256 functions when printing out human-readable dumps of ASN.1 data. Therefore
7257 applications that print this data could be vulnerable if the data is from
7258 untrusted sources. OpenSSL command line applications could also be
7259 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7260 as command line arguments.
7262 Libssl is not considered directly vulnerable. Additionally certificates etc
7263 received via remote connections via libssl are also unlikely to be able to
7264 trigger these issues because of message size limits enforced within libssl.
7266 This issue was reported to OpenSSL Guido Vranken.
7271 * Side channel attack on modular exponentiation
7273 A side-channel attack was found which makes use of cache-bank conflicts on
7274 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7275 of RSA keys. The ability to exploit this issue is limited as it relies on
7276 an attacker who has control of code in a thread running on the same
7277 hyper-threaded core as the victim thread which is performing decryptions.
7279 This issue was reported to OpenSSL by Yuval Yarom, The University of
7280 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7281 Nadia Heninger, University of Pennsylvania with more information at
7282 <http://cachebleed.info>.
7287 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7288 if no keysize is specified with default_bits. This fixes an
7289 omission in an earlier change that changed all RSA/DSA key generation
7290 commands to use 2048 bits by default.
7294 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7296 * Protection for DH small subgroup attacks
7298 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7299 switched on by default and cannot be disabled. This could have some
7304 * SSLv2 doesn't block disabled ciphers
7306 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7307 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7308 been disabled, provided that the SSLv2 protocol was not also disabled via
7311 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7312 and Sebastian Schinzel.
7317 * Reject DH handshakes with parameters shorter than 1024 bits.
7321 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7323 * Certificate verify crash with missing PSS parameter
7325 The signature verification routines will crash with a NULL pointer
7326 dereference if presented with an ASN.1 signature using the RSA PSS
7327 algorithm and absent mask generation function parameter. Since these
7328 routines are used to verify certificate signature algorithms this can be
7329 used to crash any certificate verification operation and exploited in a
7330 DoS attack. Any application which performs certificate verification is
7331 vulnerable including OpenSSL clients and servers which enable client
7334 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7339 * X509_ATTRIBUTE memory leak
7341 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7342 memory. This structure is used by the PKCS#7 and CMS routines so any
7343 application which reads PKCS#7 or CMS data from untrusted sources is
7344 affected. SSL/TLS is not affected.
7346 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7352 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7353 This changes the decoding behaviour for some invalid messages,
7354 though the change is mostly in the more lenient direction, and
7355 legacy behaviour is preserved as much as possible.
7359 * In DSA_generate_parameters_ex, if the provided seed is too short,
7360 use a random seed, as already documented.
7362 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7364 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7366 * Alternate chains certificate forgery
7368 During certificate verfification, OpenSSL will attempt to find an
7369 alternative certificate chain if the first attempt to build such a chain
7370 fails. An error in the implementation of this logic can mean that an
7371 attacker could cause certain checks on untrusted certificates to be
7372 bypassed, such as the CA flag, enabling them to use a valid leaf
7373 certificate to act as a CA and "issue" an invalid certificate.
7375 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7381 * Race condition handling PSK identify hint
7383 If PSK identity hints are received by a multi-threaded client then
7384 the values are wrongly updated in the parent SSL_CTX structure. This can
7385 result in a race condition potentially leading to a double free of the
7391 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7393 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7394 incompatibility in the handling of HMAC. The previous ABI has now been
7397 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7399 * Malformed ECParameters causes infinite loop
7401 When processing an ECParameters structure OpenSSL enters an infinite loop
7402 if the curve specified is over a specially malformed binary polynomial
7405 This can be used to perform denial of service against any
7406 system which processes public keys, certificate requests or
7407 certificates. This includes TLS clients and TLS servers with
7408 client authentication enabled.
7410 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7415 * Exploitable out-of-bounds read in X509_cmp_time
7417 X509_cmp_time does not properly check the length of the ASN1_TIME
7418 string and can read a few bytes out of bounds. In addition,
7419 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7422 An attacker can use this to craft malformed certificates and CRLs of
7423 various sizes and potentially cause a segmentation fault, resulting in
7424 a DoS on applications that verify certificates or CRLs. TLS clients
7425 that verify CRLs are affected. TLS clients and servers with client
7426 authentication enabled may be affected if they use custom verification
7429 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7430 independently by Hanno Böck.
7435 * PKCS7 crash with missing EnvelopedContent
7437 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7438 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7439 with missing content and trigger a NULL pointer dereference on parsing.
7441 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7442 structures from untrusted sources are affected. OpenSSL clients and
7443 servers are not affected.
7445 This issue was reported to OpenSSL by Michal Zalewski (Google).
7450 * CMS verify infinite loop with unknown hash function
7452 When verifying a signedData message the CMS code can enter an infinite loop
7453 if presented with an unknown hash function OID. This can be used to perform
7454 denial of service against any system which verifies signedData messages using
7456 This issue was reported to OpenSSL by Johannes Bauer.
7461 * Race condition handling NewSessionTicket
7463 If a NewSessionTicket is received by a multi-threaded client when attempting to
7464 reuse a previous ticket then a race condition can occur potentially leading to
7465 a double free of the ticket data.
7470 * Reject DH handshakes with parameters shorter than 768 bits.
7472 *Kurt Roeckx and Emilia Kasper*
7474 * dhparam: generate 2048-bit parameters by default.
7476 *Kurt Roeckx and Emilia Kasper*
7478 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7480 * Segmentation fault in ASN1_TYPE_cmp fix
7482 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7483 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7484 certificate signature algorithm consistency this can be used to crash any
7485 certificate verification operation and exploited in a DoS attack. Any
7486 application which performs certificate verification is vulnerable including
7487 OpenSSL clients and servers which enable client authentication.
7492 * ASN.1 structure reuse memory corruption fix
7494 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7495 memory corruption via an invalid write. Such reuse is and has been
7496 strongly discouraged and is believed to be rare.
7498 Applications that parse structures containing CHOICE or ANY DEFINED BY
7499 components may be affected. Certificate parsing (d2i_X509 and related
7500 functions) are however not affected. OpenSSL clients and servers are
7506 * PKCS7 NULL pointer dereferences fix
7508 The PKCS#7 parsing code does not handle missing outer ContentInfo
7509 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7510 missing content and trigger a NULL pointer dereference on parsing.
7512 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7513 otherwise parse PKCS#7 structures from untrusted sources are
7514 affected. OpenSSL clients and servers are not affected.
7516 This issue was reported to OpenSSL by Michal Zalewski (Google).
7521 * DoS via reachable assert in SSLv2 servers fix
7523 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7524 servers that both support SSLv2 and enable export cipher suites by sending
7525 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7527 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7528 (OpenSSL development team).
7533 * Use After Free following d2i_ECPrivatekey error fix
7535 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7536 could cause a use after free condition. This, in turn, could cause a double
7537 free in several private key parsing functions (such as d2i_PrivateKey
7538 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7539 for applications that receive EC private keys from untrusted
7540 sources. This scenario is considered rare.
7542 This issue was discovered by the BoringSSL project and fixed in their
7548 * X509_to_X509_REQ NULL pointer deref fix
7550 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7551 the certificate key is invalid. This function is rarely used in practice.
7553 This issue was discovered by Brian Carpenter.
7558 * Removed the export ciphers from the DEFAULT ciphers
7562 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7564 * Build fixes for the Windows and OpenVMS platforms
7566 *Matt Caswell and Richard Levitte*
7568 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7570 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7571 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7572 dereference. This could lead to a Denial Of Service attack. Thanks to
7573 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7578 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7579 dtls1_buffer_record function under certain conditions. In particular this
7580 could occur if an attacker sent repeated DTLS records with the same
7581 sequence number but for the next epoch. The memory leak could be exploited
7582 by an attacker in a Denial of Service attack through memory exhaustion.
7583 Thanks to Chris Mueller for reporting this issue.
7588 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7589 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7590 method would be set to NULL which could later result in a NULL pointer
7591 dereference. Thanks to Frank Schmirler for reporting this issue.
7596 * Abort handshake if server key exchange message is omitted for ephemeral
7599 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7600 reporting this issue.
7605 * Remove non-export ephemeral RSA code on client and server. This code
7606 violated the TLS standard by allowing the use of temporary RSA keys in
7607 non-export ciphersuites and could be used by a server to effectively
7608 downgrade the RSA key length used to a value smaller than the server
7609 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7610 INRIA or reporting this issue.
7615 * Fixed issue where DH client certificates are accepted without verification.
7616 An OpenSSL server will accept a DH certificate for client authentication
7617 without the certificate verify message. This effectively allows a client to
7618 authenticate without the use of a private key. This only affects servers
7619 which trust a client certificate authority which issues certificates
7620 containing DH keys: these are extremely rare and hardly ever encountered.
7621 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7627 * Ensure that the session ID context of an SSL is updated when its
7628 SSL_CTX is updated via SSL_set_SSL_CTX.
7630 The session ID context is typically set from the parent SSL_CTX,
7631 and can vary with the CTX.
7635 * Fix various certificate fingerprint issues.
7637 By using non-DER or invalid encodings outside the signed portion of a
7638 certificate the fingerprint can be changed without breaking the signature.
7639 Although no details of the signed portion of the certificate can be changed
7640 this can cause problems with some applications: e.g. those using the
7641 certificate fingerprint for blacklists.
7643 1. Reject signatures with non zero unused bits.
7645 If the BIT STRING containing the signature has non zero unused bits reject
7646 the signature. All current signature algorithms require zero unused bits.
7648 2. Check certificate algorithm consistency.
7650 Check the AlgorithmIdentifier inside TBS matches the one in the
7651 certificate signature. NB: this will result in signature failure
7652 errors for some broken certificates.
7654 Thanks to Konrad Kraszewski from Google for reporting this issue.
7656 3. Check DSA/ECDSA signatures use DER.
7658 Re-encode DSA/ECDSA signatures and compare with the original received
7659 signature. Return an error if there is a mismatch.
7661 This will reject various cases including garbage after signature
7662 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7663 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7664 (negative or with leading zeroes).
7666 Further analysis was conducted and fixes were developed by Stephen Henson
7667 of the OpenSSL core team.
7673 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7674 results on some platforms, including x86_64. This bug occurs at random
7675 with a very low probability, and is not known to be exploitable in any
7676 way, though its exact impact is difficult to determine. Thanks to Pieter
7677 Wuille (Blockstream) who reported this issue and also suggested an initial
7678 fix. Further analysis was conducted by the OpenSSL development team and
7679 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7680 the OpenSSL core team.
7685 * Do not resume sessions on the server if the negotiated protocol
7686 version does not match the session's version. Resuming with a different
7687 version, while not strictly forbidden by the RFC, is of questionable
7688 sanity and breaks all known clients.
7690 *David Benjamin, Emilia Käsper*
7692 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7693 early CCS messages during renegotiation. (Note that because
7694 renegotiation is encrypted, this early CCS was not exploitable.)
7698 * Tighten client-side session ticket handling during renegotiation:
7699 ensure that the client only accepts a session ticket if the server sends
7700 the extension anew in the ServerHello. Previously, a TLS client would
7701 reuse the old extension state and thus accept a session ticket if one was
7702 announced in the initial ServerHello.
7704 Similarly, ensure that the client requires a session ticket if one
7705 was advertised in the ServerHello. Previously, a TLS client would
7706 ignore a missing NewSessionTicket message.
7710 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7714 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7715 sends a carefully crafted handshake message, to cause OpenSSL to fail
7716 to free up to 64k of memory causing a memory leak. This could be
7717 exploited in a Denial Of Service attack. This issue affects OpenSSL
7718 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7719 whether SRTP is used or configured. Implementations of OpenSSL that
7720 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7722 The fix was developed by the OpenSSL team.
7727 * Session Ticket Memory Leak.
7729 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7730 integrity of that ticket is first verified. In the event of a session
7731 ticket integrity check failing, OpenSSL will fail to free memory
7732 causing a memory leak. By sending a large number of invalid session
7733 tickets an attacker could exploit this issue in a Denial Of Service
7739 * Build option no-ssl3 is incomplete.
7741 When OpenSSL is configured with "no-ssl3" as a build option, servers
7742 could accept and complete an SSL 3.0 handshake, and clients could be
7743 configured to send them.
7746 *Akamai and the OpenSSL team*
7748 * Add support for TLS_FALLBACK_SCSV.
7749 Client applications doing fallback retries should call
7750 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7753 *Adam Langley, Bodo Moeller*
7755 * Add additional DigestInfo checks.
7757 Re-encode DigestInto in DER and check against the original when
7758 verifying RSA signature: this will reject any improperly encoded
7759 DigestInfo structures.
7761 Note: this is a precautionary measure and no attacks are currently known.
7765 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7767 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7768 SRP code can be overrun an internal buffer. Add sanity check that
7769 g, A, B < N to SRP code.
7771 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7772 Group for discovering this issue.
7777 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7778 TLS 1.0 instead of higher protocol versions when the ClientHello message
7779 is badly fragmented. This allows a man-in-the-middle attacker to force a
7780 downgrade to TLS 1.0 even if both the server and the client support a
7781 higher protocol version, by modifying the client's TLS records.
7783 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7784 researching this issue.
7789 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7790 to a denial of service attack. A malicious server can crash the client
7791 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7792 ciphersuite and sending carefully crafted handshake messages.
7794 Thanks to Felix Gröbert (Google) for discovering and researching this
7800 * By sending carefully crafted DTLS packets an attacker could cause openssl
7801 to leak memory. This can be exploited through a Denial of Service attack.
7802 Thanks to Adam Langley for discovering and researching this issue.
7807 * An attacker can force openssl to consume large amounts of memory whilst
7808 processing DTLS handshake messages. This can be exploited through a
7809 Denial of Service attack.
7810 Thanks to Adam Langley for discovering and researching this issue.
7815 * An attacker can force an error condition which causes openssl to crash
7816 whilst processing DTLS packets due to memory being freed twice. This
7817 can be exploited through a Denial of Service attack.
7818 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7824 * If a multithreaded client connects to a malicious server using a resumed
7825 session and the server sends an ec point format extension it could write
7826 up to 255 bytes to freed memory.
7828 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7834 * A malicious server can crash an OpenSSL client with a null pointer
7835 dereference (read) by specifying an SRP ciphersuite even though it was not
7836 properly negotiated with the client. This can be exploited through a
7837 Denial of Service attack.
7839 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7840 discovering and researching this issue.
7845 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7846 X509_name_oneline, X509_name_print_ex et al. to leak some information
7847 from the stack. Applications may be affected if they echo pretty printing
7848 output to the attacker.
7850 Thanks to Ivan Fratric (Google) for discovering this issue.
7853 *Emilia Käsper, and Steve Henson*
7855 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7856 for corner cases. (Certain input points at infinity could lead to
7857 bogus results, with non-infinity inputs mapped to infinity too.)
7861 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7863 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7864 handshake can force the use of weak keying material in OpenSSL
7865 SSL/TLS clients and servers.
7867 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7868 researching this issue. ([CVE-2014-0224])
7870 *KIKUCHI Masashi, Steve Henson*
7872 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7873 OpenSSL DTLS client the code can be made to recurse eventually crashing
7876 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7879 *Imre Rad, Steve Henson*
7881 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7882 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7883 client or server. This is potentially exploitable to run arbitrary
7884 code on a vulnerable client or server.
7886 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7888 *Jüri Aedla, Steve Henson*
7890 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7891 are subject to a denial of service attack.
7893 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7894 this issue. ([CVE-2014-3470])
7896 *Felix Gröbert, Ivan Fratric, Steve Henson*
7898 * Harmonize version and its documentation. -f flag is used to display
7901 *mancha <mancha1@zoho.com>*
7903 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7904 in i2d_ECPrivateKey.
7906 *mancha <mancha1@zoho.com>*
7908 * Fix some double frees. These are not thought to be exploitable.
7910 *mancha <mancha1@zoho.com>*
7912 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7914 * A missing bounds check in the handling of the TLS heartbeat extension
7915 can be used to reveal up to 64k of memory to a connected client or
7918 Thanks for Neel Mehta of Google Security for discovering this bug and to
7919 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7920 preparing the fix ([CVE-2014-0160])
7922 *Adam Langley, Bodo Moeller*
7924 * Fix for the attack described in the paper "Recovering OpenSSL
7925 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7926 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7927 <http://eprint.iacr.org/2014/140>
7929 Thanks to Yuval Yarom and Naomi Benger for discovering this
7930 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7932 *Yuval Yarom and Naomi Benger*
7934 * TLS pad extension: draft-agl-tls-padding-03
7936 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7937 TLS client Hello record length value would otherwise be > 255 and
7938 less that 512 pad with a dummy extension containing zeroes so it
7939 is at least 512 bytes long.
7941 *Adam Langley, Steve Henson*
7943 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7945 * Fix for TLS record tampering bug. A carefully crafted invalid
7946 handshake could crash OpenSSL with a NULL pointer exception.
7947 Thanks to Anton Johansson for reporting this issues.
7950 * Keep original DTLS digest and encryption contexts in retransmission
7951 structures so we can use the previous session parameters if they need
7952 to be resent. ([CVE-2013-6450])
7956 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7957 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7958 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7959 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7960 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7961 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7963 *Rob Stradling, Adam Langley*
7965 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7967 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7968 supporting platforms or when small records were transferred.
7970 *Andy Polyakov, Steve Henson*
7972 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7974 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7976 This addresses the flaw in CBC record processing discovered by
7977 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7978 at: <http://www.isg.rhul.ac.uk/tls/>
7980 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7981 Security Group at Royal Holloway, University of London
7982 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7983 Emilia Käsper for the initial patch.
7986 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7988 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7989 ciphersuites which can be exploited in a denial of service attack.
7990 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7991 and detecting this bug and to Wolfgang Ettlinger
7992 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7997 * Return an error when checking OCSP signatures when key is NULL.
7998 This fixes a DoS attack. ([CVE-2013-0166])
8002 * Make openssl verify return errors.
8004 *Chris Palmer <palmer@google.com> and Ben Laurie*
8006 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8007 the right response is stapled. Also change SSL_get_certificate()
8008 so it returns the certificate actually sent.
8009 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8011 *Rob Stradling <rob.stradling@comodo.com>*
8013 * Fix possible deadlock when decoding public keys.
8017 * Don't use TLS 1.0 record version number in initial client hello
8022 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
8024 * Sanity check record length before skipping explicit IV in TLS
8025 1.2, 1.1 and DTLS to fix DoS attack.
8027 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8028 fuzzing as a service testing platform.
8033 * Initialise tkeylen properly when encrypting CMS messages.
8034 Thanks to Solar Designer of Openwall for reporting this issue.
8038 * In FIPS mode don't try to use composite ciphers as they are not
8043 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8045 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8046 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8047 mean any application compiled against OpenSSL 1.0.0 headers setting
8048 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8049 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8050 0x10000000L Any application which was previously compiled against
8051 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8052 will need to be recompiled as a result. Letting be results in
8053 inability to disable specifically TLS 1.1 and in client context,
8054 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8058 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8059 disable just protocol X, but all protocols above X *if* there are
8060 protocols *below* X still enabled. In more practical terms it means
8061 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8062 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8063 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8068 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8070 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8071 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8072 in CRYPTO_realloc_clean.
8074 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8075 issue and to Adam Langley <agl@chromium.org> for fixing it.
8078 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8080 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8084 * Workarounds for some broken servers that "hang" if a client hello
8085 record length exceeds 255 bytes.
8087 1. Do not use record version number > TLS 1.0 in initial client
8088 hello: some (but not all) hanging servers will now work.
8089 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8090 the number of ciphers sent in the client hello. This should be
8091 set to an even number, such as 50, for example by passing:
8092 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8093 Most broken servers should now work.
8094 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8095 TLS 1.2 client support entirely.
8099 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8103 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8105 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8106 STRING form instead of a DigestInfo.
8110 * The format used for MDC2 RSA signatures is inconsistent between EVP
8111 and the RSA_sign/RSA_verify functions. This was made more apparent when
8112 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8113 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8114 the correct format in RSA_verify so both forms transparently work.
8118 * Some servers which support TLS 1.0 can choke if we initially indicate
8119 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8120 encrypted premaster secret. As a workaround use the maximum permitted
8121 client version in client hello, this should keep such servers happy
8122 and still work with previous versions of OpenSSL.
8126 * Add support for TLS/DTLS heartbeats.
8128 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8130 * Add support for SCTP.
8132 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8134 * Improved PRNG seeding for VOS.
8136 *Paul Green <Paul.Green@stratus.com>*
8138 * Extensive assembler packs updates, most notably:
8140 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8141 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8142 - x86_64: bit-sliced AES implementation;
8143 - ARM: NEON support, contemporary platforms optimizations;
8144 - s390x: z196 support;
8145 - `*`: GHASH and GF(2^m) multiplication implementations;
8149 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8150 (removal of unnecessary code)
8152 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8154 * Add TLS key material exporter from RFC 5705.
8158 * Add DTLS-SRTP negotiation from RFC 5764.
8162 * Add Next Protocol Negotiation,
8163 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8164 disabled with a no-npn flag to config or Configure. Code donated
8167 *Adam Langley <agl@google.com> and Ben Laurie*
8169 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8170 NIST-P256, NIST-P521, with constant-time single point multiplication on
8171 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8172 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8173 Code made available under Apache License version 2.0.
8175 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8176 line to include this in your build of OpenSSL, and run "make depend" (or
8177 "make update"). This enables the following EC_METHODs:
8179 EC_GFp_nistp224_method()
8180 EC_GFp_nistp256_method()
8181 EC_GFp_nistp521_method()
8183 EC_GROUP_new_by_curve_name() will automatically use these (while
8184 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8187 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8189 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8190 all platforms. Move ssize_t definition from e_os.h to the public
8191 header file e_os2.h as it now appears in public header file cms.h
8195 * New -sigopt option to the ca, req and x509 utilities. Additional
8196 signature parameters can be passed using this option and in
8201 * Add RSA PSS signing function. This will generate and set the
8202 appropriate AlgorithmIdentifiers for PSS based on those in the
8203 corresponding EVP_MD_CTX structure. No application support yet.
8207 * Support for companion algorithm specific ASN1 signing routines.
8208 New function ASN1_item_sign_ctx() signs a pre-initialised
8209 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8210 the appropriate parameters.
8214 * Add new algorithm specific ASN1 verification initialisation function
8215 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8216 handling will be the same no matter what EVP_PKEY_METHOD is used.
8217 Add a PSS handler to support verification of PSS signatures: checked
8218 against a number of sample certificates.
8222 * Add signature printing for PSS. Add PSS OIDs.
8224 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8226 * Add algorithm specific signature printing. An individual ASN1 method
8227 can now print out signatures instead of the standard hex dump.
8229 More complex signatures (e.g. PSS) can print out more meaningful
8230 information. Include DSA version that prints out the signature
8235 * Password based recipient info support for CMS library: implementing
8240 * Split password based encryption into PBES2 and PBKDF2 functions. This
8241 neatly separates the code into cipher and PBE sections and is required
8242 for some algorithms that split PBES2 into separate pieces (such as
8243 password based CMS).
8247 * Session-handling fixes:
8248 - Fix handling of connections that are resuming with a session ID,
8249 but also support Session Tickets.
8250 - Fix a bug that suppressed issuing of a new ticket if the client
8251 presented a ticket with an expired session.
8252 - Try to set the ticket lifetime hint to something reasonable.
8253 - Make tickets shorter by excluding irrelevant information.
8254 - On the client side, don't ignore renewed tickets.
8256 *Adam Langley, Bodo Moeller (Google)*
8258 * Fix PSK session representation.
8262 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8264 This work was sponsored by Intel.
8268 * Add GCM support to TLS library. Some custom code is needed to split
8269 the IV between the fixed (from PRF) and explicit (from TLS record)
8270 portions. This adds all GCM ciphersuites supported by RFC5288 and
8271 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8272 add a special AESGCM string for GCM only.
8276 * Expand range of ctrls for AES GCM. Permit setting invocation
8277 field on decrypt and retrieval of invocation field only on encrypt.
8281 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8282 As required by RFC5289 these ciphersuites cannot be used if for
8283 versions of TLS earlier than 1.2.
8287 * For FIPS capable OpenSSL interpret a NULL default public key method
8288 as unset and return the appropriate default but do *not* set the default.
8289 This means we can return the appropriate method in applications that
8290 switch between FIPS and non-FIPS modes.
8294 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8295 ENGINE is used then we cannot handle that in the FIPS module so we
8296 keep original code iff non-FIPS operations are allowed.
8300 * Add -attime option to openssl utilities.
8302 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8304 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8308 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8309 FIPS EC methods unconditionally for now.
8313 * New build option no-ec2m to disable characteristic 2 code.
8317 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8318 all cases can be covered as some introduce binary incompatibilities.
8322 * Redirect RSA operations to FIPS module including keygen,
8323 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8327 * Add similar low-level API blocking to ciphers.
8331 * low-level digest APIs are not approved in FIPS mode: any attempt
8332 to use these will cause a fatal error. Applications that *really* want
8333 to use them can use the `private_*` version instead.
8337 * Redirect cipher operations to FIPS module for FIPS builds.
8341 * Redirect digest operations to FIPS module for FIPS builds.
8345 * Update build system to add "fips" flag which will link in fipscanister.o
8346 for static and shared library builds embedding a signature if needed.
8350 * Output TLS supported curves in preference order instead of numerical
8351 order. This is currently hardcoded for the highest order curves first.
8352 This should be configurable so applications can judge speed vs strength.
8356 * Add TLS v1.2 server support for client authentication.
8360 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8365 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8366 FIPS modules versions.
8370 * Add TLS v1.2 client side support for client authentication. Keep cache
8371 of handshake records longer as we don't know the hash algorithm to use
8372 until after the certificate request message is received.
8376 * Initial TLS v1.2 client support. Add a default signature algorithms
8377 extension including all the algorithms we support. Parse new signature
8378 format in client key exchange. Relax some ECC signing restrictions for
8379 TLS v1.2 as indicated in RFC5246.
8383 * Add server support for TLS v1.2 signature algorithms extension. Switch
8384 to new signature format when needed using client digest preference.
8385 All server ciphersuites should now work correctly in TLS v1.2. No client
8386 support yet and no support for client certificates.
8390 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8391 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8392 ciphersuites. At present only RSA key exchange ciphersuites work with
8393 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8394 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8395 and version checking.
8399 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8400 with this defined it will not be affected by any changes to ssl internal
8401 structures. Add several utility functions to allow openssl application
8402 to work with OPENSSL_NO_SSL_INTERN defined.
8406 * A long standing patch to add support for SRP from EdelWeb (Peter
8407 Sylvester and Christophe Renou) was integrated.
8408 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8409 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8412 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8416 * Permit abbreviated handshakes when renegotiating using the function
8417 SSL_renegotiate_abbreviated().
8419 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8421 * Add call to ENGINE_register_all_complete() to
8422 ENGINE_load_builtin_engines(), so some implementations get used
8423 automatically instead of needing explicit application support.
8427 * Add support for TLS key exporter as described in RFC5705.
8429 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8431 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8432 a few changes are required:
8434 Add SSL_OP_NO_TLSv1_1 flag.
8435 Add TLSv1_1 methods.
8436 Update version checking logic to handle version 1.1.
8437 Add explicit IV handling (ported from DTLS code).
8438 Add command line options to s_client/s_server.
8445 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8447 * X509_ATTRIBUTE memory leak
8449 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8450 memory. This structure is used by the PKCS#7 and CMS routines so any
8451 application which reads PKCS#7 or CMS data from untrusted sources is
8452 affected. SSL/TLS is not affected.
8454 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8460 * Race condition handling PSK identify hint
8462 If PSK identity hints are received by a multi-threaded client then
8463 the values are wrongly updated in the parent SSL_CTX structure. This can
8464 result in a race condition potentially leading to a double free of the
8470 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8472 * Malformed ECParameters causes infinite loop
8474 When processing an ECParameters structure OpenSSL enters an infinite loop
8475 if the curve specified is over a specially malformed binary polynomial
8478 This can be used to perform denial of service against any
8479 system which processes public keys, certificate requests or
8480 certificates. This includes TLS clients and TLS servers with
8481 client authentication enabled.
8483 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8488 * Exploitable out-of-bounds read in X509_cmp_time
8490 X509_cmp_time does not properly check the length of the ASN1_TIME
8491 string and can read a few bytes out of bounds. In addition,
8492 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8495 An attacker can use this to craft malformed certificates and CRLs of
8496 various sizes and potentially cause a segmentation fault, resulting in
8497 a DoS on applications that verify certificates or CRLs. TLS clients
8498 that verify CRLs are affected. TLS clients and servers with client
8499 authentication enabled may be affected if they use custom verification
8502 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8503 independently by Hanno Böck.
8508 * PKCS7 crash with missing EnvelopedContent
8510 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8511 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8512 with missing content and trigger a NULL pointer dereference on parsing.
8514 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8515 structures from untrusted sources are affected. OpenSSL clients and
8516 servers are not affected.
8518 This issue was reported to OpenSSL by Michal Zalewski (Google).
8523 * CMS verify infinite loop with unknown hash function
8525 When verifying a signedData message the CMS code can enter an infinite loop
8526 if presented with an unknown hash function OID. This can be used to perform
8527 denial of service against any system which verifies signedData messages using
8529 This issue was reported to OpenSSL by Johannes Bauer.
8534 * Race condition handling NewSessionTicket
8536 If a NewSessionTicket is received by a multi-threaded client when attempting to
8537 reuse a previous ticket then a race condition can occur potentially leading to
8538 a double free of the ticket data.
8543 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8545 * Segmentation fault in ASN1_TYPE_cmp fix
8547 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8548 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8549 certificate signature algorithm consistency this can be used to crash any
8550 certificate verification operation and exploited in a DoS attack. Any
8551 application which performs certificate verification is vulnerable including
8552 OpenSSL clients and servers which enable client authentication.
8557 * ASN.1 structure reuse memory corruption fix
8559 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8560 memory corruption via an invalid write. Such reuse is and has been
8561 strongly discouraged and is believed to be rare.
8563 Applications that parse structures containing CHOICE or ANY DEFINED BY
8564 components may be affected. Certificate parsing (d2i_X509 and related
8565 functions) are however not affected. OpenSSL clients and servers are
8571 * PKCS7 NULL pointer dereferences fix
8573 The PKCS#7 parsing code does not handle missing outer ContentInfo
8574 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8575 missing content and trigger a NULL pointer dereference on parsing.
8577 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8578 otherwise parse PKCS#7 structures from untrusted sources are
8579 affected. OpenSSL clients and servers are not affected.
8581 This issue was reported to OpenSSL by Michal Zalewski (Google).
8586 * DoS via reachable assert in SSLv2 servers fix
8588 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8589 servers that both support SSLv2 and enable export cipher suites by sending
8590 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8592 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8593 (OpenSSL development team).
8598 * Use After Free following d2i_ECPrivatekey error fix
8600 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8601 could cause a use after free condition. This, in turn, could cause a double
8602 free in several private key parsing functions (such as d2i_PrivateKey
8603 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8604 for applications that receive EC private keys from untrusted
8605 sources. This scenario is considered rare.
8607 This issue was discovered by the BoringSSL project and fixed in their
8613 * X509_to_X509_REQ NULL pointer deref fix
8615 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8616 the certificate key is invalid. This function is rarely used in practice.
8618 This issue was discovered by Brian Carpenter.
8623 * Removed the export ciphers from the DEFAULT ciphers
8627 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8629 * Build fixes for the Windows and OpenVMS platforms
8631 *Matt Caswell and Richard Levitte*
8633 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8635 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8636 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8637 dereference. This could lead to a Denial Of Service attack. Thanks to
8638 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8643 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8644 dtls1_buffer_record function under certain conditions. In particular this
8645 could occur if an attacker sent repeated DTLS records with the same
8646 sequence number but for the next epoch. The memory leak could be exploited
8647 by an attacker in a Denial of Service attack through memory exhaustion.
8648 Thanks to Chris Mueller for reporting this issue.
8653 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8654 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8655 method would be set to NULL which could later result in a NULL pointer
8656 dereference. Thanks to Frank Schmirler for reporting this issue.
8661 * Abort handshake if server key exchange message is omitted for ephemeral
8664 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8665 reporting this issue.
8670 * Remove non-export ephemeral RSA code on client and server. This code
8671 violated the TLS standard by allowing the use of temporary RSA keys in
8672 non-export ciphersuites and could be used by a server to effectively
8673 downgrade the RSA key length used to a value smaller than the server
8674 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8675 INRIA or reporting this issue.
8680 * Fixed issue where DH client certificates are accepted without verification.
8681 An OpenSSL server will accept a DH certificate for client authentication
8682 without the certificate verify message. This effectively allows a client to
8683 authenticate without the use of a private key. This only affects servers
8684 which trust a client certificate authority which issues certificates
8685 containing DH keys: these are extremely rare and hardly ever encountered.
8686 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8692 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8693 results on some platforms, including x86_64. This bug occurs at random
8694 with a very low probability, and is not known to be exploitable in any
8695 way, though its exact impact is difficult to determine. Thanks to Pieter
8696 Wuille (Blockstream) who reported this issue and also suggested an initial
8697 fix. Further analysis was conducted by the OpenSSL development team and
8698 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8699 the OpenSSL core team.
8704 * Fix various certificate fingerprint issues.
8706 By using non-DER or invalid encodings outside the signed portion of a
8707 certificate the fingerprint can be changed without breaking the signature.
8708 Although no details of the signed portion of the certificate can be changed
8709 this can cause problems with some applications: e.g. those using the
8710 certificate fingerprint for blacklists.
8712 1. Reject signatures with non zero unused bits.
8714 If the BIT STRING containing the signature has non zero unused bits reject
8715 the signature. All current signature algorithms require zero unused bits.
8717 2. Check certificate algorithm consistency.
8719 Check the AlgorithmIdentifier inside TBS matches the one in the
8720 certificate signature. NB: this will result in signature failure
8721 errors for some broken certificates.
8723 Thanks to Konrad Kraszewski from Google for reporting this issue.
8725 3. Check DSA/ECDSA signatures use DER.
8727 Reencode DSA/ECDSA signatures and compare with the original received
8728 signature. Return an error if there is a mismatch.
8730 This will reject various cases including garbage after signature
8731 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8732 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8733 (negative or with leading zeroes).
8735 Further analysis was conducted and fixes were developed by Stephen Henson
8736 of the OpenSSL core team.
8742 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8744 * Session Ticket Memory Leak.
8746 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8747 integrity of that ticket is first verified. In the event of a session
8748 ticket integrity check failing, OpenSSL will fail to free memory
8749 causing a memory leak. By sending a large number of invalid session
8750 tickets an attacker could exploit this issue in a Denial Of Service
8756 * Build option no-ssl3 is incomplete.
8758 When OpenSSL is configured with "no-ssl3" as a build option, servers
8759 could accept and complete an SSL 3.0 handshake, and clients could be
8760 configured to send them.
8763 *Akamai and the OpenSSL team*
8765 * Add support for TLS_FALLBACK_SCSV.
8766 Client applications doing fallback retries should call
8767 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8770 *Adam Langley, Bodo Moeller*
8772 * Add additional DigestInfo checks.
8774 Reencode DigestInto in DER and check against the original when
8775 verifying RSA signature: this will reject any improperly encoded
8776 DigestInfo structures.
8778 Note: this is a precautionary measure and no attacks are currently known.
8782 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8784 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8785 to a denial of service attack. A malicious server can crash the client
8786 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8787 ciphersuite and sending carefully crafted handshake messages.
8789 Thanks to Felix Gröbert (Google) for discovering and researching this
8795 * By sending carefully crafted DTLS packets an attacker could cause openssl
8796 to leak memory. This can be exploited through a Denial of Service attack.
8797 Thanks to Adam Langley for discovering and researching this issue.
8802 * An attacker can force openssl to consume large amounts of memory whilst
8803 processing DTLS handshake messages. This can be exploited through a
8804 Denial of Service attack.
8805 Thanks to Adam Langley for discovering and researching this issue.
8810 * An attacker can force an error condition which causes openssl to crash
8811 whilst processing DTLS packets due to memory being freed twice. This
8812 can be exploited through a Denial of Service attack.
8813 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8819 * If a multithreaded client connects to a malicious server using a resumed
8820 session and the server sends an ec point format extension it could write
8821 up to 255 bytes to freed memory.
8823 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8829 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8830 X509_name_oneline, X509_name_print_ex et al. to leak some information
8831 from the stack. Applications may be affected if they echo pretty printing
8832 output to the attacker.
8834 Thanks to Ivan Fratric (Google) for discovering this issue.
8837 *Emilia Käsper, and Steve Henson*
8839 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8840 for corner cases. (Certain input points at infinity could lead to
8841 bogus results, with non-infinity inputs mapped to infinity too.)
8845 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8847 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8848 handshake can force the use of weak keying material in OpenSSL
8849 SSL/TLS clients and servers.
8851 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8852 researching this issue. ([CVE-2014-0224])
8854 *KIKUCHI Masashi, Steve Henson*
8856 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8857 OpenSSL DTLS client the code can be made to recurse eventually crashing
8860 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8863 *Imre Rad, Steve Henson*
8865 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8866 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8867 client or server. This is potentially exploitable to run arbitrary
8868 code on a vulnerable client or server.
8870 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8872 *Jüri Aedla, Steve Henson*
8874 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8875 are subject to a denial of service attack.
8877 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8878 this issue. ([CVE-2014-3470])
8880 *Felix Gröbert, Ivan Fratric, Steve Henson*
8882 * Harmonize version and its documentation. -f flag is used to display
8885 *mancha <mancha1@zoho.com>*
8887 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8888 in i2d_ECPrivateKey.
8890 *mancha <mancha1@zoho.com>*
8892 * Fix some double frees. These are not thought to be exploitable.
8894 *mancha <mancha1@zoho.com>*
8896 * Fix for the attack described in the paper "Recovering OpenSSL
8897 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8898 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8899 <http://eprint.iacr.org/2014/140>
8901 Thanks to Yuval Yarom and Naomi Benger for discovering this
8902 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8904 *Yuval Yarom and Naomi Benger*
8906 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8908 * Keep original DTLS digest and encryption contexts in retransmission
8909 structures so we can use the previous session parameters if they need
8910 to be resent. ([CVE-2013-6450])
8914 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8915 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8916 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8917 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8918 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8919 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8921 *Rob Stradling, Adam Langley*
8923 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8925 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8927 This addresses the flaw in CBC record processing discovered by
8928 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8929 at: <http://www.isg.rhul.ac.uk/tls/>
8931 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8932 Security Group at Royal Holloway, University of London
8933 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8934 Emilia Käsper for the initial patch.
8937 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8939 * Return an error when checking OCSP signatures when key is NULL.
8940 This fixes a DoS attack. ([CVE-2013-0166])
8944 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8945 the right response is stapled. Also change SSL_get_certificate()
8946 so it returns the certificate actually sent.
8947 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8948 (This is a backport)
8950 *Rob Stradling <rob.stradling@comodo.com>*
8952 * Fix possible deadlock when decoding public keys.
8956 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8958 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8961 * Sanity check record length before skipping explicit IV in DTLS
8964 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8965 fuzzing as a service testing platform.
8970 * Initialise tkeylen properly when encrypting CMS messages.
8971 Thanks to Solar Designer of Openwall for reporting this issue.
8975 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8977 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8978 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8979 in CRYPTO_realloc_clean.
8981 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8982 issue and to Adam Langley <agl@chromium.org> for fixing it.
8985 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8987 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8989 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8990 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8991 content decryption and always return the same error. Note: this attack
8992 needs on average 2^20 messages so it only affects automated senders. The
8993 old behaviour can be re-enabled in the CMS code by setting the
8994 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8995 an MMA defence is not necessary.
8996 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8997 this issue. ([CVE-2012-0884])
9001 * Fix CVE-2011-4619: make sure we really are receiving a
9002 client hello before rejecting multiple SGC restarts. Thanks to
9003 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
9007 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
9009 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
9010 Thanks to Antonio Martin, Enterprise Secure Access Research and
9011 Development, Cisco Systems, Inc. for discovering this bug and
9012 preparing a fix. ([CVE-2012-0050])
9016 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
9018 * Nadhem Alfardan and Kenny Paterson have discovered an extension
9019 of the Vaudenay padding oracle attack on CBC mode encryption
9020 which enables an efficient plaintext recovery attack against
9021 the OpenSSL implementation of DTLS. Their attack exploits timing
9022 differences arising during decryption processing. A research
9023 paper describing this attack can be found at:
9024 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
9025 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9026 Security Group at Royal Holloway, University of London
9027 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9028 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9029 for preparing the fix. ([CVE-2011-4108])
9031 *Robin Seggelmann, Michael Tuexen*
9033 * Clear bytes used for block padding of SSL 3.0 records.
9036 *Adam Langley (Google)*
9038 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9039 Kadianakis <desnacked@gmail.com> for discovering this issue and
9040 Adam Langley for preparing the fix. ([CVE-2011-4619])
9042 *Adam Langley (Google)*
9044 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9046 *Andrey Kulikov <amdeich@gmail.com>*
9048 * Prevent malformed RFC3779 data triggering an assertion failure.
9049 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9050 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9052 *Rob Austein <sra@hactrn.net>*
9054 * Improved PRNG seeding for VOS.
9056 *Paul Green <Paul.Green@stratus.com>*
9058 * Fix ssl_ciph.c set-up race.
9060 *Adam Langley (Google)*
9062 * Fix spurious failures in ecdsatest.c.
9064 *Emilia Käsper (Google)*
9066 * Fix the BIO_f_buffer() implementation (which was mixing different
9067 interpretations of the `..._len` fields).
9069 *Adam Langley (Google)*
9071 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9072 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9073 threads won't reuse the same blinding coefficients.
9075 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9076 lock to call BN_BLINDING_invert_ex, and avoids one use of
9077 BN_BLINDING_update for each BN_BLINDING structure (previously,
9078 the last update always remained unused).
9080 *Emilia Käsper (Google)*
9082 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9084 *Bob Buckholz (Google)*
9086 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9088 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9089 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9091 *Kaspar Brand <ossl@velox.ch>*
9093 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9094 for multi-threaded use of ECDH. ([CVE-2011-3210])
9096 *Adam Langley (Google)*
9098 * Fix x509_name_ex_d2i memory leak on bad inputs.
9102 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9103 signature public key algorithm by using OID xref utilities instead.
9104 Before this you could only use some ECC ciphersuites with SHA1 only.
9108 * Add protection against ECDSA timing attacks as mentioned in the paper
9109 by Billy Bob Brumley and Nicola Tuveri, see:
9110 <http://eprint.iacr.org/2011/232.pdf>
9112 *Billy Bob Brumley and Nicola Tuveri*
9114 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9116 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9118 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9120 * Fix bug in string printing code: if *any* escaping is enabled we must
9121 escape the escape character (backslash) or the resulting string is
9126 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9128 * Disable code workaround for ancient and obsolete Netscape browsers
9129 and servers: an attacker can use it in a ciphersuite downgrade attack.
9130 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9134 * Fixed J-PAKE implementation error, originally discovered by
9135 Sebastien Martini, further info and confirmation from Stefan
9136 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9140 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9142 * Fix extension code to avoid race conditions which can result in a buffer
9143 overrun vulnerability: resumed sessions must not be modified as they can
9144 be shared by multiple threads. CVE-2010-3864
9148 * Fix WIN32 build system to correctly link an ENGINE directory into
9153 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9155 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9158 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9160 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9162 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9163 context. The operation can be customised via the ctrl mechanism in
9164 case ENGINEs want to include additional functionality.
9168 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9172 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9173 output hashes compatible with older versions of OpenSSL.
9175 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9177 * Fix compression algorithm handling: if resuming a session use the
9178 compression algorithm of the resumed session instead of determining
9179 it from client hello again. Don't allow server to change algorithm.
9183 * Add load_crls() function to commands tidying load_certs() too. Add option
9184 to verify utility to allow additional CRLs to be included.
9188 * Update OCSP request code to permit adding custom headers to the request:
9189 some responders need this.
9193 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9196 *Julia Lawall <julia@diku.dk>*
9198 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9199 needlessly dereferenced structures, used obsolete functions and
9200 didn't handle all updated verify codes correctly.
9204 * Disable MD2 in the default configuration.
9208 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9209 indicate the initial BIO being pushed or popped. This makes it possible
9210 to determine whether the BIO is the one explicitly called or as a result
9211 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9212 it handles reference counts correctly and doesn't zero out the I/O bio
9213 when it is not being explicitly popped. WARNING: applications which
9214 included workarounds for the old buggy behaviour will need to be modified
9215 or they could free up already freed BIOs.
9219 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9220 renaming to all platforms (within the 0.9.8 branch, this was
9221 done conditionally on Netware platforms to avoid a name clash).
9223 *Guenter <lists@gknw.net>*
9225 * Add ECDHE and PSK support to DTLS.
9227 *Michael Tuexen <tuexen@fh-muenster.de>*
9229 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9234 * Add "missing" function EVP_MD_flags() (without this the only way to
9235 retrieve a digest flags is by accessing the structure directly. Update
9236 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9237 or cipher is registered as in the "from" argument. Print out all
9238 registered digests in the dgst usage message instead of manually
9239 attempting to work them out.
9243 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9244 this allows the use of compression and extensions. Change default cipher
9245 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9246 by default unless an application cipher string requests it.
9250 * Alter match criteria in PKCS12_parse(). It used to try to use local
9251 key ids to find matching certificates and keys but some PKCS#12 files
9252 don't follow the (somewhat unwritten) rules and this strategy fails.
9253 Now just gather all certificates together and the first private key
9254 then look for the first certificate that matches the key.
9258 * Support use of registered digest and cipher names for dgst and cipher
9259 commands instead of having to add each one as a special case. So now
9266 openssl dgst -sha256 foo
9268 and this works for ENGINE based algorithms too.
9272 * Update Gost ENGINE to support parameter files.
9274 *Victor B. Wagner <vitus@cryptocom.ru>*
9276 * Support GeneralizedTime in ca utility.
9278 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9280 * Enhance the hash format used for certificate directory links. The new
9281 form uses the canonical encoding (meaning equivalent names will work
9282 even if they aren't identical) and uses SHA1 instead of MD5. This form
9283 is incompatible with the older format and as a result c_rehash should
9284 be used to rebuild symbolic links.
9288 * Make PKCS#8 the default write format for private keys, replacing the
9289 traditional format. This form is standardised, more secure and doesn't
9290 include an implicit MD5 dependency.
9294 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9295 committed to OpenSSL should pass this lot as a minimum.
9299 * Add session ticket override functionality for use by EAP-FAST.
9301 *Jouni Malinen <j@w1.fi>*
9303 * Modify HMAC functions to return a value. Since these can be implemented
9304 in an ENGINE errors can occur.
9308 * Type-checked OBJ_bsearch_ex.
9312 * Type-checked OBJ_bsearch. Also some constification necessitated
9313 by type-checking. Still to come: TXT_DB, bsearch(?),
9314 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9319 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9320 seconds to a tm structure directly, instead of going through OS
9321 specific date routines. This avoids any issues with OS routines such
9322 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9323 and X509_time_adj_ex() to cover the extended range. The existing
9324 X509_time_adj() is still usable and will no longer have any date issues.
9328 * Delta CRL support. New use deltas option which will attempt to locate
9329 and search any appropriate delta CRLs available.
9331 This work was sponsored by Google.
9335 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9336 code and add additional score elements. Validate alternate CRL paths
9337 as part of the CRL checking and indicate a new error "CRL path validation
9338 error" in this case. Applications wanting additional details can use
9339 the verify callback and check the new "parent" field. If this is not
9340 NULL CRL path validation is taking place. Existing applications won't
9341 see this because it requires extended CRL support which is off by
9344 This work was sponsored by Google.
9348 * Support for freshest CRL extension.
9350 This work was sponsored by Google.
9354 * Initial indirect CRL support. Currently only supported in the CRLs
9355 passed directly and not via lookup. Process certificate issuer
9356 CRL entry extension and lookup CRL entries by bother issuer name
9357 and serial number. Check and process CRL issuer entry in IDP extension.
9359 This work was sponsored by Google.
9363 * Add support for distinct certificate and CRL paths. The CRL issuer
9364 certificate is validated separately in this case. Only enabled if
9365 an extended CRL support flag is set: this flag will enable additional
9366 CRL functionality in future.
9368 This work was sponsored by Google.
9372 * Add support for policy mappings extension.
9374 This work was sponsored by Google.
9378 * Fixes to pathlength constraint, self issued certificate handling,
9379 policy processing to align with RFC3280 and PKITS tests.
9381 This work was sponsored by Google.
9385 * Support for name constraints certificate extension. DN, email, DNS
9386 and URI types are currently supported.
9388 This work was sponsored by Google.
9392 * To cater for systems that provide a pointer-based thread ID rather
9393 than numeric, deprecate the current numeric thread ID mechanism and
9394 replace it with a structure and associated callback type. This
9395 mechanism allows a numeric "hash" to be extracted from a thread ID in
9396 either case, and on platforms where pointers are larger than 'long',
9397 mixing is done to help ensure the numeric 'hash' is usable even if it
9398 can't be guaranteed unique. The default mechanism is to use "&errno"
9399 as a pointer-based thread ID to distinguish between threads.
9401 Applications that want to provide their own thread IDs should now use
9402 CRYPTO_THREADID_set_callback() to register a callback that will call
9403 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9405 Note that ERR_remove_state() is now deprecated, because it is tied
9406 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9407 to free the current thread's error state should be replaced by
9408 ERR_remove_thread_state(NULL).
9410 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9411 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9412 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9413 application was previously providing a numeric thread callback that
9414 was inappropriate for distinguishing threads, then uniqueness might
9415 have been obtained with &errno that happened immediately in the
9416 intermediate development versions of OpenSSL; this is no longer the
9417 case, the numeric thread callback will now override the automatic use
9420 *Geoff Thorpe, with help from Bodo Moeller*
9422 * Initial support for different CRL issuing certificates. This covers a
9423 simple case where the self issued certificates in the chain exist and
9424 the real CRL issuer is higher in the existing chain.
9426 This work was sponsored by Google.
9430 * Removed effectively defunct crypto/store from the build.
9434 * Revamp of STACK to provide stronger type-checking. Still to come:
9435 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9436 ASN1_STRING, CONF_VALUE.
9440 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9441 RAM on SSL connections. This option can save about 34k per idle SSL.
9445 * Revamp of LHASH to provide stronger type-checking. Still to come:
9446 STACK, TXT_DB, bsearch, qsort.
9450 * Initial support for Cryptographic Message Syntax (aka CMS) based
9451 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9452 support for data, signedData, compressedData, digestedData and
9453 encryptedData, envelopedData types included. Scripts to check against
9454 RFC4134 examples draft and interop and consistency checks of many
9455 content types and variants.
9459 * Add options to enc utility to support use of zlib compression BIO.
9463 * Extend mk1mf to support importing of options and assembly language
9464 files from Configure script, currently only included in VC-WIN32.
9465 The assembly language rules can now optionally generate the source
9466 files from the associated perl scripts.
9470 * Implement remaining functionality needed to support GOST ciphersuites.
9471 Interop testing has been performed using CryptoPro implementations.
9473 *Victor B. Wagner <vitus@cryptocom.ru>*
9475 * s390x assembler pack.
9479 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9484 * Implement Opaque PRF Input TLS extension as specified in
9485 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9486 official specification yet and no extension type assignment by
9487 IANA exists, this extension (for now) will have to be explicitly
9488 enabled when building OpenSSL by providing the extension number
9489 to use. For example, specify an option
9491 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9493 to the "config" or "Configure" script to enable the extension,
9494 assuming extension number 0x9527 (which is a completely arbitrary
9495 and unofficial assignment based on the MD5 hash of the Internet
9496 Draft). Note that by doing so, you potentially lose
9497 interoperability with other TLS implementations since these might
9498 be using the same extension number for other purposes.
9500 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9501 opaque PRF input value to use in the handshake. This will create
9502 an internal copy of the length-'len' string at 'src', and will
9503 return non-zero for success.
9505 To get more control and flexibility, provide a callback function
9508 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9509 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9513 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9516 Callback function 'cb' will be called in handshakes, and is
9517 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9518 Argument 'arg' is for application purposes (the value as given to
9519 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9520 be provided to the callback function). The callback function
9521 has to return non-zero to report success: usually 1 to use opaque
9522 PRF input just if possible, or 2 to enforce use of the opaque PRF
9523 input. In the latter case, the library will abort the handshake
9524 if opaque PRF input is not successfully negotiated.
9526 Arguments 'peerinput' and 'len' given to the callback function
9527 will always be NULL and 0 in the case of a client. A server will
9528 see the client's opaque PRF input through these variables if
9529 available (NULL and 0 otherwise). Note that if the server
9530 provides an opaque PRF input, the length must be the same as the
9531 length of the client's opaque PRF input.
9533 Note that the callback function will only be called when creating
9534 a new session (session resumption can resume whatever was
9535 previously negotiated), and will not be called in SSL 2.0
9536 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9537 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9538 for applications that need to enforce opaque PRF input.
9542 * Update ssl code to support digests other than SHA1+MD5 for handshake
9545 *Victor B. Wagner <vitus@cryptocom.ru>*
9547 * Add RFC4507 support to OpenSSL. This includes the corrections in
9548 RFC4507bis. The encrypted ticket format is an encrypted encoded
9549 SSL_SESSION structure, that way new session features are automatically
9552 If a client application caches session in an SSL_SESSION structure
9553 support is transparent because tickets are now stored in the encoded
9556 The SSL_CTX structure automatically generates keys for ticket
9557 protection in servers so again support should be possible
9558 with no application modification.
9560 If a client or server wishes to disable RFC4507 support then the option
9561 SSL_OP_NO_TICKET can be set.
9563 Add a TLS extension debugging callback to allow the contents of any client
9564 or server extensions to be examined.
9566 This work was sponsored by Google.
9570 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9571 OpenSSL should now compile cleanly on gcc 4.2
9573 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9575 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9576 support including streaming MAC support: this is required for GOST
9577 ciphersuite support.
9579 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9581 * Add option -stream to use PKCS#7 streaming in smime utility. New
9582 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9583 to output in BER and PEM format.
9587 * Experimental support for use of HMAC via EVP_PKEY interface. This
9588 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9589 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9590 ENGINE support for HMAC keys which are unextractable. New -mac and
9591 -macopt options to dgst utility.
9595 * New option -sigopt to dgst utility. Update dgst to use
9596 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9597 alternative signing parameters such as X9.31 or PSS in the dgst
9602 * Change ssl_cipher_apply_rule(), the internal function that does
9603 the work each time a ciphersuite string requests enabling
9604 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9605 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9606 the order of disabled ciphersuites such that those ciphersuites
9607 that most recently went from enabled to disabled not only stay
9608 in order with respect to each other, but also have higher priority
9609 than other disabled ciphersuites the next time ciphersuites are
9612 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9613 the same ciphersuites as with "HIGH" alone, but in a specific
9614 order where the PSK ciphersuites come first (since they are the
9615 most recently disabled ciphersuites when "HIGH" is parsed).
9617 Also, change ssl_create_cipher_list() (using this new
9618 functionality) such that between otherwise identical
9619 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9624 * Change ssl_create_cipher_list() so that it automatically
9625 arranges the ciphersuites in reasonable order before starting
9626 to process the rule string. Thus, the definition for "DEFAULT"
9627 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9628 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9629 This makes it much easier to arrive at a reasonable default order
9630 in applications for which anonymous ciphers are OK (meaning
9631 that you can't actually use DEFAULT).
9633 *Bodo Moeller; suggested by Victor Duchovni*
9635 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9636 processing) into multiple integers instead of setting
9637 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9638 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9639 (These masks as well as the individual bit definitions are hidden
9640 away into the non-exported interface ssl/ssl_locl.h, so this
9641 change to the definition of the SSL_CIPHER structure shouldn't
9642 affect applications.) This give us more bits for each of these
9643 categories, so there is no longer a need to coagulate AES128 and
9644 AES256 into a single algorithm bit, and to coagulate Camellia128
9645 and Camellia256 into a single algorithm bit, which has led to all
9648 Thus, among other things, the kludge introduced in 0.9.7m and
9649 0.9.8e for masking out AES256 independently of AES128 or masking
9650 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9652 With the change, we also introduce new ciphersuite aliases that
9653 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9658 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9659 Use the leftmost N bytes of the signature input if the input is
9660 larger than the prime q (with N being the size in bytes of q).
9664 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9665 it yet and it is largely untested.
9669 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9673 * Initial incomplete changes to avoid need for function casts in OpenSSL
9674 some compilers (gcc 4.2 and later) reject their use. Safestack is
9675 reimplemented. Update ASN1 to avoid use of legacy functions.
9679 * Win32/64 targets are linked with Winsock2.
9683 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9684 to external functions. This can be used to increase CRL handling
9685 efficiency especially when CRLs are very large by (for example) storing
9686 the CRL revoked certificates in a database.
9690 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9691 new CRLs added to a directory can be used. New command line option
9692 -verify_return_error to s_client and s_server. This causes real errors
9693 to be returned by the verify callback instead of carrying on no matter
9694 what. This reflects the way a "real world" verify callback would behave.
9698 * GOST engine, supporting several GOST algorithms and public key formats.
9699 Kindly donated by Cryptocom.
9703 * Partial support for Issuing Distribution Point CRL extension. CRLs
9704 partitioned by DP are handled but no indirect CRL or reason partitioning
9705 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9706 selected via a scoring technique which handles IDP and AKID in CRLs.
9710 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9711 will ultimately be used for all verify operations: this will remove the
9712 X509_STORE dependency on certificate verification and allow alternative
9713 lookup methods. X509_STORE based implementations of these two callbacks.
9717 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9718 Modify get_crl() to find a valid (unexpired) CRL if possible.
9722 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9723 this would be called X509_CRL_cmp() but that name is already used by
9724 a function that just compares CRL issuer names. Cache several CRL
9725 extensions in X509_CRL structure and cache CRLDP in X509.
9729 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9730 this maps equivalent X509_NAME structures into a consistent structure.
9731 Name comparison can then be performed rapidly using memcmp().
9735 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9740 * Allow digests to supply their own micalg string for S/MIME type using
9741 the ctrl EVP_MD_CTRL_MICALG.
9745 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9746 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9747 ctrl. It can then customise the structure before and/or after signing
9752 * New function OBJ_add_sigid() to allow application defined signature OIDs
9753 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9754 to free up any added signature OIDs.
9758 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9759 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9760 digest and cipher tables. New options added to openssl utility:
9761 list-message-digest-algorithms and list-cipher-algorithms.
9765 * Change the array representation of binary polynomials: the list
9766 of degrees of non-zero coefficients is now terminated with -1.
9767 Previously it was terminated with 0, which was also part of the
9768 value; thus, the array representation was not applicable to
9769 polynomials where t^0 has coefficient zero. This change makes
9770 the array representation useful in a more general context.
9774 * Various modifications and fixes to SSL/TLS cipher string
9775 handling. For ECC, the code now distinguishes between fixed ECDH
9776 with RSA certificates on the one hand and with ECDSA certificates
9777 on the other hand, since these are separate ciphersuites. The
9778 unused code for Fortezza ciphersuites has been removed.
9780 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9781 (not "ECDHE"). For consistency with the code for DH
9782 certificates, use of ECDH certificates is now considered ECDH
9783 authentication, not RSA or ECDSA authentication (the latter is
9784 merely the CA's signing algorithm and not actively used in the
9787 The temporary ciphersuite alias "ECCdraft" is no longer
9788 available, and ECC ciphersuites are no longer excluded from "ALL"
9789 and "DEFAULT". The following aliases now exist for RFC 4492
9790 ciphersuites, most of these by analogy with the DH case:
9792 kECDHr - ECDH cert, signed with RSA
9793 kECDHe - ECDH cert, signed with ECDSA
9794 kECDH - ECDH cert (signed with either RSA or ECDSA)
9795 kEECDH - ephemeral ECDH
9796 ECDH - ECDH cert or ephemeral ECDH
9802 AECDH - anonymous ECDH
9803 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9807 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9808 Use correct micalg parameters depending on digest(s) in signed message.
9812 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9813 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9817 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9818 an engine to register a method. Add ENGINE lookups for methods and
9819 functional reference processing.
9823 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9824 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9829 * New -resign option to smime utility. This adds one or more signers
9830 to an existing PKCS#7 signedData structure. Also -md option to use an
9831 alternative message digest algorithm for signing.
9835 * Tidy up PKCS#7 routines and add new functions to make it easier to
9836 create PKCS7 structures containing multiple signers. Update smime
9837 application to support multiple signers.
9841 * New -macalg option to pkcs12 utility to allow setting of an alternative
9846 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9847 Reorganize PBE internals to lookup from a static table using NIDs,
9848 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9849 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9850 PRF which will be automatically used with PBES2.
9854 * Replace the algorithm specific calls to generate keys in "req" with the
9859 * Update PKCS#7 enveloped data routines to use new API. This is now
9860 supported by any public key method supporting the encrypt operation. A
9861 ctrl is added to allow the public key algorithm to examine or modify
9862 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9867 * Add a ctrl to asn1 method to allow a public key algorithm to express
9868 a default digest type to use. In most cases this will be SHA1 but some
9869 algorithms (such as GOST) need to specify an alternative digest. The
9870 return value indicates how strong the preference is 1 means optional and
9871 2 is mandatory (that is it is the only supported type). Modify
9872 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9873 use the default md. Update openssl utilities to use the default digest
9874 type for signing if it is not explicitly indicated.
9878 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9879 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9880 signing method from the key type. This effectively removes the link
9881 between digests and public key types.
9885 * Add an OID cross reference table and utility functions. Its purpose is to
9886 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9887 rsaEncryption. This will allow some of the algorithm specific hackery
9888 needed to use the correct OID to be removed.
9892 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9893 structures for PKCS7_sign(). They are now set up by the relevant public
9898 * Add provisional EC pkey method with support for ECDSA and ECDH.
9902 * Add support for key derivation (agreement) in the API, DH method and
9907 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9908 public and private key formats. As a side effect these add additional
9909 command line functionality not previously available: DSA signatures can be
9910 generated and verified using pkeyutl and DH key support and generation in
9917 *Oliver Tappe <zooey@hirschkaefer.de>*
9919 * New make target "install_html_docs" installs HTML renditions of the
9922 *Oliver Tappe <zooey@hirschkaefer.de>*
9924 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9925 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9926 support key and parameter generation and add initial key generation
9927 functionality for RSA.
9931 * Add functions for main EVP_PKEY_method operations. The undocumented
9932 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9933 `EVP_PKEY_{encrypt,decrypt}_old`.
9937 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9938 key API, doesn't do much yet.
9942 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9943 public key algorithms. New option to openssl utility:
9944 "list-public-key-algorithms" to print out info.
9948 * Implement the Supported Elliptic Curves Extension for
9949 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9953 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9954 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9958 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9959 utilities such as rsa, dsa, dsaparam etc except they process any key
9964 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9965 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9966 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9971 * Initial support for pluggable public key ASN1.
9972 De-spaghettify the public key ASN1 handling. Move public and private
9973 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9974 algorithm specific handling to a single module within the relevant
9975 algorithm directory. Add functions to allow (near) opaque processing
9976 of public and private key structures.
9980 * Implement the Supported Point Formats Extension for
9981 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9985 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9986 for the psk identity [hint] and the psk callback functions to the
9987 SSL_SESSION, SSL and SSL_CTX structure.
9990 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9994 SSL_CTX_use_psk_identity_hint
9995 SSL_get_psk_identity_hint
9996 SSL_get_psk_identity
9997 SSL_use_psk_identity_hint
9999 *Mika Kousa and Pasi Eronen of Nokia Corporation*
10001 * Add RFC 3161 compliant time stamp request creation, response generation
10002 and response verification functionality.
10004 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
10006 * Add initial support for TLS extensions, specifically for the server_name
10007 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10008 have new members for a hostname. The SSL data structure has an
10009 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10010 stored in that context to allow for session resumption, even after the
10011 SSL has been switched to a new SSL_CTX in reaction to a client's
10012 server_name extension.
10014 New functions (subject to change):
10016 SSL_get_servername()
10017 SSL_get_servername_type()
10020 New CTRL codes and macros (subject to change):
10022 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10023 - SSL_CTX_set_tlsext_servername_callback()
10024 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10025 - SSL_CTX_set_tlsext_servername_arg()
10026 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10028 openssl s_client has a new '-servername ...' option.
10030 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10031 '-key2 ...', '-servername_fatal' (subject to change). This allows
10032 testing the HostName extension for a specific single hostname ('-cert'
10033 and '-key' remain fallbacks for handshakes without HostName
10034 negotiation). If the unrecognized_name alert has to be sent, this by
10035 default is a warning; it becomes fatal with the '-servername_fatal'
10038 *Peter Sylvester, Remy Allais, Christophe Renou*
10040 * Whirlpool hash implementation is added.
10044 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10045 bn(64,32). Because of instruction set limitations it doesn't have
10046 any negative impact on performance. This was done mostly in order
10047 to make it possible to share assembler modules, such as bn_mul_mont
10048 implementations, between 32- and 64-bit builds without hassle.
10052 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10053 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10058 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10059 dedicated Montgomery multiplication procedure, is introduced.
10060 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10061 "64-bit" performance on certain 32-bit targets.
10065 * New option SSL_OP_NO_COMP to disable use of compression selectively
10066 in SSL structures. New SSL ctrl to set maximum send fragment size.
10067 Save memory by setting the I/O buffer sizes dynamically instead of
10068 using the maximum available value.
10072 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10073 in addition to the text details.
10077 * Very, very preliminary EXPERIMENTAL support for printing of general
10078 ASN1 structures. This currently produces rather ugly output and doesn't
10079 handle several customised structures at all.
10083 * Integrated support for PVK file format and some related formats such
10084 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10085 these in the 'rsa' and 'dsa' utilities.
10089 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10093 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10094 place for the (very old) "NETSCAPE" format certificates which are now
10095 handled using new ASN1 code equivalents.
10099 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10100 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10101 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10105 * Modify CRL distribution points extension code to print out previously
10106 unsupported fields. Enhance extension setting code to allow setting of
10111 * Add print and set support for Issuing Distribution Point CRL extension.
10115 * Change 'Configure' script to enable Camellia by default.
10122 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10124 * When rejecting SSL/TLS records due to an incorrect version number, never
10125 update s->server with a new major version number. As of
10126 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10127 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10128 the previous behavior could result in a read attempt at NULL when
10129 receiving specific incorrect SSL/TLS records once record payload
10130 protection is active. ([CVE-2010-0740])
10132 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10134 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10135 could be crashed if the relevant tables were not present (e.g. chrooted).
10137 *Tomas Hoger <thoger@redhat.com>*
10139 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10141 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10143 *Martin Olsson, Neel Mehta*
10145 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10146 accommodate for stack sorting, always a write lock!).
10150 * On some versions of WIN32 Heap32Next is very slow. This can cause
10151 excessive delays in the RAND_poll(): over a minute. As a workaround
10152 include a time check in the inner Heap32Next loop too.
10156 * The code that handled flushing of data in SSL/TLS originally used the
10157 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10158 the problem outlined in PR#1949. The fix suggested there however can
10159 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10160 of Apache). So instead simplify the code to flush unconditionally.
10161 This should be fine since flushing with no data to flush is a no op.
10165 * Handle TLS versions 2.0 and later properly and correctly use the
10166 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10167 off ancient servers have a habit of sticking around for a while...
10171 * Modify compression code so it frees up structures without using the
10172 ex_data callbacks. This works around a problem where some applications
10173 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10174 restarting) then use compression (e.g. SSL with compression) later.
10175 This results in significant per-connection memory leaks and
10176 has caused some security issues including CVE-2008-1678 and
10181 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10182 change when encrypting or decrypting.
10186 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10187 connect and renegotiate with servers which do not support RI.
10188 Until RI is more widely deployed this option is enabled by default.
10192 * Add "missing" ssl ctrls to clear options and mode.
10196 * If client attempts to renegotiate and doesn't support RI respond with
10197 a no_renegotiation alert as required by RFC5746. Some renegotiating
10198 TLS clients will continue a connection gracefully when they receive
10199 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10200 waiting for a server hello which it will never receive. Now we treat a
10201 received no_renegotiation alert as a fatal error. This is because
10202 applications requesting a renegotiation might well expect it to succeed
10203 and would have no code in place to handle the server denying it so the
10204 only safe thing to do is to terminate the connection.
10208 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10209 peer supports secure renegotiation and 0 otherwise. Print out peer
10210 renegotiation support in s_client/s_server.
10214 * Replace the highly broken and deprecated SPKAC certification method with
10215 the updated NID creation version. This should correctly handle UTF8.
10219 * Implement RFC5746. Re-enable renegotiation but require the extension
10220 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10221 turns out to be a bad idea. It has been replaced by
10222 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10223 SSL_CTX_set_options(). This is really not recommended unless you
10224 know what you are doing.
10226 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10228 * Fixes to stateless session resumption handling. Use initial_ctx when
10229 issuing and attempting to decrypt tickets in case it has changed during
10230 servername handling. Use a non-zero length session ID when attempting
10231 stateless session resumption: this makes it possible to determine if
10232 a resumption has occurred immediately after receiving server hello
10233 (several places in OpenSSL subtly assume this) instead of later in
10238 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10239 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10240 fixes for a few places where the return code is not checked
10243 *Julia Lawall <julia@diku.dk>*
10245 * Add --strict-warnings option to Configure script to include devteam
10246 warnings in other configurations.
10250 * Add support for --libdir option and LIBDIR variable in makefiles. This
10251 makes it possible to install openssl libraries in locations which
10252 have names other than "lib", for example "/usr/lib64" which some
10255 *Steve Henson, based on patch from Jeremy Utley*
10257 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10258 X690 8.9.12 and can produce some misleading textual output of OIDs.
10260 *Steve Henson, reported by Dan Kaminsky*
10262 * Delete MD2 from algorithm tables. This follows the recommendation in
10263 several standards that it is not used in new applications due to
10264 several cryptographic weaknesses. For binary compatibility reasons
10265 the MD2 API is still compiled in by default.
10269 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10274 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10275 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10278 *Guenter <lists@gknw.net>*
10280 * Fix the server certificate chain building code to use X509_verify_cert(),
10281 it used to have an ad-hoc builder which was unable to cope with anything
10282 other than a simple chain.
10284 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10286 * Don't check self signed certificate signatures in X509_verify_cert()
10287 by default (a flag can override this): it just wastes time without
10288 adding any security. As a useful side effect self signed root CAs
10289 with non-FIPS digests are now usable in FIPS mode.
10293 * In dtls1_process_out_of_seq_message() the check if the current message
10294 is already buffered was missing. For every new message was memory
10295 allocated, allowing an attacker to perform an denial of service attack
10296 with sending out of seq handshake messages until there is no memory
10297 left. Additionally every future message was buffered, even if the
10298 sequence number made no sense and would be part of another handshake.
10299 So only messages with sequence numbers less than 10 in advance will be
10300 buffered. ([CVE-2009-1378])
10302 *Robin Seggelmann, discovered by Daniel Mentz*
10304 * Records are buffered if they arrive with a future epoch to be
10305 processed after finishing the corresponding handshake. There is
10306 currently no limitation to this buffer allowing an attacker to perform
10307 a DOS attack with sending records with future epochs until there is no
10308 memory left. This patch adds the pqueue_size() function to determine
10309 the size of a buffer and limits the record buffer to 100 entries.
10312 *Robin Seggelmann, discovered by Daniel Mentz*
10314 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10315 parent structure is freed. ([CVE-2009-1379])
10319 * Handle non-blocking I/O properly in SSL_shutdown() call.
10321 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10323 * Add `2.5.4.*` OIDs
10325 *Ilya O. <vrghost@gmail.com>*
10327 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10329 * Disable renegotiation completely - this fixes a severe security
10330 problem ([CVE-2009-3555]) at the cost of breaking all
10331 renegotiation. Renegotiation can be re-enabled by setting
10332 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10333 run-time. This is really not recommended unless you know what
10338 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10340 * Don't set val to NULL when freeing up structures, it is freed up by
10341 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10342 zeroing past the valid field. ([CVE-2009-0789])
10344 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10346 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10347 checked correctly. This would allow some invalid signed attributes to
10348 appear to verify correctly. ([CVE-2009-0591])
10350 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10352 * Reject UniversalString and BMPString types with invalid lengths. This
10353 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10354 a legal length. ([CVE-2009-0590])
10358 * Set S/MIME signing as the default purpose rather than setting it
10359 unconditionally. This allows applications to override it at the store
10364 * Permit restricted recursion of ASN1 strings. This is needed in practice
10365 to handle some structures.
10369 * Improve efficiency of mem_gets: don't search whole buffer each time
10372 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10374 * New -hex option for openssl rand.
10378 * Print out UTF8String and NumericString when parsing ASN1.
10382 * Support NumericString type for name components.
10386 * Allow CC in the environment to override the automatically chosen
10387 compiler. Note that nothing is done to ensure flags work with the
10392 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10394 * Properly check EVP_VerifyFinal() and similar return values
10397 *Ben Laurie, Bodo Moeller, Google Security Team*
10399 * Enable TLS extensions by default.
10403 * Allow the CHIL engine to be loaded, whether the application is
10404 multithreaded or not. (This does not release the developer from the
10405 obligation to set up the dynamic locking callbacks.)
10407 *Sander Temme <sander@temme.net>*
10409 * Use correct exit code if there is an error in dgst command.
10411 *Steve Henson; problem pointed out by Roland Dirlewanger*
10413 * Tweak Configure so that you need to say "experimental-jpake" to enable
10414 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10418 * Add experimental JPAKE support, including demo authentication in
10419 s_client and s_server.
10423 * Set the comparison function in v3_addr_canonize().
10425 *Rob Austein <sra@hactrn.net>*
10427 * Add support for XMPP STARTTLS in s_client.
10429 *Philip Paeps <philip@freebsd.org>*
10431 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10432 to ensure that even with this option, only ciphersuites in the
10433 server's preference list will be accepted. (Note that the option
10434 applies only when resuming a session, so the earlier behavior was
10435 just about the algorithm choice for symmetric cryptography.)
10439 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10441 * Fix NULL pointer dereference if a DTLS server received
10442 ChangeCipherSpec as first record ([CVE-2009-1386]).
10446 * Fix a state transition in s3_srvr.c and d1_srvr.c
10447 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10449 *Nagendra Modadugu*
10451 * The fix in 0.9.8c that supposedly got rid of unsafe
10452 double-checked locking was incomplete for RSA blinding,
10453 addressing just one layer of what turns out to have been
10454 doubly unsafe triple-checked locking.
10456 So now fix this for real by retiring the MONT_HELPER macro
10457 in crypto/rsa/rsa_eay.c.
10459 *Bodo Moeller; problem pointed out by Marius Schilder*
10461 * Various precautionary measures:
10463 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10465 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10466 (NB: This would require knowledge of the secret session ticket key
10467 to exploit, in which case you'd be SOL either way.)
10469 - Change bn_nist.c so that it will properly handle input BIGNUMs
10470 outside the expected range.
10472 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10475 *Neel Mehta, Bodo Moeller*
10477 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10478 the load fails. Useful for distros.
10480 *Ben Laurie and the FreeBSD team*
10482 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10486 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10490 * Expand ENGINE to support engine supplied SSL client certificate functions.
10492 This work was sponsored by Logica.
10496 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10497 keystores. Support for SSL/TLS client authentication too.
10498 Not compiled unless enable-capieng specified to Configure.
10500 This work was sponsored by Logica.
10504 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10505 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10506 attribute creation routines such as certificate requests and PKCS#12
10511 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10513 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10514 handshake which could lead to a client crash as found using the
10515 Codenomicon TLS test suite ([CVE-2008-1672])
10517 *Steve Henson, Mark Cox*
10519 * Fix double free in TLS server name extensions which could lead to
10520 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10524 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10526 Clear the error queue to ensure that error entries left from
10527 older function calls do not interfere with the correct operation.
10529 *Lutz Jaenicke, Erik de Castro Lopo*
10531 * Remove root CA certificates of commercial CAs:
10533 The OpenSSL project does not recommend any specific CA and does not
10534 have any policy with respect to including or excluding any CA.
10535 Therefore, it does not make any sense to ship an arbitrary selection
10536 of root CA certificates with the OpenSSL software.
10540 * RSA OAEP patches to fix two separate invalid memory reads.
10541 The first one involves inputs when 'lzero' is greater than
10542 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10543 before the beginning of from). The second one involves inputs where
10544 the 'db' section contains nothing but zeroes (there is a one-byte
10545 invalid read after the end of 'db').
10547 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10549 * Partial backport from 0.9.9-dev:
10551 Introduce bn_mul_mont (dedicated Montgomery multiplication
10552 procedure) as a candidate for BIGNUM assembler implementation.
10553 While 0.9.9-dev uses assembler for various architectures, only
10554 x86_64 is available by default here in the 0.9.8 branch, and
10555 32-bit x86 is available through a compile-time setting.
10557 To try the 32-bit x86 assembler implementation, use Configure
10558 option "enable-montasm" (which exists only for this backport).
10560 As "enable-montasm" for 32-bit x86 disclaims code stability
10561 anyway, in this constellation we activate additional code
10562 backported from 0.9.9-dev for further performance improvements,
10563 namely BN_from_montgomery_word. (To enable this otherwise,
10564 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10566 *Andy Polyakov (backport partially by Bodo Moeller)*
10568 * Add TLS session ticket callback. This allows an application to set
10569 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10570 values. This is useful for key rollover for example where several key
10571 sets may exist with different names.
10575 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10576 This was broken until now in 0.9.8 releases, such that the only way
10577 a registered ENGINE could be used (assuming it initialises
10578 successfully on the host) was to explicitly set it as the default
10579 for the relevant algorithms. This is in contradiction with 0.9.7
10580 behaviour and the documentation. With this fix, when an ENGINE is
10581 registered into a given algorithm's table of implementations, the
10582 'uptodate' flag is reset so that auto-discovery will be used next
10583 time a new context for that algorithm attempts to select an
10586 *Ian Lister (tweaked by Geoff Thorpe)*
10588 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10589 implementation in the following ways:
10591 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10594 Lack of BER streaming support means one pass streaming processing is
10595 only supported if data is detached: setting the streaming flag is
10596 ignored for embedded content.
10598 CMS support is disabled by default and must be explicitly enabled
10599 with the enable-cms configuration option.
10603 * Update the GMP engine glue to do direct copies between BIGNUM and
10604 mpz_t when openssl and GMP use the same limb size. Otherwise the
10605 existing "conversion via a text string export" trick is still used.
10607 *Paul Sheer <paulsheer@gmail.com>*
10609 * Zlib compression BIO. This is a filter BIO which compressed and
10610 uncompresses any data passed through it.
10614 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10615 RFC3394 compatible AES key wrapping.
10619 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10620 sets string data without copying. X509_ALGOR_set0() and
10621 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10622 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10623 from an X509_ATTRIBUTE structure optionally checking it occurs only
10624 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10629 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10630 to get the expected BN_FLG_CONSTTIME behavior.
10632 *Bodo Moeller (Google)*
10636 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10637 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10638 - added some more tests to do_tests.pl
10639 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10640 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10641 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10642 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10643 - various changes to netware.pl to enable gcc-cross builds on Win32
10645 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10646 - various changes to fix missing prototype warnings
10647 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10648 - added AES, WHIRLPOOL and CPUID assembler code to build files
10649 - added missing AES assembler make rules to mk1mf.pl
10650 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10652 *Guenter Knauf <eflash@gmx.net>*
10654 * Implement certificate status request TLS extension defined in RFC3546.
10655 A client can set the appropriate parameters and receive the encoded
10656 OCSP response via a callback. A server can query the supplied parameters
10657 and set the encoded OCSP response in the callback. Add simplified examples
10658 to s_client and s_server.
10662 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10664 * Fix various bugs:
10665 + Binary incompatibility of ssl_ctx_st structure
10666 + DTLS interoperation with non-compliant servers
10667 + Don't call get_session_cb() without proposed session
10668 + Fix ia64 assembler code
10670 *Andy Polyakov, Steve Henson*
10672 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10674 * DTLS Handshake overhaul. There were longstanding issues with
10675 OpenSSL DTLS implementation, which were making it impossible for
10676 RFC 4347 compliant client to communicate with OpenSSL server.
10677 Unfortunately just fixing these incompatibilities would "cut off"
10678 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10679 server keeps tolerating non RFC compliant syntax. The opposite is
10680 not true, 0.9.8f client can not communicate with earlier server.
10681 This update even addresses CVE-2007-4995.
10685 * Changes to avoid need for function casts in OpenSSL: some compilers
10686 (gcc 4.2 and later) reject their use.
10687 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10690 * Add RFC4507 support to OpenSSL. This includes the corrections in
10691 RFC4507bis. The encrypted ticket format is an encrypted encoded
10692 SSL_SESSION structure, that way new session features are automatically
10695 If a client application caches session in an SSL_SESSION structure
10696 support is transparent because tickets are now stored in the encoded
10699 The SSL_CTX structure automatically generates keys for ticket
10700 protection in servers so again support should be possible
10701 with no application modification.
10703 If a client or server wishes to disable RFC4507 support then the option
10704 SSL_OP_NO_TICKET can be set.
10706 Add a TLS extension debugging callback to allow the contents of any client
10707 or server extensions to be examined.
10709 This work was sponsored by Google.
10713 * Add initial support for TLS extensions, specifically for the server_name
10714 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10715 have new members for a hostname. The SSL data structure has an
10716 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10717 stored in that context to allow for session resumption, even after the
10718 SSL has been switched to a new SSL_CTX in reaction to a client's
10719 server_name extension.
10721 New functions (subject to change):
10723 SSL_get_servername()
10724 SSL_get_servername_type()
10727 New CTRL codes and macros (subject to change):
10729 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10730 - SSL_CTX_set_tlsext_servername_callback()
10731 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10732 - SSL_CTX_set_tlsext_servername_arg()
10733 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10735 openssl s_client has a new '-servername ...' option.
10737 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10738 '-key2 ...', '-servername_fatal' (subject to change). This allows
10739 testing the HostName extension for a specific single hostname ('-cert'
10740 and '-key' remain fallbacks for handshakes without HostName
10741 negotiation). If the unrecognized_name alert has to be sent, this by
10742 default is a warning; it becomes fatal with the '-servername_fatal'
10745 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10747 * Add AES and SSE2 assembly language support to VC++ build.
10751 * Mitigate attack on final subtraction in Montgomery reduction.
10755 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10756 (which previously caused an internal error).
10760 * Squeeze another 10% out of IGE mode when in != out.
10764 * AES IGE mode speedup.
10766 *Dean Gaudet (Google)*
10768 * Add the Korean symmetric 128-bit cipher SEED (see
10769 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10770 add SEED ciphersuites from RFC 4162:
10772 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10773 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10774 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10775 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10777 To minimize changes between patchlevels in the OpenSSL 0.9.8
10778 series, SEED remains excluded from compilation unless OpenSSL
10779 is configured with 'enable-seed'.
10781 *KISA, Bodo Moeller*
10783 * Mitigate branch prediction attacks, which can be practical if a
10784 single processor is shared, allowing a spy process to extract
10785 information. For detailed background information, see
10786 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10787 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10788 and Necessary Software Countermeasures"). The core of the change
10789 are new versions BN_div_no_branch() and
10790 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10791 respectively, which are slower, but avoid the security-relevant
10792 conditional branches. These are automatically called by BN_div()
10793 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10794 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10795 remove a conditional branch.
10797 BN_FLG_CONSTTIME is the new name for the previous
10798 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10799 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10800 in the exponent causes BN_mod_exp_mont() to use the alternative
10801 implementation in BN_mod_exp_mont_consttime().) The old name
10802 remains as a deprecated alias.
10804 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10805 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10806 constant-time implementations for more than just exponentiation.
10807 Here too the old name is kept as a deprecated alias.
10809 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10810 the BN_BLINDING structure gets an independent copy of the
10811 modulus. This means that the previous `BIGNUM *m` argument to
10812 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10813 essentially becomes `const BIGNUM *m`, although we can't actually
10814 change this in the header file before 0.9.9. It allows
10815 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10816 enable BN_FLG_CONSTTIME.
10818 *Matthew D Wood (Intel Corp)*
10820 * In the SSL/TLS server implementation, be strict about session ID
10821 context matching (which matters if an application uses a single
10822 external cache for different purposes). Previously,
10823 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10824 set. This did ensure strict client verification, but meant that,
10825 with applications using a single external cache for quite
10826 different requirements, clients could circumvent ciphersuite
10827 restrictions for a given session ID context by starting a session
10828 in a different context.
10832 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10833 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10834 authentication-only ciphersuites.
10838 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10839 not complete and could lead to a possible single byte overflow
10840 ([CVE-2007-5135]) [Ben Laurie]
10842 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10844 * Since AES128 and AES256 (and similarly Camellia128 and
10845 Camellia256) share a single mask bit in the logic of
10846 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10847 kludge to work properly if AES128 is available and AES256 isn't
10848 (or if Camellia128 is available and Camellia256 isn't).
10852 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10853 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10854 When a point or a seed is encoded in a BIT STRING, we need to
10855 prevent the removal of trailing zero bits to get the proper DER
10856 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10857 of a NamedBitList, for which trailing 0 bits need to be removed.)
10861 * Have SSL/TLS server implementation tolerate "mismatched" record
10862 protocol version while receiving ClientHello even if the
10863 ClientHello is fragmented. (The server can't insist on the
10864 particular protocol version it has chosen before the ServerHello
10865 message has informed the client about his choice.)
10869 * Add RFC 3779 support.
10871 *Rob Austein for ARIN, Ben Laurie*
10873 * Load error codes if they are not already present instead of using a
10874 static variable. This allows them to be cleanly unloaded and reloaded.
10875 Improve header file function name parsing.
10879 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10880 or CAPABILITY handshake as required by RFCs.
10882 *Goetz Babin-Ebell*
10884 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10886 * Introduce limits to prevent malicious keys being able to
10887 cause a denial of service. ([CVE-2006-2940])
10889 *Steve Henson, Bodo Moeller*
10891 * Fix ASN.1 parsing of certain invalid structures that can result
10892 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10894 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10895 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10897 * Fix SSL client code which could crash if connecting to a
10898 malicious SSLv2 server. ([CVE-2006-4343])
10900 *Tavis Ormandy and Will Drewry, Google Security Team*
10902 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10903 match only those. Before that, "AES256-SHA" would be interpreted
10904 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10905 the same strength classification in 0.9.7h) as we currently only
10906 have a single AES bit in the ciphersuite description bitmap.
10907 That change, however, also applied to ciphersuite strings such as
10908 "RC4-MD5" that intentionally matched multiple ciphersuites --
10909 namely, SSL 2.0 ciphersuites in addition to the more common ones
10910 from SSL 3.0/TLS 1.0.
10912 So we change the selection algorithm again: Naming an explicit
10913 ciphersuite selects this one ciphersuite, and any other similar
10914 ciphersuite (same bitmap) from *other* protocol versions.
10915 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10916 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10918 Since SSL 2.0 does not have any ciphersuites for which the
10919 128/256 bit distinction would be relevant, this works for now.
10920 The proper fix will be to use different bits for AES128 and
10921 AES256, which would have avoided the problems from the beginning;
10922 however, bits are scarce, so we can only do this in a new release
10923 (not just a patchlevel) when we can change the SSL_CIPHER
10924 definition to split the single 'unsigned long mask' bitmap into
10925 multiple values to extend the available space.
10929 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10931 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10932 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10934 * Add AES IGE and biIGE modes.
10938 * Change the Unix randomness entropy gathering to use poll() when
10939 possible instead of select(), since the latter has some
10940 undesirable limitations.
10942 *Darryl Miles via Richard Levitte and Bodo Moeller*
10944 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10945 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10946 cannot be implicitly activated as part of, e.g., the "AES" alias.
10947 However, please upgrade to OpenSSL 0.9.9[-dev] for
10948 non-experimental use of the ECC ciphersuites to get TLS extension
10949 support, which is required for curve and point format negotiation
10950 to avoid potential handshake problems.
10954 * Disable rogue ciphersuites:
10956 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10957 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10958 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10960 The latter two were purportedly from
10961 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10964 Also deactivate the remaining ciphersuites from
10965 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10966 unofficial, and the ID has long expired.
10970 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10971 dual-core machines) and other potential thread-safety issues.
10975 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10976 versions), which is now available for royalty-free use
10977 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10978 Also, add Camellia TLS ciphersuites from RFC 4132.
10980 To minimize changes between patchlevels in the OpenSSL 0.9.8
10981 series, Camellia remains excluded from compilation unless OpenSSL
10982 is configured with 'enable-camellia'.
10986 * Disable the padding bug check when compression is in use. The padding
10987 bug check assumes the first packet is of even length, this is not
10988 necessarily true if compression is enabled and can result in false
10989 positives causing handshake failure. The actual bug test is ancient
10990 code so it is hoped that implementations will either have fixed it by
10991 now or any which still have the bug do not support compression.
10995 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10997 * When applying a cipher rule check to see if string match is an explicit
10998 cipher suite and only match that one cipher suite if it is.
11002 * Link in manifests for VC++ if needed.
11004 *Austin Ziegler <halostatue@gmail.com>*
11006 * Update support for ECC-based TLS ciphersuites according to
11007 draft-ietf-tls-ecc-12.txt with proposed changes (but without
11008 TLS extensions, which are supported starting with the 0.9.9
11009 branch, not in the OpenSSL 0.9.8 branch).
11013 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
11014 opaque EVP_CIPHER_CTX handling.
11018 * Fixes and enhancements to zlib compression code. We now only use
11019 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
11020 to conform with the standards mentioned here:
11021 <http://www.zlib.net/DLL_FAQ.txt>
11022 Static zlib linking now works on Windows and the new --with-zlib-include
11023 --with-zlib-lib options to Configure can be used to supply the location
11024 of the headers and library. Gracefully handle case where zlib library
11029 * Several fixes and enhancements to the OID generation code. The old code
11030 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11031 handle numbers larger than ULONG_MAX, truncated printing and had a
11032 non standard OBJ_obj2txt() behaviour.
11036 * Add support for building of engines under engine/ as shared libraries
11037 under VC++ build system.
11041 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11042 Hopefully, we will not see any false combination of paths any more.
11046 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11048 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11049 (part of SSL_OP_ALL). This option used to disable the
11050 countermeasure against man-in-the-middle protocol-version
11051 rollback in the SSL 2.0 server implementation, which is a bad
11052 idea. ([CVE-2005-2969])
11054 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11055 for Information Security, National Institute of Advanced Industrial
11056 Science and Technology [AIST], Japan)*
11058 * Add two function to clear and return the verify parameter flags.
11062 * Keep cipherlists sorted in the source instead of sorting them at
11063 runtime, thus removing the need for a lock.
11067 * Avoid some small subgroup attacks in Diffie-Hellman.
11069 *Nick Mathewson and Ben Laurie*
11071 * Add functions for well-known primes.
11075 * Extended Windows CE support.
11077 *Satoshi Nakamura and Andy Polyakov*
11079 * Initialize SSL_METHOD structures at compile time instead of during
11080 runtime, thus removing the need for a lock.
11084 * Make PKCS7_decrypt() work even if no certificate is supplied by
11085 attempting to decrypt each encrypted key in turn. Add support to
11090 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11092 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11095 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11099 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11100 key into the same file any more.
11104 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11108 * Add -utf8 command line and config file option to 'ca'.
11110 *Stefan <stf@udoma.org*
11112 * Removed the macro des_crypt(), as it seems to conflict with some
11113 libraries. Use DES_crypt().
11117 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11118 involves renaming the source and generated shared-libs for
11119 both. The engines will accept the corrected or legacy ids
11120 ('ncipher' and '4758_cca' respectively) when binding. NB,
11121 this only applies when building 'shared'.
11123 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11125 * Add attribute functions to EVP_PKEY structure. Modify
11126 PKCS12_create() to recognize a CSP name attribute and
11127 use it. Make -CSP option work again in pkcs12 utility.
11131 * Add new functionality to the bn blinding code:
11132 - automatic re-creation of the BN_BLINDING parameters after
11133 a fixed number of uses (currently 32)
11134 - add new function for parameter creation
11135 - introduce flags to control the update behaviour of the
11136 BN_BLINDING parameters
11137 - hide BN_BLINDING structure
11138 Add a second BN_BLINDING slot to the RSA structure to improve
11139 performance when a single RSA object is shared among several
11144 * Add support for DTLS.
11146 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11148 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11149 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11153 * Remove buggy and incomplete DH cert support from
11154 ssl/ssl_rsa.c and ssl/s3_both.c
11158 * Use SHA-1 instead of MD5 as the default digest algorithm for
11159 the `apps/openssl` commands.
11163 * Compile clean with "-Wall -Wmissing-prototypes
11164 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11165 DEBUG_SAFESTACK must also be set.
11169 * Change ./Configure so that certain algorithms can be disabled by default.
11170 The new counterpiece to "no-xxx" is "enable-xxx".
11172 The patented RC5 and MDC2 algorithms will now be disabled unless
11173 "enable-rc5" and "enable-mdc2", respectively, are specified.
11175 (IDEA remains enabled despite being patented. This is because IDEA
11176 is frequently required for interoperability, and there is no license
11177 fee for non-commercial use. As before, "no-idea" can be used to
11178 avoid this algorithm.)
11182 * Add processing of proxy certificates (see RFC 3820). This work was
11183 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11184 EGEE (Enabling Grids for E-science in Europe).
11188 * RC4 performance overhaul on modern architectures/implementations, such
11189 as Intel P4, IA-64 and AMD64.
11193 * New utility extract-section.pl. This can be used specify an alternative
11194 section number in a pod file instead of having to treat each file as
11195 a separate case in Makefile. This can be done by adding two lines to the
11198 =for comment openssl_section:XXX
11200 The blank line is mandatory.
11204 * New arguments -certform, -keyform and -pass for s_client and s_server
11205 to allow alternative format key and certificate files and passphrase
11210 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11211 update associated structures and add various utility functions.
11213 Add new policy related verify parameters, include policy checking in
11214 standard verify code. Enhance 'smime' application with extra parameters
11215 to support policy checking and print out.
11219 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11220 Nehemiah processors. These extensions support AES encryption in hardware
11221 as well as RNG (though RNG support is currently disabled).
11223 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11225 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11229 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11231 *Andy Polyakov and a number of other people*
11233 * Improved PowerPC platform support. Most notably BIGNUM assembler
11234 implementation contributed by IBM.
11236 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11238 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11239 exponent rather than 'unsigned long'. There is a corresponding change to
11240 the new 'rsa_keygen' element of the RSA_METHOD structure.
11242 *Jelte Jansen, Geoff Thorpe*
11244 * Functionality for creating the initial serial number file is now
11245 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11247 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11248 number file to 1, which is bound to cause problems. To avoid
11249 the problems while respecting compatibility between different 0.9.7
11250 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11251 CA.pl for serial number initialization. With the new release 0.9.8,
11252 we can fix the problem directly in the 'ca' utility.)
11256 * Reduced header interdependencies by declaring more opaque objects in
11257 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11258 give fewer recursive includes, which could break lazy source code - so
11259 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11260 developers should define this symbol when building and using openssl to
11261 ensure they track the recommended behaviour, interfaces, [etc], but
11262 backwards-compatible behaviour prevails when this isn't defined.
11266 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11270 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11271 This will generate a random key of the appropriate length based on the
11272 cipher context. The EVP_CIPHER can provide its own random key generation
11273 routine to support keys of a specific form. This is used in the des and
11274 3des routines to generate a key of the correct parity. Update S/MIME
11275 code to use new functions and hence generate correct parity DES keys.
11276 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11277 valid (weak or incorrect parity).
11281 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11282 as looking them up. This is useful when the verified structure may contain
11283 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11284 present unless the new PKCS7_NO_CRL flag is asserted.
11288 * Extend ASN1 oid configuration module. It now additionally accepts the
11291 shortName = some long name, 1.2.3.4
11295 * Reimplemented the BN_CTX implementation. There is now no more static
11296 limitation on the number of variables it can handle nor the depth of the
11297 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11298 information can now expand as required, and rather than having a single
11299 static array of bignums, BN_CTX now uses a linked-list of such arrays
11300 allowing it to expand on demand whilst maintaining the usefulness of
11301 BN_CTX's "bundling".
11305 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11306 to allow all RSA operations to function using a single BN_CTX.
11310 * Preliminary support for certificate policy evaluation and checking. This
11311 is initially intended to pass the tests outlined in "Conformance Testing
11312 of Relying Party Client Certificate Path Processing Logic" v1.07.
11316 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11317 remained unused and not that useful. A variety of other little bignum
11318 tweaks and fixes have also been made continuing on from the audit (see
11323 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11324 associated ASN1, EVP and SSL functions and old ASN1 macros.
11328 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11329 and this should never fail. So the return value from the use of
11330 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11331 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11335 * BN_CTX_get() should return zero-valued bignums, providing the same
11336 initialised value as BN_new().
11338 *Geoff Thorpe, suggested by Ulf Möller*
11340 * Support for inhibitAnyPolicy certificate extension.
11344 * An audit of the BIGNUM code is underway, for which debugging code is
11345 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11346 is considered valid when processing BIGNUMs, and causes execution to
11347 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11348 further steps are taken to deliberately pollute unused data in BIGNUM
11349 structures to try and expose faulty code further on. For now, openssl will
11350 (in its default mode of operation) continue to tolerate the inconsistent
11351 forms that it has tolerated in the past, but authors and packagers should
11352 consider trying openssl and their own applications when compiled with
11353 these debugging symbols defined. It will help highlight potential bugs in
11354 their own code, and will improve the test coverage for OpenSSL itself. At
11355 some point, these tighter rules will become openssl's default to improve
11356 maintainability, though the assert()s and other overheads will remain only
11357 in debugging configurations. See bn.h for more details.
11359 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11361 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11362 that can only be obtained through BN_CTX_new() (which implicitly
11363 initialises it). The presence of this function only made it possible
11364 to overwrite an existing structure (and cause memory leaks).
11368 * Because of the callback-based approach for implementing LHASH as a
11369 template type, lh_insert() adds opaque objects to hash-tables and
11370 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11371 to clean up those corresponding objects before destroying the hash table
11372 (and losing the object pointers). So some over-zealous constifications in
11373 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11374 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11375 prototyped to have "const" restrictions on the object pointers they are
11376 given (and so aren't required to cast them away any more).
11380 * The tmdiff.h API was so ugly and minimal that our own timing utility
11381 (speed) prefers to use its own implementation. The two implementations
11382 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11383 its object type properly exposed (MS_TM) instead of casting to/from
11384 `char *`. This may still change yet if someone realises MS_TM and
11386 aren't necessarily the greatest nomenclatures - but this is what was used
11387 internally to the implementation so I've used that for now.
11391 * Ensure that deprecated functions do not get compiled when
11392 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11393 the self-tests were still using deprecated key-generation functions so
11394 these have been updated also.
11398 * Reorganise PKCS#7 code to separate the digest location functionality
11399 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11400 New function PKCS7_set_digest() to set the digest type for PKCS#7
11401 digestedData type. Add additional code to correctly generate the
11402 digestedData type and add support for this type in PKCS7 initialization
11407 * New function PKCS7_set0_type_other() this initializes a PKCS7
11408 structure of type "other".
11412 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11413 sure the loop does correctly stop and breaking ("division by zero")
11414 modulus operations are not performed. The (pre-generated) prime
11415 table crypto/bn/bn_prime.h was already correct, but it could not be
11416 re-generated on some platforms because of the "division by zero"
11417 situation in the script.
11419 *Ralf S. Engelschall*
11421 * Update support for ECC-based TLS ciphersuites according to
11422 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11423 SHA-1 now is only used for "small" curves (where the
11424 representation of a field element takes up to 24 bytes); for
11425 larger curves, the field element resulting from ECDH is directly
11426 used as premaster secret.
11428 *Douglas Stebila (Sun Microsystems Laboratories)*
11430 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11431 curve secp160r1 to the tests.
11433 *Douglas Stebila (Sun Microsystems Laboratories)*
11435 * Add the possibility to load symbols globally with DSO.
11437 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11439 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11440 control of the error stack.
11444 * Add support for STORE in ENGINE.
11448 * Add the STORE type. The intention is to provide a common interface
11449 to certificate and key stores, be they simple file-based stores, or
11450 HSM-type store, or LDAP stores, or...
11451 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11455 * Add a generic structure called OPENSSL_ITEM. This can be used to
11456 pass a list of arguments to any function as well as provide a way
11457 for a function to pass data back to the caller.
11461 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11462 works like BUF_strdup() but can be used to duplicate a portion of
11463 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11468 * Add the function sk_find_ex() which works like sk_find(), but will
11469 return an index to an element even if an exact match couldn't be
11470 found. The index is guaranteed to point at the element where the
11471 searched-for key would be inserted to preserve sorting order.
11475 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11476 takes an extra flags argument for optional functionality. Currently,
11477 the following flags are defined:
11479 OBJ_BSEARCH_VALUE_ON_NOMATCH
11480 This one gets OBJ_bsearch_ex() to return a pointer to the first
11481 element where the comparing function returns a negative or zero
11484 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11485 This one gets OBJ_bsearch_ex() to return a pointer to the first
11486 element where the comparing function returns zero. This is useful
11487 if there are more than one element where the comparing function
11492 * Make it possible to create self-signed certificates with 'openssl ca'
11493 in such a way that the self-signed certificate becomes part of the
11494 CA database and uses the same mechanisms for serial number generation
11495 as all other certificate signing. The new flag '-selfsign' enables
11496 this functionality. Adapt CA.sh and CA.pl.in.
11500 * Add functionality to check the public key of a certificate request
11501 against a given private. This is useful to check that a certificate
11502 request can be signed by that key (self-signing).
11506 * Make it possible to have multiple active certificates with the same
11507 subject in the CA index file. This is done only if the keyword
11508 'unique_subject' is set to 'no' in the main CA section (default
11509 if 'CA_default') of the configuration file. The value is saved
11510 with the database itself in a separate index attribute file,
11511 named like the index file with '.attr' appended to the name.
11515 * Generate multi-valued AVAs using '+' notation in config files for
11520 * Support for nameConstraints certificate extension.
11524 * Support for policyConstraints certificate extension.
11528 * Support for policyMappings certificate extension.
11532 * Make sure the default DSA_METHOD implementation only uses its
11533 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11534 and change its own handlers to be NULL so as to remove unnecessary
11535 indirection. This lets alternative implementations fallback to the
11536 default implementation more easily.
11540 * Support for directoryName in GeneralName related extensions
11545 * Make it possible to link applications using Makefile.shared.
11546 Make that possible even when linking against static libraries!
11550 * Support for single pass processing for S/MIME signing. This now
11551 means that S/MIME signing can be done from a pipe, in addition
11552 cleartext signing (multipart/signed type) is effectively streaming
11553 and the signed data does not need to be all held in memory.
11555 This is done with a new flag PKCS7_STREAM. When this flag is set
11556 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11557 is done after the data is output (and digests calculated) in
11558 SMIME_write_PKCS7().
11562 * Add full support for -rpath/-R, both in shared libraries and
11563 applications, at least on the platforms where it's known how
11568 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11569 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11570 will now compute a table of multiples of the generator that
11571 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11572 faster (notably in the case of a single point multiplication,
11573 scalar * generator).
11575 *Nils Larsch, Bodo Moeller*
11577 * IPv6 support for certificate extensions. The various extensions
11578 which use the IP:a.b.c.d can now take IPv6 addresses using the
11579 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11584 * Added an ENGINE that implements RSA by performing private key
11585 exponentiations with the GMP library. The conversions to and from
11586 GMP's mpz_t format aren't optimised nor are any montgomery forms
11587 cached, and on x86 it appears OpenSSL's own performance has caught up.
11588 However there are likely to be other architectures where GMP could
11589 provide a boost. This ENGINE is not built in by default, but it can be
11590 specified at Configure time and should be accompanied by the necessary
11591 linker additions, eg;
11592 ./config -DOPENSSL_USE_GMP -lgmp
11596 * "openssl engine" will not display ENGINE/DSO load failure errors when
11597 testing availability of engines with "-t" - the old behaviour is
11598 produced by increasing the feature's verbosity with "-tt".
11602 * ECDSA routines: under certain error conditions uninitialized BN objects
11603 could be freed. Solution: make sure initialization is performed early
11604 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11609 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11610 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11611 software implementations. For DSA and DH, parameter generation can
11612 also be overridden by providing the appropriate method callbacks.
11616 * Change the "progress" mechanism used in key-generation and
11617 primality testing to functions that take a new BN_GENCB pointer in
11618 place of callback/argument pairs. The new API functions have `_ex`
11619 postfixes and the older functions are reimplemented as wrappers for
11620 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11621 declarations of the old functions to help (graceful) attempts to
11622 migrate to the new functions. Also, the new key-generation API
11623 functions operate on a caller-supplied key-structure and return
11624 success/failure rather than returning a key or NULL - this is to
11625 help make "keygen" another member function of RSA_METHOD etc.
11627 Example for using the new callback interface:
11629 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11630 void *my_arg = ...;
11633 BN_GENCB_set(&my_cb, my_callback, my_arg);
11635 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11636 /* For the meaning of a, b in calls to my_callback(), see the
11637 * documentation of the function that calls the callback.
11638 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11639 * my_callback should return 1 if it wants BN_is_prime_ex()
11640 * to continue, or 0 to stop.
11645 * Change the ZLIB compression method to be stateful, and make it
11646 available to TLS with the number defined in
11647 draft-ietf-tls-compression-04.txt.
11651 * Add the ASN.1 structures and functions for CertificatePair, which
11652 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11654 CertificatePair ::= SEQUENCE {
11655 forward [0] Certificate OPTIONAL,
11656 reverse [1] Certificate OPTIONAL,
11657 -- at least one of the pair shall be present -- }
11659 Also implement the PEM functions to read and write certificate
11660 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11662 This needed to be defined, mostly for the sake of the LDAP
11663 attribute crossCertificatePair, but may prove useful elsewhere as
11668 * Make it possible to inhibit symlinking of shared libraries in
11669 Makefile.shared, for Cygwin's sake.
11673 * Extend the BIGNUM API by creating a function
11674 void BN_set_negative(BIGNUM *a, int neg);
11675 and a macro that behave like
11676 int BN_is_negative(const BIGNUM *a);
11678 to avoid the need to access 'a->neg' directly in applications.
11682 * Implement fast modular reduction for pseudo-Mersenne primes
11683 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11684 EC_GROUP_new_curve_GFp() will now automatically use this
11687 *Nils Larsch <nla@trustcenter.de>*
11689 * Add new lock type (CRYPTO_LOCK_BN).
11693 * Change the ENGINE framework to automatically load engines
11694 dynamically from specific directories unless they could be
11695 found to already be built in or loaded. Move all the
11696 current engines except for the cryptodev one to a new
11697 directory engines/.
11698 The engines in engines/ are built as shared libraries if
11699 the "shared" options was given to ./Configure or ./config.
11700 Otherwise, they are inserted in libcrypto.a.
11701 /usr/local/ssl/engines is the default directory for dynamic
11702 engines, but that can be overridden at configure time through
11703 the usual use of --prefix and/or --openssldir, and at run
11704 time with the environment variable OPENSSL_ENGINES.
11706 *Geoff Thorpe and Richard Levitte*
11708 * Add Makefile.shared, a helper makefile to build shared
11709 libraries. Adapt Makefile.org.
11713 * Add version info to Win32 DLLs.
11715 *Peter 'Luna' Runestig" <peter@runestig.com>*
11717 * Add new 'medium level' PKCS#12 API. Certificates and keys
11718 can be added using this API to created arbitrary PKCS#12
11719 files while avoiding the low-level API.
11721 New options to PKCS12_create(), key or cert can be NULL and
11722 will then be omitted from the output file. The encryption
11723 algorithm NIDs can be set to -1 for no encryption, the mac
11724 iteration count can be set to 0 to omit the mac.
11726 Enhance pkcs12 utility by making the -nokeys and -nocerts
11727 options work when creating a PKCS#12 file. New option -nomac
11728 to omit the mac, NONE can be set for an encryption algorithm.
11729 New code is modified to use the enhanced PKCS12_create()
11730 instead of the low-level API.
11734 * Extend ASN1 encoder to support indefinite length constructed
11735 encoding. This can output sequences tags and octet strings in
11736 this form. Modify pk7_asn1.c to support indefinite length
11737 encoding. This is experimental and needs additional code to
11738 be useful, such as an ASN1 bio and some enhanced streaming
11741 Extend template encode functionality so that tagging is passed
11742 down to the template encoder.
11746 * Let 'openssl req' fail if an argument to '-newkey' is not
11747 recognized instead of using RSA as a default.
11751 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11752 As these are not official, they are not included in "ALL";
11753 the "ECCdraft" ciphersuite group alias can be used to select them.
11755 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11757 * Add ECDH engine support.
11759 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11761 * Add ECDH in new directory crypto/ecdh/.
11763 *Douglas Stebila (Sun Microsystems Laboratories)*
11765 * Let BN_rand_range() abort with an error after 100 iterations
11766 without success (which indicates a broken PRNG).
11770 * Change BN_mod_sqrt() so that it verifies that the input value
11771 is really the square of the return value. (Previously,
11772 BN_mod_sqrt would show GIGO behaviour.)
11776 * Add named elliptic curves over binary fields from X9.62, SECG,
11777 and WAP/WTLS; add OIDs that were still missing.
11779 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11781 * Extend the EC library for elliptic curves over binary fields
11782 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11785 EC_GF2m_simple_method
11789 EC_GROUP_new_curve_GF2m
11790 EC_GROUP_set_curve_GF2m
11791 EC_GROUP_get_curve_GF2m
11792 EC_POINT_set_affine_coordinates_GF2m
11793 EC_POINT_get_affine_coordinates_GF2m
11794 EC_POINT_set_compressed_coordinates_GF2m
11796 Point compression for binary fields is disabled by default for
11797 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11800 As binary polynomials are represented as BIGNUMs, various members
11801 of the EC_GROUP and EC_POINT data structures can be shared
11802 between the implementations for prime fields and binary fields;
11803 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11804 are essentially identical to their `..._GFp` counterparts.
11805 (For simplicity, the `..._GFp` prefix has been dropped from
11806 various internal method names.)
11808 An internal 'field_div' method (similar to 'field_mul' and
11809 'field_sqr') has been added; this is used only for binary fields.
11811 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11813 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11814 through methods ('mul', 'precompute_mult').
11816 The generic implementations (now internally called 'ec_wNAF_mul'
11817 and 'ec_wNAF_precomputed_mult') remain the default if these
11818 methods are undefined.
11820 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11822 * New function EC_GROUP_get_degree, which is defined through
11823 EC_METHOD. For curves over prime fields, this returns the bit
11824 length of the modulus.
11826 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11828 * New functions EC_GROUP_dup, EC_POINT_dup.
11829 (These simply call ..._new and ..._copy).
11831 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11833 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11834 Polynomials are represented as BIGNUMs (where the sign bit is not
11835 used) in the following functions [macros]:
11838 BN_GF2m_sub [= BN_GF2m_add]
11839 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11840 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11841 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11843 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11844 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11845 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11846 BN_GF2m_cmp [= BN_ucmp]
11848 (Note that only the 'mod' functions are actually for fields GF(2^m).
11849 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11851 For some functions, an the irreducible polynomial defining a
11852 field can be given as an 'unsigned int[]' with strictly
11853 decreasing elements giving the indices of those bits that are set;
11854 i.e., p[] represents the polynomial
11855 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11857 p[0] > p[1] > ... > p[k] = 0.
11858 This applies to the following functions:
11861 BN_GF2m_mod_mul_arr
11862 BN_GF2m_mod_sqr_arr
11863 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11864 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11865 BN_GF2m_mod_exp_arr
11866 BN_GF2m_mod_sqrt_arr
11867 BN_GF2m_mod_solve_quad_arr
11871 Conversion can be performed by the following functions:
11876 bntest.c has additional tests for binary polynomial arithmetic.
11878 Two implementations for BN_GF2m_mod_div() are available.
11879 The default algorithm simply uses BN_GF2m_mod_inv() and
11880 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11881 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11882 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11884 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11886 * Add new error code 'ERR_R_DISABLED' that can be used when some
11887 functionality is disabled at compile-time.
11889 *Douglas Stebila <douglas.stebila@sun.com>*
11891 * Change default behaviour of 'openssl asn1parse' so that more
11892 information is visible when viewing, e.g., a certificate:
11894 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11895 mode the content of non-printable OCTET STRINGs is output in a
11896 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11897 avoid the appearance of a printable string.
11899 *Nils Larsch <nla@trustcenter.de>*
11901 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11903 EC_GROUP_set_asn1_flag()
11904 EC_GROUP_get_asn1_flag()
11905 EC_GROUP_set_point_conversion_form()
11906 EC_GROUP_get_point_conversion_form()
11907 These control ASN1 encoding details:
11908 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11909 has been set to OPENSSL_EC_NAMED_CURVE.
11910 - Points are encoded in uncompressed form by default; options for
11911 asn1_for are as for point2oct, namely
11912 POINT_CONVERSION_COMPRESSED
11913 POINT_CONVERSION_UNCOMPRESSED
11914 POINT_CONVERSION_HYBRID
11916 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11918 EC_GROUP_set_seed()
11919 EC_GROUP_get0_seed()
11920 EC_GROUP_get_seed_len()
11921 This is used only for ASN1 purposes (so far).
11923 *Nils Larsch <nla@trustcenter.de>*
11925 * Add 'field_type' member to EC_METHOD, which holds the NID
11926 of the appropriate field type OID. The new function
11927 EC_METHOD_get_field_type() returns this value.
11929 *Nils Larsch <nla@trustcenter.de>*
11932 EC_POINT_point2bn()
11933 EC_POINT_bn2point()
11934 EC_POINT_point2hex()
11935 EC_POINT_hex2point()
11936 providing useful interfaces to EC_POINT_point2oct() and
11937 EC_POINT_oct2point().
11939 *Nils Larsch <nla@trustcenter.de>*
11941 * Change internals of the EC library so that the functions
11942 EC_GROUP_set_generator()
11943 EC_GROUP_get_generator()
11944 EC_GROUP_get_order()
11945 EC_GROUP_get_cofactor()
11946 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11947 to methods, which would lead to unnecessary code duplication when
11948 adding different types of curves.
11950 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11952 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11953 arithmetic, and such that modified wNAFs are generated
11954 (which avoid length expansion in many cases).
11958 * Add a function EC_GROUP_check_discriminant() (defined via
11959 EC_METHOD) that verifies that the curve discriminant is non-zero.
11961 Add a function EC_GROUP_check() that makes some sanity tests
11962 on a EC_GROUP, its generator and order. This includes
11963 EC_GROUP_check_discriminant().
11965 *Nils Larsch <nla@trustcenter.de>*
11967 * Add ECDSA in new directory crypto/ecdsa/.
11969 Add applications 'openssl ecparam' and 'openssl ecdsa'
11970 (these are based on 'openssl dsaparam' and 'openssl dsa').
11972 ECDSA support is also included in various other files across the
11973 library. Most notably,
11974 - 'openssl req' now has a '-newkey ecdsa:file' option;
11975 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11976 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11977 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11978 them suitable for ECDSA where domain parameters must be
11979 extracted before the specific public key;
11980 - ECDSA engine support has been added.
11982 *Nils Larsch <nla@trustcenter.de>*
11984 * Include some named elliptic curves, and add OIDs from X9.62,
11985 SECG, and WAP/WTLS. Each curve can be obtained from the new
11987 EC_GROUP_new_by_curve_name(),
11988 and the list of available named curves can be obtained with
11989 EC_get_builtin_curves().
11990 Also add a 'curve_name' member to EC_GROUP objects, which can be
11992 EC_GROUP_set_curve_name()
11993 EC_GROUP_get_curve_name()
11995 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11997 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11998 was actually never needed) and in BN_mul(). The removal in BN_mul()
11999 required a small change in bn_mul_part_recursive() and the addition
12000 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12001 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12002 bn_sub_words() and bn_add_words() except they take arrays with
12007 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
12009 * Cleanse PEM buffers before freeing them since they may contain
12012 *Benjamin Bennett <ben@psc.edu>*
12014 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
12015 a ciphersuite string such as "DEFAULT:RSA" cannot enable
12016 authentication-only ciphersuites.
12020 * Since AES128 and AES256 share a single mask bit in the logic of
12021 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
12022 kludge to work properly if AES128 is available and AES256 isn't.
12026 * Expand security boundary to match 1.1.1 module.
12030 * Remove redundant features: hash file source, editing of test vectors
12031 modify fipsld to use external fips_premain.c signature.
12035 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12036 run algorithm test programs.
12040 * Make algorithm test programs more tolerant of whitespace.
12044 * Have SSL/TLS server implementation tolerate "mismatched" record
12045 protocol version while receiving ClientHello even if the
12046 ClientHello is fragmented. (The server can't insist on the
12047 particular protocol version it has chosen before the ServerHello
12048 message has informed the client about his choice.)
12052 * Load error codes if they are not already present instead of using a
12053 static variable. This allows them to be cleanly unloaded and reloaded.
12057 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12059 * Introduce limits to prevent malicious keys being able to
12060 cause a denial of service. ([CVE-2006-2940])
12062 *Steve Henson, Bodo Moeller*
12064 * Fix ASN.1 parsing of certain invalid structures that can result
12065 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12067 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12068 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12070 * Fix SSL client code which could crash if connecting to a
12071 malicious SSLv2 server. ([CVE-2006-4343])
12073 *Tavis Ormandy and Will Drewry, Google Security Team*
12075 * Change ciphersuite string processing so that an explicit
12076 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12077 will no longer include "AES128-SHA"), and any other similar
12078 ciphersuite (same bitmap) from *other* protocol versions (so that
12079 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12080 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12081 changes from 0.9.8b and 0.9.8d.
12085 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12087 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12088 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12090 * Change the Unix randomness entropy gathering to use poll() when
12091 possible instead of select(), since the latter has some
12092 undesirable limitations.
12094 *Darryl Miles via Richard Levitte and Bodo Moeller*
12096 * Disable rogue ciphersuites:
12098 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12099 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12100 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12102 The latter two were purportedly from
12103 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12106 Also deactivate the remaining ciphersuites from
12107 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12108 unofficial, and the ID has long expired.
12112 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12113 dual-core machines) and other potential thread-safety issues.
12117 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12119 * Adapt fipsld and the build system to link against the validated FIPS
12120 module in FIPS mode.
12124 * Fixes for VC++ 2005 build under Windows.
12128 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12129 from a Windows bash shell such as MSYS. It is autodetected from the
12130 "config" script when run from a VC++ environment. Modify standard VC++
12131 build to use fipscanister.o from the GNU make build.
12135 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12137 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12138 The value now differs depending on if you build for FIPS or not.
12139 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12140 safely run with a non-FIPSed libcrypto, as it may crash because of
12141 the difference induced by this change.
12145 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12147 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12148 (part of SSL_OP_ALL). This option used to disable the
12149 countermeasure against man-in-the-middle protocol-version
12150 rollback in the SSL 2.0 server implementation, which is a bad
12151 idea. ([CVE-2005-2969])
12153 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12154 for Information Security, National Institute of Advanced Industrial
12155 Science and Technology [AIST, Japan)]*
12157 * Minimal support for X9.31 signatures and PSS padding modes. This is
12158 mainly for FIPS compliance and not fully integrated at this stage.
12162 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12163 the exponentiation using a fixed-length exponent. (Otherwise,
12164 the information leaked through timing could expose the secret key
12165 after many signatures; cf. Bleichenbacher's attack on DSA with
12170 * Make a new fixed-window mod_exp implementation the default for
12171 RSA, DSA, and DH private-key operations so that the sequence of
12172 squares and multiplies and the memory access pattern are
12173 independent of the particular secret key. This will mitigate
12174 cache-timing and potential related attacks.
12176 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12177 and this is automatically used by BN_mod_exp_mont() if the new flag
12178 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12179 will use this BN flag for private exponents unless the flag
12180 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12181 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12183 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12185 * Change the client implementation for SSLv23_method() and
12186 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12187 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12188 (Previously, the SSL 2.0 backwards compatible Client Hello
12189 message format would be used even with SSL_OP_NO_SSLv2.)
12193 * Add support for smime-type MIME parameter in S/MIME messages which some
12198 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12199 a threadsafe manner. Modify rsa code to use new function and add calls
12200 to dsa and dh code (which had race conditions before).
12204 * Include the fixed error library code in the C error file definitions
12205 instead of fixing them up at runtime. This keeps the error code
12206 structures constant.
12210 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12212 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12215 * Fixes for newer kerberos headers. NB: the casts are needed because
12216 the 'length' field is signed on one version and unsigned on another
12217 with no (?) obvious way to tell the difference, without these VC++
12218 complains. Also the "definition" of FAR (blank) is no longer included
12219 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12220 some needed definitions.
12224 * Undo Cygwin change.
12228 * Added support for proxy certificates according to RFC 3820.
12229 Because they may be a security thread to unaware applications,
12230 they must be explicitly allowed in run-time. See
12231 docs/HOWTO/proxy_certificates.txt for further information.
12235 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12237 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12238 server and client random values. Previously
12239 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12240 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12242 This change has negligible security impact because:
12244 1. Server and client random values still have 24 bytes of pseudo random
12247 2. Server and client random values are sent in the clear in the initial
12250 3. The master secret is derived using the premaster secret (48 bytes in
12251 size for static RSA ciphersuites) as well as client server and random
12254 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12257 *Stephen Henson, reported by UK NISCC*
12259 * Use Windows randomness collection on Cygwin.
12263 * Fix hang in EGD/PRNGD query when communication socket is closed
12264 prematurely by EGD/PRNGD.
12266 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12268 * Prompt for pass phrases when appropriate for PKCS12 input format.
12272 * Back-port of selected performance improvements from development
12273 branch, as well as improved support for PowerPC platforms.
12277 * Add lots of checks for memory allocation failure, error codes to indicate
12278 failure and freeing up memory if a failure occurs.
12280 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12282 * Add new -passin argument to dgst.
12286 * Perform some character comparisons of different types in X509_NAME_cmp:
12287 this is needed for some certificates that re-encode DNs into UTF8Strings
12288 (in violation of RFC3280) and can't or won't issue name rollover
12293 * Make an explicit check during certificate validation to see that
12294 the CA setting in each certificate on the chain is correct. As a
12295 side effect always do the following basic checks on extensions,
12296 not just when there's an associated purpose to the check:
12298 - if there is an unhandled critical extension (unless the user
12299 has chosen to ignore this fault)
12300 - if the path length has been exceeded (if one is set at all)
12301 - that certain extensions fit the associated purpose (if one has
12306 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12308 * Avoid a race condition when CRLs are checked in a multi threaded
12309 environment. This would happen due to the reordering of the revoked
12310 entries during signature checking and serial number lookup. Now the
12311 encoding is cached and the serial number sort performed under a lock.
12312 Add new STACK function sk_is_sorted().
12316 * Add Delta CRL to the extension code.
12320 * Various fixes to s3_pkt.c so alerts are sent properly.
12322 *David Holmes <d.holmes@f5.com>*
12324 * Reduce the chances of duplicate issuer name and serial numbers (in
12325 violation of RFC3280) using the OpenSSL certificate creation utilities.
12326 This is done by creating a random 64 bit value for the initial serial
12327 number when a serial number file is created or when a self signed
12328 certificate is created using 'openssl req -x509'. The initial serial
12329 number file is created using 'openssl x509 -next_serial' in CA.pl
12330 rather than being initialized to 1.
12334 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12336 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12337 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12339 *Joe Orton, Steve Henson*
12341 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12344 *Joe Orton, Steve Henson*
12346 * Make it possible to have multiple active certificates with the same
12347 subject in the CA index file. This is done only if the keyword
12348 'unique_subject' is set to 'no' in the main CA section (default
12349 if 'CA_default') of the configuration file. The value is saved
12350 with the database itself in a separate index attribute file,
12351 named like the index file with '.attr' appended to the name.
12355 * X509 verify fixes. Disable broken certificate workarounds when
12356 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12357 keyUsage extension present. Don't accept CRLs with unhandled critical
12358 extensions: since verify currently doesn't process CRL extensions this
12359 rejects a CRL with *any* critical extensions. Add new verify error codes
12364 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12365 A clarification of RFC2560 will require the use of OCTET STRINGs and
12366 some implementations cannot handle the current raw format. Since OpenSSL
12367 copies and compares OCSP nonces as opaque blobs without any attempt at
12368 parsing them this should not create any compatibility issues.
12372 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12373 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12374 this HMAC (and other) operations are several times slower than OpenSSL
12379 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12381 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12383 * Use the correct content when signing type "other".
12387 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12389 * Fix various bugs revealed by running the NISCC test suite:
12391 Stop out of bounds reads in the ASN1 code when presented with
12392 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12394 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12396 If verify callback ignores invalid public key errors don't try to check
12397 certificate signature with the NULL public key.
12401 * New -ignore_err option in ocsp application to stop the server
12402 exiting on the first error in a request.
12406 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12407 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12412 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12413 extra data after the compression methods not only for TLS 1.0
12414 but also for SSL 3.0 (as required by the specification).
12416 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12418 * Change X509_certificate_type() to mark the key as exported/exportable
12419 when it's 512 *bits* long, not 512 bytes.
12423 * Change AES_cbc_encrypt() so it outputs exact multiple of
12424 blocks during encryption.
12428 * Various fixes to base64 BIO and non blocking I/O. On write
12429 flushes were not handled properly if the BIO retried. On read
12430 data was not being buffered properly and had various logic bugs.
12431 This also affects blocking I/O when the data being decoded is a
12436 * Various S/MIME bugfixes and compatibility changes:
12437 output correct application/pkcs7 MIME type if
12438 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12439 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12440 of files as .eml work). Correctly handle very long lines in MIME
12445 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12447 * Countermeasure against the Klima-Pokorny-Rosa extension of
12448 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12449 a protocol version number mismatch like a decryption error
12450 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12454 * Turn on RSA blinding by default in the default implementation
12455 to avoid a timing attack. Applications that don't want it can call
12456 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12457 They would be ill-advised to do so in most cases.
12459 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12461 * Change RSA blinding code so that it works when the PRNG is not
12462 seeded (in this case, the secret RSA exponent is abused as
12463 an unpredictable seed -- if it is not unpredictable, there
12464 is no point in blinding anyway). Make RSA blinding thread-safe
12465 by remembering the creator's thread ID in rsa->blinding and
12466 having all other threads use local one-time blinding factors
12467 (this requires more computation than sharing rsa->blinding, but
12468 avoids excessive locking; and if an RSA object is not shared
12469 between threads, blinding will still be very fast).
12473 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12474 ENGINE as defaults for all supported algorithms irrespective of
12475 the 'flags' parameter. 'flags' is now honoured, so applications
12476 should make sure they are passing it correctly.
12480 * Target "mingw" now allows native Windows code to be generated in
12481 the Cygwin environment as well as with the MinGW compiler.
12485 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12487 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12488 via timing by performing a MAC computation even if incorrect
12489 block cipher padding has been found. This is a countermeasure
12490 against active attacks where the attacker has to distinguish
12491 between bad padding and a MAC verification error. ([CVE-2003-0078])
12493 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12494 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12495 Martin Vuagnoux (EPFL, Ilion)*
12497 * Make the no-err option work as intended. The intention with no-err
12498 is not to have the whole error stack handling routines removed from
12499 libcrypto, it's only intended to remove all the function name and
12500 reason texts, thereby removing some of the footprint that may not
12501 be interesting if those errors aren't displayed anyway.
12503 NOTE: it's still possible for any application or module to have its
12504 own set of error texts inserted. The routines are there, just not
12505 used by default when no-err is given.
12509 * Add support for FreeBSD on IA64.
12511 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12513 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12514 Kerberos function mit_des_cbc_cksum(). Before this change,
12515 the value returned by DES_cbc_cksum() was like the one from
12516 mit_des_cbc_cksum(), except the bytes were swapped.
12518 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12520 * Allow an application to disable the automatic SSL chain building.
12521 Before this a rather primitive chain build was always performed in
12522 ssl3_output_cert_chain(): an application had no way to send the
12523 correct chain if the automatic operation produced an incorrect result.
12525 Now the chain builder is disabled if either:
12527 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12529 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12531 The reasoning behind this is that an application would not want the
12532 auto chain building to take place if extra chain certificates are
12533 present and it might also want a means of sending no additional
12534 certificates (for example the chain has two certificates and the
12539 * Add the possibility to build without the ENGINE framework.
12541 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12543 * Under Win32 gmtime() can return NULL: check return value in
12544 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12548 * DSA routines: under certain error conditions uninitialized BN objects
12549 could be freed. Solution: make sure initialization is performed early
12550 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12551 Nils Larsch <nla@trustcenter.de> via PR#459)
12555 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12556 checked on reconnect on the client side, therefore session resumption
12557 could still fail with a "ssl session id is different" error. This
12558 behaviour is masked when SSL_OP_ALL is used due to
12559 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12560 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12561 followup to PR #377.
12565 * IA-32 assembler support enhancements: unified ELF targets, support
12566 for SCO/Caldera platforms, fix for Cygwin shared build.
12570 * Add support for FreeBSD on sparc64. As a consequence, support for
12571 FreeBSD on non-x86 processors is separate from x86 processors on
12572 the config script, much like the NetBSD support.
12574 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12576 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12578 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12581 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12582 code (06) was taken as the first octet of the session ID and the last
12583 octet was ignored consequently. As a result SSLv2 client side session
12584 caching could not have worked due to the session ID mismatch between
12586 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12591 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12592 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12597 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12598 seems that in spite of existing for more than a year, many application
12599 author have done nothing to provide the necessary callbacks, which
12600 means that this particular engine will not work properly anywhere.
12601 This is a very unfortunate situation which forces us, in the name
12602 of usability, to give the hw_ncipher.c a static lock, which is part
12604 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12605 appear in 0.9.8 or later. We EXPECT application authors to have
12606 dealt properly with this when 0.9.8 is released (unless we actually
12607 make such changes in the libcrypto locking code that changes will
12608 have to be made anyway).
12612 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12613 octets have been read, EOF or an error occurs. Without this change
12614 some truncated ASN1 structures will not produce an error.
12618 * Disable Heimdal support, since it hasn't been fully implemented.
12619 Still give the possibility to force the use of Heimdal, but with
12620 warnings and a request that patches get sent to openssl-dev.
12624 * Add the VC-CE target, introduce the WINCE sysname, and add
12625 INSTALL.WCE and appropriate conditionals to make it build.
12627 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12629 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12630 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12631 edit numbers of the version.
12633 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12635 * Introduce safe string copy and catenation functions
12636 (BUF_strlcpy() and BUF_strlcat()).
12638 *Ben Laurie (CHATS) and Richard Levitte*
12640 * Avoid using fixed-size buffers for one-line DNs.
12642 *Ben Laurie (CHATS)*
12644 * Add BUF_MEM_grow_clean() to avoid information leakage when
12645 resizing buffers containing secrets, and use where appropriate.
12647 *Ben Laurie (CHATS)*
12649 * Avoid using fixed size buffers for configuration file location.
12651 *Ben Laurie (CHATS)*
12653 * Avoid filename truncation for various CA files.
12655 *Ben Laurie (CHATS)*
12657 * Use sizeof in preference to magic numbers.
12659 *Ben Laurie (CHATS)*
12661 * Avoid filename truncation in cert requests.
12663 *Ben Laurie (CHATS)*
12665 * Add assertions to check for (supposedly impossible) buffer
12668 *Ben Laurie (CHATS)*
12670 * Don't cache truncated DNS entries in the local cache (this could
12671 potentially lead to a spoofing attack).
12673 *Ben Laurie (CHATS)*
12675 * Fix various buffers to be large enough for hex/decimal
12676 representations in a platform independent manner.
12678 *Ben Laurie (CHATS)*
12680 * Add CRYPTO_realloc_clean() to avoid information leakage when
12681 resizing buffers containing secrets, and use where appropriate.
12683 *Ben Laurie (CHATS)*
12685 * Add BIO_indent() to avoid much slightly worrying code to do
12688 *Ben Laurie (CHATS)*
12690 * Convert sprintf()/BIO_puts() to BIO_printf().
12692 *Ben Laurie (CHATS)*
12694 * buffer_gets() could terminate with the buffer only half
12697 *Ben Laurie (CHATS)*
12699 * Add assertions to prevent user-supplied crypto functions from
12700 overflowing internal buffers by having large block sizes, etc.
12702 *Ben Laurie (CHATS)*
12704 * New OPENSSL_assert() macro (similar to assert(), but enabled
12707 *Ben Laurie (CHATS)*
12709 * Eliminate unused copy of key in RC4.
12711 *Ben Laurie (CHATS)*
12713 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12715 *Ben Laurie (CHATS)*
12717 * Fix off-by-one error in EGD path.
12719 *Ben Laurie (CHATS)*
12721 * If RANDFILE path is too long, ignore instead of truncating.
12723 *Ben Laurie (CHATS)*
12725 * Eliminate unused and incorrectly sized X.509 structure
12728 *Ben Laurie (CHATS)*
12730 * Eliminate unused and dangerous function knumber().
12732 *Ben Laurie (CHATS)*
12734 * Eliminate unused and dangerous structure, KSSL_ERR.
12736 *Ben Laurie (CHATS)*
12738 * Protect against overlong session ID context length in an encoded
12739 session object. Since these are local, this does not appear to be
12742 *Ben Laurie (CHATS)*
12744 * Change from security patch (see 0.9.6e below) that did not affect
12745 the 0.9.6 release series:
12747 Remote buffer overflow in SSL3 protocol - an attacker could
12748 supply an oversized master key in Kerberos-enabled versions.
12751 *Ben Laurie (CHATS)*
12753 * Change the SSL kerb5 codes to match RFC 2712.
12757 * Make -nameopt work fully for req and add -reqopt switch.
12759 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12761 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12763 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12765 * Make sure tests can be performed even if the corresponding algorithms
12766 have been removed entirely. This was also the last step to make
12767 OpenSSL compilable with DJGPP under all reasonable conditions.
12769 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12771 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12772 to allow version independent disabling of normally unselected ciphers,
12773 which may be activated as a side-effect of selecting a single cipher.
12775 (E.g., cipher list string "RSA" enables ciphersuites that are left
12776 out of "ALL" because they do not provide symmetric encryption.
12777 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12779 *Lutz Jaenicke, Bodo Moeller*
12781 * Add appropriate support for separate platform-dependent build
12782 directories. The recommended way to make a platform-dependent
12783 build directory is the following (tested on Linux), maybe with
12786 # Place yourself outside of the OpenSSL source tree. In
12787 # this example, the environment variable OPENSSL_SOURCE
12788 # is assumed to contain the absolute OpenSSL source directory.
12789 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12790 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12791 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12792 mkdir -p `dirname $F`
12793 ln -s $OPENSSL_SOURCE/$F $F
12796 To be absolutely sure not to disturb the source tree, a "make clean"
12797 is a good thing. If it isn't successful, don't worry about it,
12798 it probably means the source directory is very clean.
12802 * Make sure any ENGINE control commands make local copies of string
12803 pointers passed to them whenever necessary. Otherwise it is possible
12804 the caller may have overwritten (or deallocated) the original string
12805 data when a later ENGINE operation tries to use the stored values.
12807 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12809 * Improve diagnostics in file reading and command-line digests.
12811 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12813 * Add AES modes CFB and OFB to the object database. Correct an
12814 error in AES-CFB decryption.
12818 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12819 allows existing EVP_CIPHER_CTX structures to be reused after
12820 calling `EVP_*Final()`. This behaviour is used by encryption
12821 BIOs and some applications. This has the side effect that
12822 applications must explicitly clean up cipher contexts with
12823 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12827 * Check the values of dna and dnb in bn_mul_recursive before calling
12828 bn_mul_comba (a non zero value means the a or b arrays do not contain
12829 n2 elements) and fallback to bn_mul_normal if either is not zero.
12833 * Fix escaping of non-ASCII characters when using the -subj option
12834 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12838 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12839 form for "surname", serialNumber has no short form.
12840 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12841 therefore remove "mail" short name for "internet 7".
12842 The OID for unique identifiers in X509 certificates is
12843 x500UniqueIdentifier, not uniqueIdentifier.
12844 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12848 * Add an "init" command to the ENGINE config module and auto initialize
12849 ENGINEs. Without any "init" command the ENGINE will be initialized
12850 after all ctrl commands have been executed on it. If init=1 the
12851 ENGINE is initialized at that point (ctrls before that point are run
12852 on the uninitialized ENGINE and after on the initialized one). If
12853 init=0 then the ENGINE will not be initialized at all.
12857 * Fix the 'app_verify_callback' interface so that the user-defined
12858 argument is actually passed to the callback: In the
12859 SSL_CTX_set_cert_verify_callback() prototype, the callback
12860 declaration has been changed from
12863 int (*cb)(X509_STORE_CTX *,void *);
12864 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12865 i=s->ctx->app_verify_callback(&ctx)
12866 has been changed into
12867 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12869 To update applications using SSL_CTX_set_cert_verify_callback(),
12870 a dummy argument can be added to their callback functions.
12872 *D. K. Smetters <smetters@parc.xerox.com>*
12874 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12876 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12878 * Add and OPENSSL_LOAD_CONF define which will cause
12879 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12880 This allows older applications to transparently support certain
12881 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12882 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12883 load the config file and OPENSSL_add_all_algorithms_conf() which will
12884 always load it have also been added.
12888 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12889 Adjust NIDs and EVP layer.
12891 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12893 * Config modules support in openssl utility.
12895 Most commands now load modules from the config file,
12896 though in a few (such as version) this isn't done
12897 because it couldn't be used for anything.
12899 In the case of ca and req the config file used is
12900 the same as the utility itself: that is the -config
12901 command line option can be used to specify an
12906 * Move default behaviour from OPENSSL_config(). If appname is NULL
12907 use "openssl_conf" if filename is NULL use default openssl config file.
12911 * Add an argument to OPENSSL_config() to allow the use of an alternative
12912 config section name. Add a new flag to tolerate a missing config file
12913 and move code to CONF_modules_load_file().
12917 * Support for crypto accelerator cards from Accelerated Encryption
12918 Processing, www.aep.ie. (Use engine 'aep')
12919 The support was copied from 0.9.6c [engine] and adapted/corrected
12920 to work with the new engine framework.
12922 *AEP Inc. and Richard Levitte*
12924 * Support for SureWare crypto accelerator cards from Baltimore
12925 Technologies. (Use engine 'sureware')
12926 The support was copied from 0.9.6c [engine] and adapted
12927 to work with the new engine framework.
12931 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12932 make the newer ENGINE framework commands for the CHIL engine work.
12934 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12936 * Make it possible to produce shared libraries on ReliantUNIX.
12938 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12940 * Add the configuration target debug-linux-ppro.
12941 Make 'openssl rsa' use the general key loading routines
12942 implemented in `apps.c`, and make those routines able to
12943 handle the key format FORMAT_NETSCAPE and the variant
12946 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12948 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12950 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12952 * Add -keyform to rsautl, and document -engine.
12954 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12956 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12957 BIO_R_NO_SUCH_FILE error code rather than the generic
12958 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12962 * Add new functions
12963 ERR_peek_last_error
12964 ERR_peek_last_error_line
12965 ERR_peek_last_error_line_data.
12966 These are similar to
12968 ERR_peek_error_line
12969 ERR_peek_error_line_data,
12970 but report on the latest error recorded rather than the first one
12971 still in the error queue.
12973 *Ben Laurie, Bodo Moeller*
12975 * default_algorithms option in ENGINE config module. This allows things
12977 default_algorithms = ALL
12978 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12982 * Preliminary ENGINE config module.
12986 * New experimental application configuration code.
12990 * Change the AES code to follow the same name structure as all other
12991 symmetric ciphers, and behave the same way. Move everything to
12992 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12994 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12996 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12998 *Ben Laurie and Theo de Raadt*
13000 * Add option to output public keys in req command.
13002 *Massimiliano Pala madwolf@openca.org*
13004 * Use wNAFs in EC_POINTs_mul() for improved efficiency
13005 (up to about 10% better than before for P-192 and P-224).
13009 * New functions/macros
13011 SSL_CTX_set_msg_callback(ctx, cb)
13012 SSL_CTX_set_msg_callback_arg(ctx, arg)
13013 SSL_set_msg_callback(ssl, cb)
13014 SSL_set_msg_callback_arg(ssl, arg)
13016 to request calling a callback function
13018 void cb(int write_p, int version, int content_type,
13019 const void *buf, size_t len, SSL *ssl, void *arg)
13021 whenever a protocol message has been completely received
13022 (write_p == 0) or sent (write_p == 1). Here 'version' is the
13023 protocol version according to which the SSL library interprets
13024 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
13025 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
13026 the content type as defined in the SSL 3.0/TLS 1.0 protocol
13027 specification (change_cipher_spec(20), alert(21), handshake(22)).
13028 'buf' and 'len' point to the actual message, 'ssl' to the
13029 SSL object, and 'arg' is the application-defined value set by
13030 SSL[_CTX]_set_msg_callback_arg().
13032 'openssl s_client' and 'openssl s_server' have new '-msg' options
13033 to enable a callback that displays all protocol messages.
13037 * Change the shared library support so shared libraries are built as
13038 soon as the corresponding static library is finished, and thereby get
13039 openssl and the test programs linked against the shared library.
13040 This still only happens when the keyword "shard" has been given to
13041 the configuration scripts.
13043 NOTE: shared library support is still an experimental thing, and
13044 backward binary compatibility is still not guaranteed.
13046 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13048 * Add support for Subject Information Access extension.
13050 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13052 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13053 additional bytes when new memory had to be allocated, not just
13054 when reusing an existing buffer.
13058 * New command line and configuration option 'utf8' for the req command.
13059 This allows field values to be specified as UTF8 strings.
13063 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13064 runs for the former and machine-readable output for the latter.
13068 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13069 of the e-mail address in the DN (i.e., it will go into a certificate
13070 extension only). The new configuration file option 'email_in_dn = no'
13071 has the same effect.
13073 *Massimiliano Pala madwolf@openca.org*
13075 * Change all functions with names starting with `des_` to be starting
13076 with `DES_` instead. Add wrappers that are compatible with libdes,
13077 but are named `_ossl_old_des_*`. Finally, add macros that map the
13078 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13079 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13080 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13083 Since we provide two compatibility mappings, the user needs to
13084 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13085 compatibility is desired. The default (i.e., when that macro
13086 isn't defined) is OpenSSL 0.9.6c compatibility.
13088 There are also macros that enable and disable the support of old
13089 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13090 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13091 are defined, the default will apply: to support the old des routines.
13093 In either case, one must include openssl/des.h to get the correct
13094 definitions. Do not try to just include openssl/des_old.h, that
13097 NOTE: This is a major break of an old API into a new one. Software
13098 authors are encouraged to switch to the `DES_` style functions. Some
13099 time in the future, des_old.h and the libdes compatibility functions
13100 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13101 default), and then completely removed.
13105 * Test for certificates which contain unsupported critical extensions.
13106 If such a certificate is found during a verify operation it is
13107 rejected by default: this behaviour can be overridden by either
13108 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13109 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13110 X509_supported_extension() has also been added which returns 1 if a
13111 particular extension is supported.
13115 * Modify the behaviour of EVP cipher functions in similar way to digests
13116 to retain compatibility with existing code.
13120 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13121 compatibility with existing code. In particular the 'ctx' parameter does
13122 not have to be to be initialized before the call to EVP_DigestInit() and
13123 it is tidied up after a call to EVP_DigestFinal(). New function
13124 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13125 EVP_MD_CTX_copy() changed to not require the destination to be
13126 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13127 requires the destination to be valid.
13129 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13130 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13134 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13135 so that complete 'Handshake' protocol structures are kept in memory
13136 instead of overwriting 'msg_type' and 'length' with 'body' data.
13140 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13142 *Massimo Santin via Richard Levitte*
13144 * Major restructuring to the underlying ENGINE code. This includes
13145 reduction of linker bloat, separation of pure "ENGINE" manipulation
13146 (initialisation, etc) from functionality dealing with implementations
13147 of specific crypto interfaces. This change also introduces integrated
13148 support for symmetric ciphers and digest implementations - so ENGINEs
13149 can now accelerate these by providing EVP_CIPHER and EVP_MD
13150 implementations of their own. This is detailed in
13151 [crypto/engine/README.md](crypto/engine/README.md)
13152 as it couldn't be adequately described here. However, there are a few
13153 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13154 were changed in the original introduction of ENGINE code have now
13155 reverted back - the hooking from this code to ENGINE is now a good
13156 deal more passive and at run-time, operations deal directly with
13157 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13158 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13159 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13160 they were not being used by the framework as there is no concept of a
13161 BIGNUM_METHOD and they could not be generalised to the new
13162 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13163 ENGINE_cpy() has been removed as it cannot be consistently defined in
13168 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13172 * Change mkdef.pl to sort symbols that get the same entry number,
13173 and make sure the automatically generated functions `ERR_load_*`
13174 become part of libeay.num as well.
13178 * New function SSL_renegotiate_pending(). This returns true once
13179 renegotiation has been requested (either SSL_renegotiate() call
13180 or HelloRequest/ClientHello received from the peer) and becomes
13181 false once a handshake has been completed.
13182 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13183 sends a HelloRequest, but does not ensure that a handshake takes
13184 place. SSL_renegotiate_pending() is useful for checking if the
13185 client has followed the request.)
13189 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13190 By default, clients may request session resumption even during
13191 renegotiation (if session ID contexts permit); with this option,
13192 session resumption is possible only in the first handshake.
13194 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13195 more bits available for options that should not be part of
13196 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13200 * Add some demos for certificate and certificate request creation.
13204 * Make maximum certificate chain size accepted from the peer application
13205 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13206 "Douglas E. Engert" <deengert@anl.gov>.
13210 * Add support for shared libraries for Unixware-7
13211 (Boyd Lynn Gerber <gerberb@zenez.com>).
13215 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13216 be done prior to destruction. Use this to unload error strings from
13217 ENGINEs that load their own error strings. NB: This adds two new API
13218 functions to "get" and "set" this destroy handler in an ENGINE.
13222 * Alter all existing ENGINE implementations (except "openssl" and
13223 "openbsd") to dynamically instantiate their own error strings. This
13224 makes them more flexible to be built both as statically-linked ENGINEs
13225 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13226 Also, add stub code to each that makes building them as self-contained
13227 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13231 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13232 implementations into applications that are completely implemented in
13233 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13234 commands that can be used to configure what shared-library to load and
13235 to control aspects of the way it is handled. Also, made an update to
13236 the [README-Engine.md](README-Engine.md) file
13237 that brings its information up-to-date and
13238 provides some information and instructions on the "dynamic" ENGINE
13239 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13243 * Make it possible to unload ranges of ERR strings with a new
13244 "ERR_unload_strings" function.
13248 * Add a copy() function to EVP_MD.
13252 * Make EVP_MD routines take a context pointer instead of just the
13253 md_data void pointer.
13257 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13258 that the digest can only process a single chunk of data
13259 (typically because it is provided by a piece of
13260 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13261 is only going to provide a single chunk of data, and hence the
13262 framework needn't accumulate the data for oneshot drivers.
13266 * As with "ERR", make it possible to replace the underlying "ex_data"
13267 functions. This change also alters the storage and management of global
13268 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13269 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13270 index counters. The API functions that use this state have been changed
13271 to take a "class_index" rather than pointers to the class's local STACK
13272 and counter, and there is now an API function to dynamically create new
13273 classes. This centralisation allows us to (a) plug a lot of the
13274 thread-safety problems that existed, and (b) makes it possible to clean
13275 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13276 such data would previously have always leaked in application code and
13277 workarounds were in place to make the memory debugging turn a blind eye
13278 to it. Application code that doesn't use this new function will still
13279 leak as before, but their memory debugging output will announce it now
13280 rather than letting it slide.
13282 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13283 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13284 has a return value to indicate success or failure.
13288 * Make it possible to replace the underlying "ERR" functions such that the
13289 global state (2 LHASH tables and 2 locks) is only used by the "default"
13290 implementation. This change also adds two functions to "get" and "set"
13291 the implementation prior to it being automatically set the first time
13292 any other ERR function takes place. Ie. an application can call "get",
13293 pass the return value to a module it has just loaded, and that module
13294 can call its own "set" function using that value. This means the
13295 module's "ERR" operations will use (and modify) the error state in the
13296 application and not in its own statically linked copy of OpenSSL code.
13300 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13301 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13302 the operation, and provides a more encapsulated way for external code
13303 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13304 to use these functions rather than manually incrementing the counts.
13306 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13310 * Add EVP test program.
13314 * Add symmetric cipher support to ENGINE. Expect the API to change!
13318 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13319 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13320 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13321 These allow a CRL to be built without having to access X509_CRL fields
13322 directly. Modify 'ca' application to use new functions.
13326 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13327 bug workarounds. Rollback attack detection is a security feature.
13328 The problem will only arise on OpenSSL servers when TLSv1 is not
13329 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13330 Software authors not wanting to support TLSv1 will have special reasons
13331 for their choice and can explicitly enable this option.
13333 *Bodo Moeller, Lutz Jaenicke*
13335 * Rationalise EVP so it can be extended: don't include a union of
13336 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13337 (similar to those existing for EVP_CIPHER_CTX).
13342 EVP_MD_CTX_init(&md); /* new function call */
13343 EVP_DigestInit(&md, EVP_sha1());
13344 EVP_DigestUpdate(&md, in, len);
13345 EVP_DigestFinal(&md, out, NULL);
13346 EVP_MD_CTX_cleanup(&md); /* new function call */
13350 * Make DES key schedule conform to the usual scheme, as well as
13351 correcting its structure. This means that calls to DES functions
13352 now have to pass a pointer to a des_key_schedule instead of a
13353 plain des_key_schedule (which was actually always a pointer
13356 des_key_schedule ks;
13358 des_set_key_checked(..., &ks);
13359 des_ncbc_encrypt(..., &ks, ...);
13361 (Note that a later change renames 'des_...' into 'DES_...'.)
13365 * Initial reduction of linker bloat: the use of some functions, such as
13366 PEM causes large amounts of unused functions to be linked in due to
13367 poor organisation. For example pem_all.c contains every PEM function
13368 which has a knock on effect of linking in large amounts of (unused)
13369 ASN1 code. Grouping together similar functions and splitting unrelated
13370 functions prevents this.
13374 * Cleanup of EVP macros.
13378 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13379 correct `_ecb suffix`.
13383 * Add initial OCSP responder support to ocsp application. The
13384 revocation information is handled using the text based index
13385 use by the ca application. The responder can either handle
13386 requests generated internally, supplied in files (for example
13387 via a CGI script) or using an internal minimal server.
13391 * Add configuration choices to get zlib compression for TLS.
13395 * Changes to Kerberos SSL for RFC 2712 compliance:
13396 1. Implemented real KerberosWrapper, instead of just using
13397 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13398 2. Implemented optional authenticator field of KerberosWrapper.
13400 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13401 and authenticator structs; see crypto/krb5/.
13403 Generalized Kerberos calls to support multiple Kerberos libraries.
13404 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13405 via Richard Levitte*
13407 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13408 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13409 values for each of the key sizes rather than having just
13410 parameters (and 'speed' generating keys each time).
13414 * Speed up EVP routines.
13417 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13418 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13419 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13420 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13422 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13423 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13424 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13427 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13429 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13433 * Added the OS2-EMX target.
13435 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13437 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13438 New functions to support `NCONF` routines in extension code.
13439 New function `CONF_set_nconf()`
13440 to allow functions which take an `NCONF` to also handle the old `LHASH`
13441 structure: this means that the old `CONF` compatible routines can be
13442 retained (in particular w.rt. extensions) without having to duplicate the
13443 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13447 * Enhance the general user interface with mechanisms for inner control
13448 and with possibilities to have yes/no kind of prompts.
13452 * Change all calls to low-level digest routines in the library and
13453 applications to use EVP. Add missing calls to HMAC_cleanup() and
13454 don't assume HMAC_CTX can be copied using memcpy().
13456 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13458 * Add the possibility to control engines through control names but with
13459 arbitrary arguments instead of just a string.
13460 Change the key loaders to take a UI_METHOD instead of a callback
13461 function pointer. NOTE: this breaks binary compatibility with earlier
13462 versions of OpenSSL [engine].
13463 Adapt the nCipher code for these new conditions and add a card insertion
13468 * Enhance the general user interface with mechanisms to better support
13469 dialog box interfaces, application-defined prompts, the possibility
13470 to use defaults (for example default passwords from somewhere else)
13471 and interrupts/cancellations.
13475 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13476 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13480 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13481 tidy up some unnecessarily weird code in 'sk_new()').
13483 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13485 * Change the key loading routines for ENGINEs to use the same kind
13486 callback (pem_password_cb) as all other routines that need this
13491 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13492 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13493 than this minimum value is recommended.
13497 * New random seeder for OpenVMS, using the system process statistics
13498 that are easily reachable.
13502 * Windows apparently can't transparently handle global
13503 variables defined in DLLs. Initialisations such as:
13505 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13507 won't compile. This is used by the any applications that need to
13508 declare their own ASN1 modules. This was fixed by adding the option
13509 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13510 needed for static libraries under Win32.
13514 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13515 setting of purpose and trust fields. New X509_STORE trust and
13516 purpose functions and tidy up setting in other SSL functions.
13520 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13521 structure. These are inherited by X509_STORE_CTX when it is
13522 initialised. This allows various defaults to be set in the
13523 X509_STORE structure (such as flags for CRL checking and custom
13524 purpose or trust settings) for functions which only use X509_STORE_CTX
13525 internally such as S/MIME.
13527 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13528 trust settings if they are not set in X509_STORE. This allows X509_STORE
13529 purposes and trust (in S/MIME for example) to override any set by default.
13531 Add command line options for CRL checking to smime, s_client and s_server
13536 * Initial CRL based revocation checking. If the CRL checking flag(s)
13537 are set then the CRL is looked up in the X509_STORE structure and
13538 its validity and signature checked, then if the certificate is found
13539 in the CRL the verify fails with a revoked error.
13541 Various new CRL related callbacks added to X509_STORE_CTX structure.
13543 Command line options added to 'verify' application to support this.
13545 This needs some additional work, such as being able to handle multiple
13546 CRLs with different times, extension based lookup (rather than just
13547 by subject name) and ultimately more complete V2 CRL extension
13552 * Add a general user interface API (crypto/ui/). This is designed
13553 to replace things like des_read_password and friends (backward
13554 compatibility functions using this new API are provided).
13555 The purpose is to remove prompting functions from the DES code
13556 section as well as provide for prompting through dialog boxes in
13557 a window system and the like.
13561 * Add "ex_data" support to ENGINE so implementations can add state at a
13562 per-structure level rather than having to store it globally.
13566 * Make it possible for ENGINE structures to be copied when retrieved by
13567 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13568 This causes the "original" ENGINE structure to act like a template,
13569 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13570 operational state can be localised to each ENGINE structure, despite the
13571 fact they all share the same "methods". New ENGINE structures returned in
13572 this case have no functional references and the return value is the single
13573 structural reference. This matches the single structural reference returned
13574 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13579 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13580 needs to match any other type at all we need to manually clear the
13585 * Changes to the "openssl engine" utility to include;
13586 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13587 about an ENGINE's available control commands.
13588 - executing control commands from command line arguments using the
13589 '-pre' and '-post' switches. '-post' is only used if '-t' is
13590 specified and the ENGINE is successfully initialised. The syntax for
13591 the individual commands are colon-separated, for example;
13592 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13596 * New dynamic control command support for ENGINEs. ENGINEs can now
13597 declare their own commands (numbers), names (strings), descriptions,
13598 and input types for run-time discovery by calling applications. A
13599 subset of these commands are implicitly classed as "executable"
13600 depending on their input type, and only these can be invoked through
13601 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13602 can be based on user input, config files, etc). The distinction is
13603 that "executable" commands cannot return anything other than a boolean
13604 result and can only support numeric or string input, whereas some
13605 discoverable commands may only be for direct use through
13606 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13607 pointers, or other custom uses. The "executable" commands are to
13608 support parameterisations of ENGINE behaviour that can be
13609 unambiguously defined by ENGINEs and used consistently across any
13610 OpenSSL-based application. Commands have been added to all the
13611 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13612 control over shared-library paths without source code alterations.
13616 * Changed all ENGINE implementations to dynamically allocate their
13617 ENGINEs rather than declaring them statically. Apart from this being
13618 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13619 this also allows the implementations to compile without using the
13620 internal engine_int.h header.
13624 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13625 'const' value. Any code that should be able to modify a RAND_METHOD
13626 should already have non-const pointers to it (ie. they should only
13627 modify their own ones).
13631 * Made a variety of little tweaks to the ENGINE code.
13632 - "atalla" and "ubsec" string definitions were moved from header files
13633 to C code. "nuron" string definitions were placed in variables
13634 rather than hard-coded - allowing parameterisation of these values
13635 later on via ctrl() commands.
13636 - Removed unused "#if 0"'d code.
13637 - Fixed engine list iteration code so it uses ENGINE_free() to release
13638 structural references.
13639 - Constified the RAND_METHOD element of ENGINE structures.
13640 - Constified various get/set functions as appropriate and added
13641 missing functions (including a catch-all ENGINE_cpy that duplicates
13642 all ENGINE values onto a new ENGINE except reference counts/state).
13643 - Removed NULL parameter checks in get/set functions. Setting a method
13644 or function to NULL is a way of cancelling out a previously set
13645 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13646 and doesn't justify the extra error symbols and code.
13647 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13648 flags from engine_int.h to engine.h.
13649 - Changed prototypes for ENGINE handler functions (init(), finish(),
13650 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13654 * Implement binary inversion algorithm for BN_mod_inverse in addition
13655 to the algorithm using long division. The binary algorithm can be
13656 used only if the modulus is odd. On 32-bit systems, it is faster
13657 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13658 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13659 up to 450 bits. In 64-bit environments, the binary algorithm
13660 appears to be advantageous for much longer moduli; here we use it
13661 for moduli up to 2048 bits.
13665 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13666 could not support the combine flag in choice fields.
13670 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13671 extensions from a certificate request to the certificate.
13675 * Allow multiple 'certopt' and 'nameopt' options to be separated
13676 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13677 file: this allows the display of the certificate about to be
13678 signed to be customised, to allow certain fields to be included
13679 or excluded and extension details. The old system didn't display
13680 multicharacter strings properly, omitted fields not in the policy
13681 and couldn't display additional details such as extensions.
13685 * Function EC_POINTs_mul for multiple scalar multiplication
13686 of an arbitrary number of elliptic curve points
13687 \sum scalars[i]*points[i],
13688 optionally including the generator defined for the EC_GROUP:
13689 scalar*generator + \sum scalars[i]*points[i].
13691 EC_POINT_mul is a simple wrapper function for the typical case
13692 that the point list has just one item (besides the optional
13697 * First EC_METHODs for curves over GF(p):
13699 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13700 operations and provides various method functions that can also
13701 operate with faster implementations of modular arithmetic.
13703 EC_GFp_mont_method() reuses most functions that are part of
13704 EC_GFp_simple_method, but uses Montgomery arithmetic.
13706 *Bodo Moeller; point addition and point doubling
13707 implementation directly derived from source code provided by
13708 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13710 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13711 crypto/ec/ec_lib.c):
13713 Curves are EC_GROUP objects (with an optional group generator)
13714 based on EC_METHODs that are built into the library.
13716 Points are EC_POINT objects based on EC_GROUP objects.
13718 Most of the framework would be able to handle curves over arbitrary
13719 finite fields, but as there are no obvious types for fields other
13720 than GF(p), some functions are limited to that for now.
13724 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13725 that the file contains a complete HTTP response.
13729 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13730 change the def and num file printf format specifier from "%-40sXXX"
13731 to "%-39s XXX". The latter will always guarantee a space after the
13732 field while the former will cause them to run together if the field
13733 is 40 of more characters long.
13737 * Constify the cipher and digest 'method' functions and structures
13738 and modify related functions to take constant EVP_MD and EVP_CIPHER
13743 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13744 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13748 * Modify `EVP_Digest*()` routines so they now return values. Although the
13749 internal software routines can never fail additional hardware versions
13754 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13756 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13757 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13760 ERR_R_NESTED_ASN1_ERROR
13762 ERR_R_MISSING_ASN1_EOS
13763 were 4 .. 9, conflicting with
13764 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13766 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13767 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13769 Add new error code 'ERR_R_INTERNAL_ERROR'.
13773 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13778 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13779 sets the subject name for a new request or supersedes the
13780 subject name in a given request. Formats that can be parsed are
13781 'CN=Some Name, OU=myOU, C=IT'
13783 'CN=Some Name/OU=myOU/C=IT'.
13785 Add options '-batch' and '-verbose' to 'openssl req'.
13787 *Massimiliano Pala <madwolf@hackmasters.net>*
13789 * Introduce the possibility to access global variables through
13790 functions on platform were that's the best way to handle exporting
13791 global variables in shared libraries. To enable this functionality,
13792 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13793 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13794 is normally done by Configure or something similar).
13796 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13797 in the source file (foo.c) like this:
13799 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13800 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13802 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13803 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13805 OPENSSL_DECLARE_GLOBAL(int,foo);
13806 #define foo OPENSSL_GLOBAL_REF(foo)
13807 OPENSSL_DECLARE_GLOBAL(double,bar);
13808 #define bar OPENSSL_GLOBAL_REF(bar)
13810 The #defines are very important, and therefore so is including the
13811 header file everywhere where the defined globals are used.
13813 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13814 of ASN.1 items, but that structure is a bit different.
13816 The largest change is in util/mkdef.pl which has been enhanced with
13817 better and easier to understand logic to choose which symbols should
13818 go into the Windows .def files as well as a number of fixes and code
13819 cleanup (among others, algorithm keywords are now sorted
13820 lexicographically to avoid constant rewrites).
13824 * In BN_div() keep a copy of the sign of 'num' before writing the
13825 result to 'rm' because if rm==num the value will be overwritten
13826 and produce the wrong result if 'num' is negative: this caused
13827 problems with BN_mod() and BN_nnmod().
13831 * Function OCSP_request_verify(). This checks the signature on an
13832 OCSP request and verifies the signer certificate. The signer
13833 certificate is just checked for a generic purpose and OCSP request
13838 * Add OCSP_check_validity() function to check the validity of OCSP
13839 responses. OCSP responses are prepared in real time and may only
13840 be a few seconds old. Simply checking that the current time lies
13841 between thisUpdate and nextUpdate max reject otherwise valid responses
13842 caused by either OCSP responder or client clock inaccuracy. Instead
13843 we allow thisUpdate and nextUpdate to fall within a certain period of
13844 the current time. The age of the response can also optionally be
13845 checked. Two new options -validity_period and -status_age added to
13850 * If signature or public key algorithm is unrecognized print out its
13851 OID rather that just UNKNOWN.
13855 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13856 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13857 ID to be generated from the issuer certificate alone which can then be
13858 passed to OCSP_id_issuer_cmp().
13862 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13863 ASN1 modules to export functions returning ASN1_ITEM pointers
13864 instead of the ASN1_ITEM structures themselves. This adds several
13865 new macros which allow the underlying ASN1 function/structure to
13866 be accessed transparently. As a result code should not use ASN1_ITEM
13867 references directly (such as &X509_it) but instead use the relevant
13868 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13869 use of the new ASN1 code on platforms where exporting structures
13870 is problematical (for example in shared libraries) but exporting
13871 functions returning pointers to structures is not.
13875 * Add support for overriding the generation of SSL/TLS session IDs.
13876 These callbacks can be registered either in an SSL_CTX or per SSL.
13877 The purpose of this is to allow applications to control, if they wish,
13878 the arbitrary values chosen for use as session IDs, particularly as it
13879 can be useful for session caching in multiple-server environments. A
13880 command-line switch for testing this (and any client code that wishes
13881 to use such a feature) has been added to "s_server".
13883 *Geoff Thorpe, Lutz Jaenicke*
13885 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13886 of the form `#if defined(...) || defined(...) || ...` and
13887 `#if !defined(...) && !defined(...) && ...`. This also avoids
13888 the growing number of special cases it was previously handling.
13892 * Make all configuration macros available for application by making
13893 sure they are available in opensslconf.h, by giving them names starting
13894 with `OPENSSL_` to avoid conflicts with other packages and by making
13895 sure e_os2.h will cover all platform-specific cases together with
13897 Additionally, it is now possible to define configuration/platform-
13898 specific names (called "system identities"). In the C code, these
13899 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13900 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13901 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13906 * New option -set_serial to 'req' and 'x509' this allows the serial
13907 number to use to be specified on the command line. Previously self
13908 signed certificates were hard coded with serial number 0 and the
13909 CA options of 'x509' had to use a serial number in a file which was
13914 * New options to 'ca' utility to support V2 CRL entry extensions.
13915 Currently CRL reason, invalidity date and hold instruction are
13916 supported. Add new CRL extensions to V3 code and some new objects.
13920 * New function EVP_CIPHER_CTX_set_padding() this is used to
13921 disable standard block padding (aka PKCS#5 padding) in the EVP
13922 API, which was previously mandatory. This means that the data is
13923 not padded in any way and so the total length much be a multiple
13924 of the block size, otherwise an error occurs.
13928 * Initial (incomplete) OCSP SSL support.
13932 * New function OCSP_parse_url(). This splits up a URL into its host,
13933 port and path components: primarily to parse OCSP URLs. New -url
13934 option to ocsp utility.
13938 * New nonce behavior. The return value of OCSP_check_nonce() now
13939 reflects the various checks performed. Applications can decide
13940 whether to tolerate certain situations such as an absent nonce
13941 in a response when one was present in a request: the ocsp application
13942 just prints out a warning. New function OCSP_add1_basic_nonce()
13943 this is to allow responders to include a nonce in a response even if
13944 the request is nonce-less.
13948 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13949 skipped when using openssl x509 multiple times on a single input file,
13950 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13954 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13955 set string type: to handle setting ASN1_TIME structures. Fix ca
13956 utility to correctly initialize revocation date of CRLs.
13960 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13961 the clients preferred ciphersuites and rather use its own preferences.
13962 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13963 Internet Explorer by ensuring unchanged hash method during stepup.
13964 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13968 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13969 to aes and add a new 'exist' option to print out symbols that don't
13974 * Additional options to ocsp utility to allow flags to be set and
13975 additional certificates supplied.
13979 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13980 OCSP client a number of certificate to only verify the response
13985 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13986 handle the new API. Currently only ECB, CBC modes supported. Add new
13989 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13990 Encryption Standard (AES) Ciphersuites for Transport Layer
13991 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13992 not enabled by default and were not part of the "ALL" ciphersuite
13993 alias because they were not yet official; they could be
13994 explicitly requested by specifying the "AESdraft" ciphersuite
13995 group alias. In the final release of OpenSSL 0.9.7, the group
13996 alias is called "AES" and is part of "ALL".)
13998 *Ben Laurie, Steve Henson, Bodo Moeller*
14000 * New function OCSP_copy_nonce() to copy nonce value (if present) from
14001 request to response.
14005 * Functions for OCSP responders. OCSP_request_onereq_count(),
14006 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
14007 extract information from a certificate request. OCSP_response_create()
14008 creates a response and optionally adds a basic response structure.
14009 OCSP_basic_add1_status() adds a complete single response to a basic
14010 response and returns the OCSP_SINGLERESP structure just added (to allow
14011 extensions to be included for example). OCSP_basic_add1_cert() adds a
14012 certificate to a basic response and OCSP_basic_sign() signs a basic
14013 response with various flags. New helper functions ASN1_TIME_check()
14014 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
14015 (converts ASN1_TIME to GeneralizedTime).
14019 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
14020 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
14021 structure from a certificate. X509_pubkey_digest() digests the public_key
14022 contents: this is used in various key identifiers.
14026 * Make sk_sort() tolerate a NULL argument.
14028 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14030 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14031 passed by the function are trusted implicitly. If any of them signed the
14032 response then it is assumed to be valid and is not verified.
14036 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14037 to data. This was previously part of the PKCS7 ASN1 code. This
14038 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14039 *Steve Henson, reported by Kenneth R. Robinette
14040 <support@securenetterm.com>*
14042 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14043 routines: without these tracing memory leaks is very painful.
14044 Fix leaks in PKCS12 and PKCS7 routines.
14048 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14049 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14050 effectively meant GeneralizedTime would never be used. Now it
14051 is initialised to -1 but X509_time_adj() now has to check the value
14052 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14053 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14054 *Steve Henson, reported by Kenneth R. Robinette
14055 <support@securenetterm.com>*
14057 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14058 result in a zero length in the ASN1_INTEGER structure which was
14059 not consistent with the structure when d2i_ASN1_INTEGER() was used
14060 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14061 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14062 where it did not print out a minus for negative ASN1_INTEGER.
14066 * Add summary printout to ocsp utility. The various functions which
14067 convert status values to strings have been renamed to:
14068 OCSP_response_status_str(), OCSP_cert_status_str() and
14069 OCSP_crl_reason_str() and are no longer static. New options
14070 to verify nonce values and to disable verification. OCSP response
14071 printout format cleaned up.
14075 * Add additional OCSP certificate checks. These are those specified
14076 in RFC2560. This consists of two separate checks: the CA of the
14077 certificate being checked must either be the OCSP signer certificate
14078 or the issuer of the OCSP signer certificate. In the latter case the
14079 OCSP signer certificate must contain the OCSP signing extended key
14080 usage. This check is performed by attempting to match the OCSP
14081 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14082 in the OCSP_CERTID structures of the response.
14086 * Initial OCSP certificate verification added to OCSP_basic_verify()
14087 and related routines. This uses the standard OpenSSL certificate
14088 verify routines to perform initial checks (just CA validity) and
14089 to obtain the certificate chain. Then additional checks will be
14090 performed on the chain. Currently the root CA is checked to see
14091 if it is explicitly trusted for OCSP signing. This is used to set
14092 a root CA as a global signing root: that is any certificate that
14093 chains to that CA is an acceptable OCSP signing certificate.
14097 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14098 extensions from a separate configuration file.
14099 As when reading extensions from the main configuration file,
14100 the '-extensions ...' option may be used for specifying the
14103 *Massimiliano Pala <madwolf@comune.modena.it>*
14105 * New OCSP utility. Allows OCSP requests to be generated or
14106 read. The request can be sent to a responder and the output
14107 parsed, outputted or printed in text form. Not complete yet:
14108 still needs to check the OCSP response validity.
14112 * New subcommands for 'openssl ca':
14113 `openssl ca -status <serial>` prints the status of the cert with
14114 the given serial number (according to the index file).
14115 `openssl ca -updatedb` updates the expiry status of certificates
14118 *Massimiliano Pala <madwolf@comune.modena.it>*
14120 * New '-newreq-nodes' command option to CA.pl. This is like
14121 '-newreq', but calls 'openssl req' with the '-nodes' option
14122 so that the resulting key is not encrypted.
14124 *Damien Miller <djm@mindrot.org>*
14126 * New configuration for the GNU Hurd.
14128 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14130 * Initial code to implement OCSP basic response verify. This
14131 is currently incomplete. Currently just finds the signer's
14132 certificate and verifies the signature on the response.
14136 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14137 value of OPENSSLDIR. This is available via the new '-d' option
14138 to 'openssl version', and is also included in 'openssl version -a'.
14142 * Allowing defining memory allocation callbacks that will be given
14143 file name and line number information in additional arguments
14144 (a `const char*` and an int). The basic functionality remains, as
14145 well as the original possibility to just replace malloc(),
14146 realloc() and free() by functions that do not know about these
14147 additional arguments. To register and find out the current
14148 settings for extended allocation functions, the following
14149 functions are provided:
14151 CRYPTO_set_mem_ex_functions
14152 CRYPTO_set_locked_mem_ex_functions
14153 CRYPTO_get_mem_ex_functions
14154 CRYPTO_get_locked_mem_ex_functions
14156 These work the same way as CRYPTO_set_mem_functions and friends.
14157 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14158 extended allocation function is enabled.
14159 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14160 a conventional allocation function is enabled.
14162 *Richard Levitte, Bodo Moeller*
14164 * Finish off removing the remaining LHASH function pointer casts.
14165 There should no longer be any prototype-casting required when using
14166 the LHASH abstraction, and any casts that remain are "bugs". See
14167 the callback types and macros at the head of lhash.h for details
14168 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14172 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14173 If /dev/[u]random devices are not available or do not return enough
14174 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14176 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14177 /etc/entropy will be queried once each in this sequence, querying stops
14178 when enough entropy was collected without querying more sockets.
14182 * Change the Unix RAND_poll() variant to be able to poll several
14183 random devices, as specified by DEVRANDOM, until a sufficient amount
14184 of data has been collected. We spend at most 10 ms on each file
14185 (select timeout) and read in non-blocking mode. DEVRANDOM now
14186 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14187 (previously it was just the string "/dev/urandom"), so on typical
14188 platforms the 10 ms delay will never occur.
14189 Also separate out the Unix variant to its own file, rand_unix.c.
14190 For VMS, there's a currently-empty rand_vms.c.
14194 * Move OCSP client related routines to ocsp_cl.c. These
14195 provide utility functions which an application needing
14196 to issue a request to an OCSP responder and analyse the
14197 response will typically need: as opposed to those which an
14198 OCSP responder itself would need which will be added later.
14200 OCSP_request_sign() signs an OCSP request with an API similar
14201 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14202 response. OCSP_response_get1_basic() extracts basic response
14203 from response. OCSP_resp_find_status(): finds and extracts status
14204 information from an OCSP_CERTID structure (which will be created
14205 when the request structure is built). These are built from lower
14206 level functions which work on OCSP_SINGLERESP structures but
14207 won't normally be used unless the application wishes to examine
14208 extensions in the OCSP response for example.
14210 Replace nonce routines with a pair of functions.
14211 OCSP_request_add1_nonce() adds a nonce value and optionally
14212 generates a random value. OCSP_check_nonce() checks the
14213 validity of the nonce in an OCSP response.
14217 * Change function OCSP_request_add() to OCSP_request_add0_id().
14218 This doesn't copy the supplied OCSP_CERTID and avoids the
14219 need to free up the newly created id. Change return type
14220 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14221 This can then be used to add extensions to the request.
14222 Deleted OCSP_request_new(), since most of its functionality
14223 is now in OCSP_REQUEST_new() (and the case insensitive name
14224 clash) apart from the ability to set the request name which
14225 will be added elsewhere.
14229 * Update OCSP API. Remove obsolete extensions argument from
14230 various functions. Extensions are now handled using the new
14231 OCSP extension code. New simple OCSP HTTP function which
14232 can be used to send requests and parse the response.
14236 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14237 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14238 uses the special reorder version of SET OF to sort the attributes
14239 and reorder them to match the encoded order. This resolves a long
14240 standing problem: a verify on a PKCS7 structure just after signing
14241 it used to fail because the attribute order did not match the
14242 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14243 it uses the received order. This is necessary to tolerate some broken
14244 software that does not order SET OF. This is handled by encoding
14245 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14246 to produce the required SET OF.
14250 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14251 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14252 files to get correct declarations of the ASN.1 item variables.
14256 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14257 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14258 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14259 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14260 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14261 ASN1_ITEM and no wrapper functions.
14265 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14266 replace the old function pointer based I/O routines. Change most of
14267 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14271 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14272 lines, recognize more "algorithms" that can be deselected, and make
14273 it complain about algorithm deselection that isn't recognised.
14277 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14278 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14279 to use new functions. Add NO_ASN1_OLD which can be set to remove
14280 some old style ASN1 functions: this can be used to determine if old
14281 code will still work when these eventually go away.
14285 * New extension functions for OCSP structures, these follow the
14286 same conventions as certificates and CRLs.
14290 * New function X509V3_add1_i2d(). This automatically encodes and
14291 adds an extension. Its behaviour can be customised with various
14292 flags to append, replace or delete. Various wrappers added for
14293 certificates and CRLs.
14297 * Fix to avoid calling the underlying ASN1 print routine when
14298 an extension cannot be parsed. Correct a typo in the
14299 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14303 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14304 entries for variables.
14308 * Add functionality to `apps/openssl.c` for detecting locking
14309 problems: As the program is single-threaded, all we have
14310 to do is register a locking callback using an array for
14311 storing which locks are currently held by the program.
14315 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14316 SSL_get_ex_data_X509_STORE_idx(), which is used in
14317 ssl_verify_cert_chain() and thus can be called at any time
14318 during TLS/SSL handshakes so that thread-safety is essential.
14319 Unfortunately, the ex_data design is not at all suited
14320 for multi-threaded use, so it probably should be abolished.
14324 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14326 *Broadcom, tweaked and integrated by Geoff Thorpe*
14328 * Move common extension printing code to new function
14329 X509V3_print_extensions(). Reorganise OCSP print routines and
14330 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14334 * New function X509_signature_print() to remove duplication in some
14339 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14340 set (this was treated exactly the same as SET OF previously). This
14341 is used to reorder the STACK representing the structure to match the
14342 encoding. This will be used to get round a problem where a PKCS7
14343 structure which was signed could not be verified because the STACK
14344 order did not reflect the encoded order.
14348 * Reimplement the OCSP ASN1 module using the new code.
14352 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14353 for its ASN1 operations. The old style function pointers still exist
14354 for now but they will eventually go away.
14358 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14359 completely replaces the old ASN1 functionality with a table driven
14360 encoder and decoder which interprets an ASN1_ITEM structure describing
14361 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14362 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14363 has also been converted to the new form.
14367 * Change BN_mod_exp_recp so that negative moduli are tolerated
14368 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14369 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14370 for negative moduli.
14374 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14375 of not touching the result's sign bit.
14379 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14384 * Changed the LHASH code to use prototypes for callbacks, and created
14385 macros to declare and implement thin (optionally static) functions
14386 that provide type-safety and avoid function pointer casting for the
14387 type-specific callbacks.
14391 * Added Kerberos Cipher Suites to be used with TLS, as written in
14393 *Veers Staats <staatsvr@asc.hpc.mil>,
14394 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14396 * Reformat the FAQ so the different questions and answers can be divided
14397 in sections depending on the subject.
14401 * Have the zlib compression code load ZLIB.DLL dynamically under
14406 * New function BN_mod_sqrt for computing square roots modulo a prime
14407 (using the probabilistic Tonelli-Shanks algorithm unless
14408 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14409 be handled deterministically).
14411 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14413 * Make BN_mod_inverse faster by explicitly handling small quotients
14414 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14415 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14419 * New function BN_kronecker.
14423 * Fix BN_gcd so that it works on negative inputs; the result is
14424 positive unless both parameters are zero.
14425 Previously something reasonably close to an infinite loop was
14426 possible because numbers could be growing instead of shrinking
14427 in the implementation of Euclid's algorithm.
14431 * Fix BN_is_word() and BN_is_one() macros to take into account the
14432 sign of the number in question.
14434 Fix BN_is_word(a,w) to work correctly for w == 0.
14436 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14437 because its test if the absolute value of 'a' equals 'w'.
14438 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14439 it exists mostly for use in the implementations of BN_is_zero(),
14440 BN_is_one(), and BN_is_word().
14444 * New function BN_swap.
14448 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14449 the exponentiation functions are more likely to produce reasonable
14450 results on negative inputs.
14454 * Change BN_mod_mul so that the result is always non-negative.
14455 Previously, it could be negative if one of the factors was negative;
14456 I don't think anyone really wanted that behaviour.
14460 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14461 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14462 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14463 and add new functions:
14472 BN_mod_lshift1_quick
14474 BN_mod_lshift_quick
14476 These functions always generate non-negative results.
14478 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14479 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14481 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14482 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14483 be reduced modulo `m`.
14485 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14488 The following entry accidentally appeared in the CHANGES file
14489 distributed with OpenSSL 0.9.7. The modifications described in
14490 it do *not* apply to OpenSSL 0.9.7.
14492 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14493 was actually never needed) and in BN_mul(). The removal in BN_mul()
14494 required a small change in bn_mul_part_recursive() and the addition
14495 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14496 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14497 bn_sub_words() and bn_add_words() except they take arrays with
14503 * In 'openssl passwd', verify passwords read from the terminal
14504 unless the '-salt' option is used (which usually means that
14505 verification would just waste user's time since the resulting
14506 hash is going to be compared with some given password hash)
14507 or the new '-noverify' option is used.
14509 This is an incompatible change, but it does not affect
14510 non-interactive use of 'openssl passwd' (passwords on the command
14511 line, '-stdin' option, '-in ...' option) and thus should not
14512 cause any problems.
14516 * Remove all references to RSAref, since there's no more need for it.
14520 * Make DSO load along a path given through an environment variable
14521 (SHLIB_PATH) with shl_load().
14525 * Constify the ENGINE code as a result of BIGNUM constification.
14526 Also constify the RSA code and most things related to it. In a
14527 few places, most notable in the depth of the ASN.1 code, ugly
14528 casts back to non-const were required (to be solved at a later
14533 * Make it so the openssl application has all engines loaded by default.
14537 * Constify the BIGNUM routines a little more.
14541 * Add the following functions:
14543 ENGINE_load_cswift()
14545 ENGINE_load_atalla()
14546 ENGINE_load_nuron()
14547 ENGINE_load_builtin_engines()
14549 That way, an application can itself choose if external engines that
14550 are built-in in OpenSSL shall ever be used or not. The benefit is
14551 that applications won't have to be linked with libdl or other dso
14552 libraries unless it's really needed.
14554 Changed 'openssl engine' to load all engines on demand.
14555 Changed the engine header files to avoid the duplication of some
14556 declarations (they differed!).
14560 * 'openssl engine' can now list capabilities.
14564 * Better error reporting in 'openssl engine'.
14568 * Never call load_dh_param(NULL) in s_server.
14572 * Add engine application. It can currently list engines by name and
14573 identity, and test if they are actually available.
14577 * Improve RPM specification file by forcing symbolic linking and making
14578 sure the installed documentation is also owned by root.root.
14580 *Damien Miller <djm@mindrot.org>*
14582 * Give the OpenSSL applications more possibilities to make use of
14583 keys (public as well as private) handled by engines.
14587 * Add OCSP code that comes from CertCo.
14591 * Add VMS support for the Rijndael code.
14595 * Added untested support for Nuron crypto accelerator.
14599 * Add support for external cryptographic devices. This code was
14600 previously distributed separately as the "engine" branch.
14602 *Geoff Thorpe, Richard Levitte*
14604 * Rework the filename-translation in the DSO code. It is now possible to
14605 have far greater control over how a "name" is turned into a filename
14606 depending on the operating environment and any oddities about the
14607 different shared library filenames on each system.
14611 * Support threads on FreeBSD-elf in Configure.
14615 * Fix for SHA1 assembly problem with MASM: it produces
14616 warnings about corrupt line number information when assembling
14617 with debugging information. This is caused by the overlapping
14620 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14623 NCONF_get_number() has no error checking at all. As a replacement,
14624 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14625 promoted strongly. The old NCONF_get_number is kept around for
14626 binary backward compatibility.
14627 Make it possible for methods to load from something other than a BIO,
14628 by providing a function pointer that is given a name instead of a BIO.
14629 For example, this could be used to load configuration data from an
14634 * Fix for non blocking accept BIOs. Added new I/O special reason
14635 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14636 with non blocking I/O was not possible because no retry code was
14637 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14642 * Added the beginnings of Rijndael support.
14646 * Fix for bug in DirectoryString mask setting. Add support for
14647 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14648 to allow certificate printing to more controllable, additional
14649 'certopt' option to 'x509' to allow new printing options to be
14654 * Clean old EAY MD5 hack from e_os.h.
14658 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14660 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14661 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14663 *Joe Orton, Steve Henson*
14665 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14667 * Fix additional bug revealed by the NISCC test suite:
14669 Stop bug triggering large recursion when presented with
14670 certain ASN.1 tags ([CVE-2003-0851])
14674 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14676 * Fix various bugs revealed by running the NISCC test suite:
14678 Stop out of bounds reads in the ASN1 code when presented with
14679 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14681 If verify callback ignores invalid public key errors don't try to check
14682 certificate signature with the NULL public key.
14686 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14687 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14692 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14693 extra data after the compression methods not only for TLS 1.0
14694 but also for SSL 3.0 (as required by the specification).
14696 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14698 * Change X509_certificate_type() to mark the key as exported/exportable
14699 when it's 512 *bits* long, not 512 bytes.
14703 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14705 * Countermeasure against the Klima-Pokorny-Rosa extension of
14706 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14707 a protocol version number mismatch like a decryption error
14708 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14712 * Turn on RSA blinding by default in the default implementation
14713 to avoid a timing attack. Applications that don't want it can call
14714 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14715 They would be ill-advised to do so in most cases.
14717 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14719 * Change RSA blinding code so that it works when the PRNG is not
14720 seeded (in this case, the secret RSA exponent is abused as
14721 an unpredictable seed -- if it is not unpredictable, there
14722 is no point in blinding anyway). Make RSA blinding thread-safe
14723 by remembering the creator's thread ID in rsa->blinding and
14724 having all other threads use local one-time blinding factors
14725 (this requires more computation than sharing rsa->blinding, but
14726 avoids excessive locking; and if an RSA object is not shared
14727 between threads, blinding will still be very fast).
14731 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14733 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14734 via timing by performing a MAC computation even if incorrect
14735 block cipher padding has been found. This is a countermeasure
14736 against active attacks where the attacker has to distinguish
14737 between bad padding and a MAC verification error. ([CVE-2003-0078])
14739 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14740 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14741 Martin Vuagnoux (EPFL, Ilion)*
14743 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14745 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14746 memory from its contents. This is done with a counter that will
14747 place alternating values in each byte. This can be used to solve
14748 two issues: 1) the removal of calls to memset() by highly optimizing
14749 compilers, and 2) cleansing with other values than 0, since those can
14750 be read through on certain media, for example a swap space on disk.
14754 * Bugfix: client side session caching did not work with external caching,
14755 because the session->cipher setting was not restored when reloading
14756 from the external cache. This problem was masked, when
14757 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14758 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14762 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14763 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14765 *Zeev Lieber <zeev-l@yahoo.com>*
14767 * Undo an undocumented change introduced in 0.9.6e which caused
14768 repeated calls to OpenSSL_add_all_ciphers() and
14769 OpenSSL_add_all_digests() to be ignored, even after calling
14774 * Change the default configuration reader to deal with last line not
14775 being properly terminated.
14779 * Change X509_NAME_cmp() so it applies the special rules on handling
14780 DN values that are of type PrintableString, as well as RDNs of type
14781 emailAddress where the value has the type ia5String.
14783 *stefank@valicert.com via Richard Levitte*
14785 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14786 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14787 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14788 the bitwise-OR of the two for use by the majority of applications
14789 wanting this behaviour, and update the docs. The documented
14790 behaviour and actual behaviour were inconsistent and had been
14791 changing anyway, so this is more a bug-fix than a behavioural
14794 *Geoff Thorpe, diagnosed by Nadav Har'El*
14796 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14797 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14801 * Fix initialization code race conditions in
14802 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14803 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14804 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14805 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14806 ssl2_get_cipher_by_char(),
14807 ssl3_get_cipher_by_char().
14809 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14811 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14812 the cached sessions are flushed, as the remove_cb() might use ex_data
14813 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14814 (see [openssl.org #212]).
14816 *Geoff Thorpe, Lutz Jaenicke*
14818 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14819 length, instead of the encoding length to d2i_ASN1_OBJECT.
14823 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14825 * [In 0.9.6g-engine release:]
14826 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14828 *Lynn Gazis <lgazis@rainbow.com>*
14830 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14832 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14833 and get fix the header length calculation.
14834 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14835 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14837 * Use proper error handling instead of 'assertions' in buffer
14838 overflow checks added in 0.9.6e. This prevents DoS (the
14839 assertions could call abort()).
14841 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14843 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14845 * Add various sanity checks to asn1_get_length() to reject
14846 the ASN1 length bytes if they exceed sizeof(long), will appear
14847 negative or the content length exceeds the length of the
14850 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14852 * Fix cipher selection routines: ciphers without encryption had no flags
14853 for the cipher strength set and where therefore not handled correctly
14854 by the selection routines (PR #130).
14858 * Fix EVP_dsa_sha macro.
14863 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14864 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14865 that was added in OpenSSL 0.9.6d.
14867 As the countermeasure turned out to be incompatible with some
14868 broken SSL implementations, the new option is part of SSL_OP_ALL.
14869 SSL_OP_ALL is usually employed when compatibility with weird SSL
14870 implementations is desired (e.g. '-bugs' option to 's_client' and
14871 's_server'), so the new option is automatically set in many
14876 * Changes in security patch:
14878 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14879 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14880 Air Force Materiel Command, USAF, under agreement number
14883 * Add various sanity checks to asn1_get_length() to reject
14884 the ASN1 length bytes if they exceed sizeof(long), will appear
14885 negative or the content length exceeds the length of the
14886 supplied buffer. ([CVE-2002-0659])
14888 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14890 * Assertions for various potential buffer overflows, not known to
14891 happen in practice.
14893 *Ben Laurie (CHATS)*
14895 * Various temporary buffers to hold ASCII versions of integers were
14896 too small for 64 bit platforms. ([CVE-2002-0655])
14897 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14899 * Remote buffer overflow in SSL3 protocol - an attacker could
14900 supply an oversized session ID to a client. ([CVE-2002-0656])
14902 *Ben Laurie (CHATS)*
14904 * Remote buffer overflow in SSL2 protocol - an attacker could
14905 supply an oversized client master key. ([CVE-2002-0656])
14907 *Ben Laurie (CHATS)*
14909 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14911 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14912 encoded as NULL) with id-dsa-with-sha1.
14914 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14916 * Check various `X509_...()` return values in `apps/req.c`.
14918 *Nils Larsch <nla@trustcenter.de>*
14920 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14921 an end-of-file condition would erroneously be flagged, when the CRLF
14922 was just at the end of a processed block. The bug was discovered when
14923 processing data through a buffering memory BIO handing the data to a
14924 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14925 <ptsekov@syntrex.com> and Nedelcho Stanev.
14929 * Implement a countermeasure against a vulnerability recently found
14930 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14931 before application data chunks to avoid the use of known IVs
14932 with data potentially chosen by the attacker.
14936 * Fix length checks in ssl3_get_client_hello().
14940 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14941 to prevent ssl3_read_internal() from incorrectly assuming that
14942 ssl3_read_bytes() found application data while handshake
14943 processing was enabled when in fact s->s3->in_read_app_data was
14944 merely automatically cleared during the initial handshake.
14946 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14948 * Fix object definitions for Private and Enterprise: they were not
14949 recognized in their shortname (=lowercase) representation. Extend
14950 obj_dat.pl to issue an error when using undefined keywords instead
14951 of silently ignoring the problem (Svenning Sorensen
14952 <sss@sss.dnsalias.net>).
14956 * Fix DH_generate_parameters() so that it works for 'non-standard'
14957 generators, i.e. generators other than 2 and 5. (Previously, the
14958 code did not properly initialise the 'add' and 'rem' values to
14959 BN_generate_prime().)
14961 In the new general case, we do not insist that 'generator' is
14962 actually a primitive root: This requirement is rather pointless;
14963 a generator of the order-q subgroup is just as good, if not
14968 * Map new X509 verification errors to alerts. Discovered and submitted by
14969 Tom Wu <tom@arcot.com>.
14973 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14974 returning non-zero before the data has been completely received
14975 when using non-blocking I/O.
14977 *Bodo Moeller; problem pointed out by John Hughes*
14979 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14981 *Ben Laurie, Lutz Jaenicke*
14983 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14984 Yoram Zahavi <YoramZ@gilian.com>).
14988 * Add information about CygWin 1.3 and on, and preserve proper
14989 configuration for the versions before that.
14991 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14993 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14994 check whether we deal with a copy of a session and do not delete from
14995 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14996 <izhar@checkpoint.com>.
15000 * Do not store session data into the internal session cache, if it
15001 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
15002 flag is set). Proposed by Aslam <aslam@funk.com>.
15006 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
15011 * [In 0.9.6d-engine release:]
15012 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
15014 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
15016 * Add the configuration target linux-s390x.
15018 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
15020 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
15021 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
15022 variable as an indication that a ClientHello message has been
15023 received. As the flag value will be lost between multiple
15024 invocations of ssl3_accept when using non-blocking I/O, the
15025 function may not be aware that a handshake has actually taken
15026 place, thus preventing a new session from being added to the
15029 To avoid this problem, we now set s->new_session to 2 instead of
15030 using a local variable.
15032 *Lutz Jaenicke, Bodo Moeller*
15034 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15035 if the SSL_R_LENGTH_MISMATCH error is detected.
15037 *Geoff Thorpe, Bodo Moeller*
15039 * New 'shared_ldflag' column in Configure platform table.
15043 * Fix EVP_CIPHER_mode macro.
15045 *"Dan S. Camper" <dan@bti.net>*
15047 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15048 type, we must throw them away by setting rr->length to 0.
15050 *D P Chang <dpc@qualys.com>*
15052 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15054 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15055 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15056 worked incorrectly for those cases where range = `10..._2` and
15057 `3*range` is two bits longer than range.)
15061 * Only add signing time to PKCS7 structures if it is not already
15066 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15067 OBJ_ld_ce should be OBJ_id_ce.
15068 Also some ip-pda OIDs in crypto/objects/objects.txt were
15069 incorrect (cf. RFC 3039).
15071 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15073 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15074 returns early because it has nothing to do.
15076 *Andy Schneider <andy.schneider@bjss.co.uk>*
15078 * [In 0.9.6c-engine release:]
15079 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15081 *Andy Schneider <andy.schneider@bjss.co.uk>*
15083 * [In 0.9.6c-engine release:]
15084 Add support for Cryptographic Appliance's keyserver technology.
15085 (Use engine 'keyclient')
15087 *Cryptographic Appliances and Geoff Thorpe*
15089 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15090 is called via tools/c89.sh because arguments have to be
15091 rearranged (all '-L' options must appear before the first object
15094 *Richard Shapiro <rshapiro@abinitio.com>*
15096 * [In 0.9.6c-engine release:]
15097 Add support for Broadcom crypto accelerator cards, backported
15100 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15102 * [In 0.9.6c-engine release:]
15103 Add support for SureWare crypto accelerator cards from
15104 Baltimore Technologies. (Use engine 'sureware')
15106 *Baltimore Technologies and Mark Cox*
15108 * [In 0.9.6c-engine release:]
15109 Add support for crypto accelerator cards from Accelerated
15110 Encryption Processing, www.aep.ie. (Use engine 'aep')
15112 *AEP Inc. and Mark Cox*
15114 * Add a configuration entry for gcc on UnixWare.
15116 *Gary Benson <gbenson@redhat.com>*
15118 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15119 messages are stored in a single piece (fixed-length part and
15120 variable-length part combined) and fix various bugs found on the way.
15124 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15125 instead. BIO_gethostbyname() does not know what timeouts are
15126 appropriate, so entries would stay in cache even when they have
15128 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15130 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15131 faced with a pathologically small ClientHello fragment that does
15132 not contain client_version: Instead of aborting with an error,
15133 simply choose the highest available protocol version (i.e.,
15134 TLS 1.0 unless it is disabled). In practice, ClientHello
15135 messages are never sent like this, but this change gives us
15136 strictly correct behaviour at least for TLS.
15140 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15141 never resets s->method to s->ctx->method when called from within
15142 one of the SSL handshake functions.
15144 *Bodo Moeller; problem pointed out by Niko Baric*
15146 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15147 (sent using the client's version number) if client_version is
15148 smaller than the protocol version in use. Also change
15149 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15150 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15151 the client will at least see that alert.
15155 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15160 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15161 client receives HelloRequest while in a handshake.
15163 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15165 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15166 should end in 'break', not 'goto end' which circumvents various
15167 cleanups done in state SSL_ST_OK. But session related stuff
15168 must be disabled for SSL_ST_OK in the case that we just sent a
15171 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15172 before just sending a HelloRequest.
15174 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15176 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15177 reveal whether illegal block cipher padding was found or a MAC
15178 verification error occurred. (Neither SSLerr() codes nor alerts
15179 are directly visible to potential attackers, but the information
15180 may leak via logfiles.)
15182 Similar changes are not required for the SSL 2.0 implementation
15183 because the number of padding bytes is sent in clear for SSL 2.0,
15184 and the extra bytes are just ignored. However ssl/s2_pkt.c
15185 failed to verify that the purported number of padding bytes is in
15190 * Add OpenUNIX-8 support including shared libraries
15191 (Boyd Lynn Gerber <gerberb@zenez.com>).
15195 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15196 'wristwatch attack' using huge encoding parameters (cf.
15197 James H. Manger's CRYPTO 2001 paper). Note that the
15198 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15199 encoding parameters and hence was not vulnerable.
15203 * BN_sqr() bug fix.
15205 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15207 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15208 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15209 followed by modular reduction.
15211 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15213 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15214 equivalent based on BN_pseudo_rand() instead of BN_rand().
15218 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15219 This function was broken, as the check for a new client hello message
15220 to handle SGC did not allow these large messages.
15221 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15225 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15229 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15230 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15234 * Rework the configuration and shared library support for Tru64 Unix.
15235 The configuration part makes use of modern compiler features and
15236 still retains old compiler behavior for those that run older versions
15237 of the OS. The shared library support part includes a variant that
15238 uses the RPATH feature, and is available through the special
15239 configuration target "alpha-cc-rpath", which will never be selected
15242 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15244 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15245 with the same message size as in ssl3_get_certificate_request().
15246 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15247 messages might inadvertently be reject as too long.
15249 *Petr Lampa <lampa@fee.vutbr.cz>*
15251 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15255 * Modified SSL library such that the verify_callback that has been set
15256 specifically for an SSL object with SSL_set_verify() is actually being
15257 used. Before the change, a verify_callback set with this function was
15258 ignored and the verify_callback() set in the SSL_CTX at the time of
15259 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15260 to allow the necessary settings.
15264 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15265 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15266 done automatically (in contradiction to the requirements of the C
15267 standard). This made problems when used from OpenSSH.
15271 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15272 dh->length and always used
15274 BN_rand_range(priv_key, dh->p).
15276 BN_rand_range() is not necessary for Diffie-Hellman, and this
15277 specific range makes Diffie-Hellman unnecessarily inefficient if
15278 dh->length (recommended exponent length) is much smaller than the
15279 length of dh->p. We could use BN_rand_range() if the order of
15280 the subgroup was stored in the DH structure, but we only have
15285 BN_rand(priv_key, l, ...)
15287 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15294 RSA_eay_public_encrypt
15295 RSA_eay_private_decrypt
15296 RSA_eay_private_encrypt (signing)
15297 RSA_eay_public_decrypt (signature verification)
15299 (default implementations for RSA_public_encrypt,
15300 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15301 always reject numbers >= n.
15305 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15306 to synchronize access to 'locking_thread'. This is necessary on
15307 systems where access to 'locking_thread' (an 'unsigned long'
15308 variable) is not atomic.
15312 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15313 *before* setting the 'crypto_lock_rand' flag. The previous code had
15314 a race condition if 0 is a valid thread ID.
15316 *Travis Vitek <vitek@roguewave.com>*
15318 * Add support for shared libraries under Irix.
15320 *Albert Chin-A-Young <china@thewrittenword.com>*
15322 * Add configuration option to build on Linux on both big-endian and
15323 little-endian MIPS.
15325 *Ralf Baechle <ralf@uni-koblenz.de>*
15327 * Add the possibility to create shared libraries on HP-UX.
15331 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15333 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15334 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15335 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15336 PRNG state recovery was possible based on the output of
15337 one PRNG request appropriately sized to gain knowledge on
15338 'md' followed by enough consecutive 1-byte PRNG requests
15339 to traverse all of 'state'.
15341 1. When updating 'md_local' (the current thread's copy of 'md')
15342 during PRNG output generation, hash all of the previous
15343 'md_local' value, not just the half used for PRNG output.
15345 2. Make the number of bytes from 'state' included into the hash
15346 independent from the number of PRNG bytes requested.
15348 The first measure alone would be sufficient to avoid
15349 Markku-Juhani's attack. (Actually it had never occurred
15350 to me that the half of 'md_local' used for chaining was the
15351 half from which PRNG output bytes were taken -- I had always
15352 assumed that the secret half would be used.) The second
15353 measure makes sure that additional data from 'state' is never
15354 mixed into 'md_local' in small portions; this heuristically
15355 further strengthens the PRNG.
15359 * Fix crypto/bn/asm/mips3.s.
15363 * When only the key is given to "enc", the IV is undefined. Print out
15364 an error message in this case.
15368 * Handle special case when X509_NAME is empty in X509 printing routines.
15372 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15373 positive and less than q.
15377 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15378 used: it isn't thread safe and the add_lock_callback should handle
15381 *Paul Rose <Paul.Rose@bridge.com>*
15383 * Verify that incoming data obeys the block size in
15384 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15390 *Ulf Möller, Bodo Möller*
15392 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15393 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15394 when fixing the server behaviour for backwards-compatible 'client
15395 hello' messages. (Note that the attack is impractical against
15396 SSL 3.0 and TLS 1.0 anyway because length and version checking
15397 means that the probability of guessing a valid ciphertext is
15398 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15401 Before 0.9.5, the countermeasure (hide the error by generating a
15402 random 'decryption result') did not work properly because
15403 ERR_clear_error() was missing, meaning that SSL_get_error() would
15404 detect the supposedly ignored error.
15406 Both problems are now fixed.
15410 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15411 (previously it was 1024).
15415 * Fix for compatibility mode trust settings: ignore trust settings
15416 unless some valid trust or reject settings are present.
15420 * Fix for blowfish EVP: its a variable length cipher.
15424 * Fix various bugs related to DSA S/MIME verification. Handle missing
15425 parameters in DSA public key structures and return an error in the
15426 DSA routines if parameters are absent.
15430 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15431 in the current directory if neither $RANDFILE nor $HOME was set.
15432 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15433 caused some confusion to Windows users who haven't defined $HOME.
15434 Thus RAND_file_name() is changed again: e_os.h can define a
15435 DEFAULT_HOME, which will be used if $HOME is not set.
15436 For Windows, we use "C:"; on other platforms, we still require
15437 environment variables.
15439 * Move 'if (!initialized) RAND_poll()' into regions protected by
15440 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15441 having multiple threads call RAND_poll() concurrently.
15445 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15446 combination of a flag and a thread ID variable.
15447 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15448 flag), *other* threads can enter ssleay_add_bytes without obeying
15449 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15450 that they do not hold after the first thread unsets add_do_not_lock).
15454 * Change bctest again: '-x' expressions are not available in all
15455 versions of 'test'.
15459 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15461 * Fix a couple of memory leaks in PKCS7_dataDecode()
15463 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15465 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15466 the default extension for executables, if any. Also, make the perl
15467 scripts that use symlink() to test if it really exists and use "cp"
15468 if it doesn't. All this made OpenSSL compilable and installable in
15473 * Fix for asn1_GetSequence() for indefinite length constructed data.
15474 If SEQUENCE is length is indefinite just set c->slen to the total
15475 amount of data available.
15477 *Steve Henson, reported by shige@FreeBSD.org*
15479 *This change does not apply to 0.9.7.*
15481 * Change bctest to avoid here-documents inside command substitution
15482 (workaround for FreeBSD /bin/sh bug).
15483 For compatibility with Ultrix, avoid shell functions (introduced
15484 in the bctest version that searches along $PATH).
15488 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15489 with des_encrypt() defined on some operating systems, like Solaris
15494 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15495 On the Importance of Eliminating Errors in Cryptographic
15496 Computations, J. Cryptology 14 (2001) 2, 101-119,
15497 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15501 * MIPS assembler BIGNUM division bug fix.
15505 * Disabled incorrect Alpha assembler code.
15509 * Fix PKCS#7 decode routines so they correctly update the length
15510 after reading an EOC for the EXPLICIT tag.
15514 *This change does not apply to 0.9.7.*
15516 * Fix bug in PKCS#12 key generation routines. This was triggered
15517 if a 3DES key was generated with a 0 initial byte. Include
15518 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15519 (but broken) behaviour.
15523 * Enhance bctest to search for a working bc along $PATH and print
15526 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15528 * Fix memory leaks in err.c: free err_data string if necessary;
15529 don't write to the wrong index in ERR_set_error_data.
15533 * Implement ssl23_peek (analogous to ssl23_read), which previously
15538 * Replace rdtsc with `_emit` statements for VC++ version 5.
15540 *Jeremy Cooper <jeremy@baymoo.org>*
15542 * Make it possible to reuse SSLv2 sessions.
15546 * In copy_email() check for >= 0 as a return value for
15547 X509_NAME_get_index_by_NID() since 0 is a valid index.
15549 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15551 * Avoid coredump with unsupported or invalid public keys by checking if
15552 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15553 PKCS7_verify() fails with non detached data.
15557 * Don't use getenv in library functions when run as setuid/setgid.
15558 New function OPENSSL_issetugid().
15562 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15563 due to incorrect handling of multi-threading:
15565 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15567 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15569 3. Count how many times MemCheck_off() has been called so that
15570 nested use can be treated correctly. This also avoids
15571 inband-signalling in the previous code (which relied on the
15572 assumption that thread ID 0 is impossible).
15576 * Add "-rand" option also to s_client and s_server.
15580 * Fix CPU detection on Irix 6.x.
15581 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15582 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15584 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15589 *This change does not apply to 0.9.7.*
15591 * Use the cached encoding of an X509_NAME structure rather than
15592 copying it. This is apparently the reason for the libsafe "errors"
15593 but the code is actually correct.
15597 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15598 Bleichenbacher's DSA attack.
15599 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15600 to be set and top=0 forces the highest bit to be set; top=-1 is new
15601 and leaves the highest bit random.
15603 *Ulf Moeller, Bodo Moeller*
15605 * In the `NCONF_...`-based implementations for `CONF_...` queries
15606 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15607 a temporary CONF structure with the data component set to NULL
15608 (which gives segmentation faults in lh_retrieve).
15609 Instead, use NULL for the CONF pointer in CONF_get_string and
15610 CONF_get_number (which may use environment variables) and directly
15611 return NULL from CONF_get_section.
15615 * Fix potential buffer overrun for EBCDIC.
15619 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15620 keyUsage if basicConstraints absent for a CA.
15624 * Make SMIME_write_PKCS7() write mail header values with a format that
15625 is more generally accepted (no spaces before the semicolon), since
15626 some programs can't parse those values properly otherwise. Also make
15627 sure BIO's that break lines after each write do not create invalid
15632 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15633 macros previously used would not encode an empty SEQUENCE OF
15634 and break the signature.
15638 *This change does not apply to 0.9.7.*
15640 * Zero the premaster secret after deriving the master secret in
15645 * Add some EVP_add_digest_alias registrations (as found in
15646 OpenSSL_add_all_digests()) to SSL_library_init()
15647 aka OpenSSL_add_ssl_algorithms(). This provides improved
15648 compatibility with peers using X.509 certificates
15649 with unconventional AlgorithmIdentifier OIDs.
15653 * Fix for Irix with NO_ASM.
15655 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15657 * ./config script fixes.
15659 *Ulf Moeller, Richard Levitte*
15661 * Fix 'openssl passwd -1'.
15665 * Change PKCS12_key_gen_asc() so it can cope with non null
15666 terminated strings whose length is passed in the passlen
15667 parameter, for example from PEM callbacks. This was done
15668 by adding an extra length parameter to asc2uni().
15670 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15672 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15673 call failed, free the DSA structure.
15677 * Fix to uni2asc() to cope with zero length Unicode strings.
15678 These are present in some PKCS#12 files.
15682 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15683 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15684 when writing a 32767 byte record.
15686 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15688 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15689 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15691 (RSA objects have a reference count access to which is protected
15692 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15693 so they are meant to be shared between threads.)
15694 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15695 "Reddie, Steven" <Steven.Reddie@ca.com>*
15697 * Fix a deadlock in CRYPTO_mem_leaks().
15701 * Use better test patterns in bntest.
15705 * rand_win.c fix for Borland C.
15709 * BN_rshift bugfix for n == 0.
15713 * Add a 'bctest' script that checks for some known 'bc' bugs
15714 so that 'make test' does not abort just because 'bc' is broken.
15718 * Store verify_result within SSL_SESSION also for client side to
15719 avoid potential security hole. (Re-used sessions on the client side
15720 always resulted in verify_result==X509_V_OK, not using the original
15721 result of the server certificate verification.)
15725 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15726 SSL3_RT_APPLICATION_DATA, return 0.
15727 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15732 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15733 releases, have been re-implemented by renaming the previous
15734 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15735 and ssl3_read_internal, respectively, and adding 'peek' parameters
15736 to them. The new ssl[23]_{read,peek} functions are calls to
15737 ssl[23]_read_internal with the 'peek' flag set appropriately.
15738 A 'peek' parameter has also been added to ssl3_read_bytes, which
15739 does the actual work for ssl3_read_internal.
15743 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15744 the method-specific "init()" handler. Also clean up ex_data after
15745 calling the method-specific "finish()" handler. Previously, this was
15746 happening the other way round.
15750 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15751 The previous value, 12, was not always sufficient for BN_mod_exp().
15755 * Make sure that shared libraries get the internal name engine with
15756 the full version number and not just 0. This should mark the
15757 shared libraries as not backward compatible. Of course, this should
15758 be changed again when we can guarantee backward binary compatibility.
15762 * Fix typo in get_cert_by_subject() in by_dir.c
15764 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15766 * Rework the system to generate shared libraries:
15768 - Make note of the expected extension for the shared libraries and
15769 if there is a need for symbolic links from for example libcrypto.so.0
15770 to libcrypto.so.0.9.7. There is extended info in Configure for
15773 - Make as few rebuilds of the shared libraries as possible.
15775 - Still avoid linking the OpenSSL programs with the shared libraries.
15777 - When installing, install the shared libraries separately from the
15782 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15784 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15785 and not in SSL_clear because the latter is also used by the
15786 accept/connect functions; previously, the settings made by
15787 SSL_set_read_ahead would be lost during the handshake.
15789 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15791 * Correct util/mkdef.pl to be selective about disabled algorithms.
15792 Previously, it would create entries for disabled algorithms no
15797 * Added several new manual pages for SSL_* function.
15801 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15803 * In ssl23_get_client_hello, generate an error message when faced
15804 with an initial SSL 3.0/TLS record that is too small to contain the
15805 first two bytes of the ClientHello message, i.e. client_version.
15806 (Note that this is a pathologic case that probably has never happened
15807 in real life.) The previous approach was to use the version number
15808 from the record header as a substitute; but our protocol choice
15809 should not depend on that one because it is not authenticated
15810 by the Finished messages.
15814 * More robust randomness gathering functions for Windows.
15816 *Jeffrey Altman <jaltman@columbia.edu>*
15818 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15819 not set then we don't setup the error code for issuer check errors
15820 to avoid possibly overwriting other errors which the callback does
15821 handle. If an application does set the flag then we assume it knows
15822 what it is doing and can handle the new informational codes
15827 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15828 a general "ANY" type, as such it should be able to decode anything
15829 including tagged types. However it didn't check the class so it would
15830 wrongly interpret tagged types in the same way as their universal
15831 counterpart and unknown types were just rejected. Changed so that the
15832 tagged and unknown types are handled in the same way as a SEQUENCE:
15833 that is the encoding is stored intact. There is also a new type
15834 "V_ASN1_OTHER" which is used when the class is not universal, in this
15835 case we have no idea what the actual type is so we just lump them all
15840 * On VMS, stdout may very well lead to a file that is written to
15841 in a record-oriented fashion. That means that every write() will
15842 write a separate record, which will be read separately by the
15843 programs trying to read from it. This can be very confusing.
15845 The solution is to put a BIO filter in the way that will buffer
15846 text until a linefeed is reached, and then write everything a
15847 line at a time, so every record written will be an actual line,
15848 not chunks of lines and not (usually doesn't happen, but I've
15849 seen it once) several lines in one record. BIO_f_linebuffer() is
15852 Currently, it's a VMS-only method, because that's where it has
15853 been tested well enough.
15857 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15858 it can return incorrect results.
15859 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15860 but it was in 0.9.6-beta[12].)
15864 * Disable the check for content being present when verifying detached
15865 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15866 include zero length content when signing messages.
15870 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15871 BIO_ctrl (for BIO pairs).
15875 * Add DSO method for VMS.
15879 * Bug fix: Montgomery multiplication could produce results with the
15884 * Add RPM specification openssl.spec and modify it to build three
15885 packages. The default package contains applications, application
15886 documentation and run-time libraries. The devel package contains
15887 include files, static libraries and function documentation. The
15888 doc package contains the contents of the doc directory. The original
15889 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15893 * Add a large number of documentation files for many SSL routines.
15895 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15897 * Add a configuration entry for Sony News 4.
15899 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15901 * Don't set the two most significant bits to one when generating a
15902 random number < q in the DSA library.
15906 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15907 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15908 the underlying transport is blocking) if a handshake took place.
15909 (The default behaviour is needed by applications such as s_client
15910 and s_server that use select() to determine when to use SSL_read;
15911 but for applications that know in advance when to expect data, it
15912 just makes things more complicated.)
15916 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15921 * Add a few more EBCDIC conditionals that make `req` and `x509`
15922 work better on such systems.
15924 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15926 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15927 Update PKCS12_parse() so it copies the friendlyName and the
15928 keyid to the certificates aux info.
15932 * Fix bug in PKCS7_verify() which caused an infinite loop
15933 if there was more than one signature.
15935 *Sven Uszpelkat <su@celocom.de>*
15937 * Major change in util/mkdef.pl to include extra information
15938 about each symbol, as well as presenting variables as well
15939 as functions. This change means that there's n more need
15940 to rebuild the .num files when some algorithms are excluded.
15944 * Allow the verify time to be set by an application,
15945 rather than always using the current time.
15949 * Phase 2 verify code reorganisation. The certificate
15950 verify code now looks up an issuer certificate by a
15951 number of criteria: subject name, authority key id
15952 and key usage. It also verifies self signed certificates
15953 by the same criteria. The main comparison function is
15954 X509_check_issued() which performs these checks.
15956 Lot of changes were necessary in order to support this
15957 without completely rewriting the lookup code.
15959 Authority and subject key identifier are now cached.
15961 The LHASH 'certs' is X509_STORE has now been replaced
15962 by a STACK_OF(X509_OBJECT). This is mainly because an
15963 LHASH can't store or retrieve multiple objects with
15964 the same hash value.
15966 As a result various functions (which were all internal
15967 use only) have changed to handle the new X509_STORE
15968 structure. This will break anything that messed round
15969 with X509_STORE internally.
15971 The functions X509_STORE_add_cert() now checks for an
15972 exact match, rather than just subject name.
15974 The X509_STORE API doesn't directly support the retrieval
15975 of multiple certificates matching a given criteria, however
15976 this can be worked round by performing a lookup first
15977 (which will fill the cache with candidate certificates)
15978 and then examining the cache for matches. This is probably
15979 the best we can do without throwing out X509_LOOKUP
15980 entirely (maybe later...).
15982 The X509_VERIFY_CTX structure has been enhanced considerably.
15984 All certificate lookup operations now go via a get_issuer()
15985 callback. Although this currently uses an X509_STORE it
15986 can be replaced by custom lookups. This is a simple way
15987 to bypass the X509_STORE hackery necessary to make this
15988 work and makes it possible to use more efficient techniques
15989 in future. A very simple version which uses a simple
15990 STACK for its trusted certificate store is also provided
15991 using X509_STORE_CTX_trusted_stack().
15993 The verify_cb() and verify() callbacks now have equivalents
15994 in the X509_STORE_CTX structure.
15996 X509_STORE_CTX also has a 'flags' field which can be used
15997 to customise the verify behaviour.
16001 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
16002 excludes S/MIME capabilities.
16006 * When a certificate request is read in keep a copy of the
16007 original encoding of the signed data and use it when outputting
16008 again. Signatures then use the original encoding rather than
16009 a decoded, encoded version which may cause problems if the
16010 request is improperly encoded.
16014 * For consistency with other BIO_puts implementations, call
16015 buffer_write(b, ...) directly in buffer_puts instead of calling
16018 In BIO_puts, increment b->num_write as in BIO_write.
16020 *Peter.Sylvester@EdelWeb.fr*
16022 * Fix BN_mul_word for the case where the word is 0. (We have to use
16023 BN_zero, we may not return a BIGNUM with an array consisting of
16024 words set to zero.)
16028 * Avoid calling abort() from within the library when problems are
16029 detected, except if preprocessor symbols have been defined
16030 (such as REF_CHECK, BN_DEBUG etc.).
16034 * New openssl application 'rsautl'. This utility can be
16035 used for low-level RSA operations. DER public key
16036 BIO/fp routines also added.
16040 * New Configure entry and patches for compiling on QNX 4.
16042 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16044 * A demo state-machine implementation was sponsored by
16045 Nuron (<http://www.nuron.com/>) and is now available in
16046 demos/state_machine.
16050 * New options added to the 'dgst' utility for signature
16051 generation and verification.
16055 * Unrecognized PKCS#7 content types are now handled via a
16056 catch all ASN1_TYPE structure. This allows unsupported
16057 types to be stored as a "blob" and an application can
16058 encode and decode it manually.
16062 * Fix various signed/unsigned issues to make a_strex.c
16063 compile under VC++.
16065 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16067 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16068 length if passed a buffer. ASN1_INTEGER_to_BN failed
16069 if passed a NULL BN and its argument was negative.
16071 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16073 * Modification to PKCS#7 encoding routines to output definite
16074 length encoding. Since currently the whole structures are in
16075 memory there's not real point in using indefinite length
16076 constructed encoding. However if OpenSSL is compiled with
16077 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16081 * Added BIO_vprintf() and BIO_vsnprintf().
16085 * Added more prefixes to parse for in the strings written
16086 through a logging bio, to cover all the levels that are available
16087 through syslog. The prefixes are now:
16089 PANIC, EMERG, EMR => LOG_EMERG
16090 ALERT, ALR => LOG_ALERT
16091 CRIT, CRI => LOG_CRIT
16092 ERROR, ERR => LOG_ERR
16093 WARNING, WARN, WAR => LOG_WARNING
16094 NOTICE, NOTE, NOT => LOG_NOTICE
16095 INFO, INF => LOG_INFO
16096 DEBUG, DBG => LOG_DEBUG
16098 and as before, if none of those prefixes are present at the
16099 beginning of the string, LOG_ERR is chosen.
16101 On Win32, the `LOG_*` levels are mapped according to this:
16103 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16104 LOG_WARNING => EVENTLOG_WARNING_TYPE
16105 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16109 * Made it possible to reconfigure with just the configuration
16110 argument "reconf" or "reconfigure". The command line arguments
16111 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16112 and are retrieved from there when reconfiguring.
16118 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16120 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16124 * The obj_dat.pl script was messing up the sorting of object
16125 names. The reason was that it compared the quoted version
16126 of strings as a result "OCSP" > "OCSP Signing" because
16127 " > SPACE. Changed script to store unquoted versions of
16128 names and add quotes on output. It was also omitting some
16129 names from the lookup table if they were given a default
16130 value (that is if SN is missing it is given the same
16131 value as LN and vice versa), these are now added on the
16132 grounds that if an object has a name we should be able to
16133 look it up. Finally added warning output when duplicate
16134 short or long names are found.
16138 * Changes needed for Tandem NSK.
16140 *Scott Uroff <scott@xypro.com>*
16142 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16143 RSA_padding_check_SSLv23(), special padding was never detected
16144 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16145 version rollback attacks was not effective.
16147 In s23_clnt.c, don't use special rollback-attack detection padding
16148 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16149 client; similarly, in s23_srvr.c, don't do the rollback check if
16150 SSL 2.0 is the only protocol enabled in the server.
16154 * Make it possible to get hexdumps of unprintable data with 'openssl
16155 asn1parse'. By implication, the functions ASN1_parse_dump() and
16156 BIO_dump_indent() are added.
16160 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16161 these print out strings and name structures based on various
16162 flags including RFC2253 support and proper handling of
16163 multibyte characters. Added options to the 'x509' utility
16164 to allow the various flags to be set.
16168 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16169 Also change the functions X509_cmp_current_time() and
16170 X509_gmtime_adj() work with an ASN1_TIME structure,
16171 this will enable certificates using GeneralizedTime in validity
16172 dates to be checked.
16176 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16177 negative public key encodings) on by default,
16178 NO_NEG_PUBKEY_BUG can be set to disable it.
16182 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16183 content octets. An i2c_ASN1_OBJECT is unnecessary because
16184 the encoding can be trivially obtained from the structure.
16188 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16189 not read locks (`CRYPTO_r_[un]lock`).
16193 * A first attempt at creating official support for shared
16194 libraries through configuration. I've kept it so the
16195 default is static libraries only, and the OpenSSL programs
16196 are always statically linked for now, but there are
16197 preparations for dynamic linking in place.
16198 This has been tested on Linux and Tru64.
16202 * Randomness polling function for Win9x, as described in:
16203 Peter Gutmann, Software Generation of Practically Strong
16208 * Fix so PRNG is seeded in req if using an already existing
16213 * New options to smime application. -inform and -outform
16214 allow alternative formats for the S/MIME message including
16215 PEM and DER. The -content option allows the content to be
16216 specified separately. This should allow things like Netscape
16217 form signing output easier to verify.
16221 * Fix the ASN1 encoding of tags using the 'long form'.
16225 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16226 STRING types. These convert content octets to and from the
16227 underlying type. The actual tag and length octets are
16228 already assumed to have been read in and checked. These
16229 are needed because all other string types have virtually
16230 identical handling apart from the tag. By having versions
16231 of the ASN1 functions that just operate on content octets
16232 IMPLICIT tagging can be handled properly. It also allows
16233 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16234 and ASN1_INTEGER are identical apart from the tag.
16238 * Change the handling of OID objects as follows:
16240 - New object identifiers are inserted in objects.txt, following
16241 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16242 - objects.pl is used to process obj_mac.num and create a new
16244 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16247 This is currently kind of a hack, and the perl code in objects.pl
16248 isn't very elegant, but it works as I intended. The simplest way
16249 to check that it worked correctly is to look in obj_dat.h and
16250 check the array nid_objs and make sure the objects haven't moved
16251 around (this is important!). Additions are OK, as well as
16252 consistent name changes.
16256 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16260 * Addition of the command line parameter '-rand file' to 'openssl req'.
16261 The given file adds to whatever has already been seeded into the
16262 random pool through the RANDFILE configuration file option or
16263 environment variable, or the default random state file.
16267 * mkstack.pl now sorts each macro group into lexical order.
16268 Previously the output order depended on the order the files
16269 appeared in the directory, resulting in needless rewriting
16274 * Patches to make OpenSSL compile under Win32 again. Mostly
16275 work arounds for the VC++ problem that it treats func() as
16276 func(void). Also stripped out the parts of mkdef.pl that
16277 added extra typesafe functions: these no longer exist.
16281 * Reorganisation of the stack code. The macros are now all
16282 collected in safestack.h . Each macro is defined in terms of
16283 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16284 DEBUG_SAFESTACK is now handled in terms of function casts,
16285 this has the advantage of retaining type safety without the
16286 use of additional functions. If DEBUG_SAFESTACK is not defined
16287 then the non typesafe macros are used instead. Also modified the
16288 mkstack.pl script to handle the new form. Needs testing to see
16289 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16290 the default if no major problems. Similar behaviour for ASN1_SET_OF
16291 and PKCS12_STACK_OF.
16295 * When some versions of IIS use the 'NET' form of private key the
16296 key derivation algorithm is different. Normally MD5(password) is
16297 used as a 128 bit RC4 key. In the modified case
16298 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16299 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16300 as the old Netscape_RSA functions except they have an additional
16301 'sgckey' parameter which uses the modified algorithm. Also added
16302 an -sgckey command line option to the rsa utility. Thanks to
16303 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16304 algorithm to openssl-dev.
16308 * The evp_local.h macros were using 'c.##kname' which resulted in
16309 invalid expansion on some systems (SCO 5.0.5 for example).
16310 Corrected to 'c.kname'.
16312 *Phillip Porch <root@theporch.com>*
16314 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16315 a STACK of email addresses from a certificate or request, these look
16316 in the subject name and the subject alternative name extensions and
16317 omit any duplicate addresses.
16321 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16322 This makes DSA verification about 2 % faster.
16326 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16327 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16328 plus overhead for 1024 bit moduli).
16329 This makes exponentiations about 0.5 % faster for 1024 bit
16330 exponents (as measured by "openssl speed rsa2048").
16334 * Rename memory handling macros to avoid conflicts with other
16336 Malloc => OPENSSL_malloc
16337 Malloc_locked => OPENSSL_malloc_locked
16338 Realloc => OPENSSL_realloc
16339 Free => OPENSSL_free
16343 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16344 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16348 * CygWin32 support.
16350 *John Jarvie <jjarvie@newsguy.com>*
16352 * The type-safe stack code has been rejigged. It is now only compiled
16353 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16354 by default all type-specific stack functions are "#define"d back to
16355 standard stack functions. This results in more streamlined output
16356 but retains the type-safety checking possibilities of the original
16361 * The STACK code has been cleaned up, and certain type declarations
16362 that didn't make a lot of sense have been brought in line. This has
16363 also involved a cleanup of sorts in safestack.h to more correctly
16364 map type-safe stack functions onto their plain stack counterparts.
16365 This work has also resulted in a variety of "const"ifications of
16366 lots of the code, especially `_cmp` operations which should normally
16367 be prototyped with "const" parameters anyway.
16371 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16372 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16373 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16374 where all of 'md' is used each time the PRNG is used, but 'state'
16375 is used only indexed by a cyclic counter. As entropy may not be
16376 well distributed from the beginning, 'md' is important as a
16377 chaining variable. However, the output function chains only half
16378 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16379 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16380 in all of 'state' being rewritten, with the new values depending
16381 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16385 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16386 the handshake is continued after ssl_verify_cert_chain();
16387 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16388 can lead to 'unexplainable' connection aborts later.
16390 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16392 * Major EVP API cipher revision.
16393 Add hooks for extra EVP features. This allows various cipher
16394 parameters to be set in the EVP interface. Support added for variable
16395 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16396 setting of RC2 and RC5 parameters.
16398 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16401 Remove lots of duplicated code from the EVP library. For example *every*
16402 cipher init() function handles the 'iv' in the same way according to the
16403 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16404 for CFB and OFB modes they zero ctx->num.
16406 New functionality allows removal of S/MIME code RC2 hack.
16408 Most of the routines have the same form and so can be declared in terms
16411 By shifting this to the top level EVP_CipherInit() it can be removed from
16412 all individual ciphers. If the cipher wants to handle IVs or keys
16413 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16416 Change lots of functions like EVP_EncryptUpdate() to now return a
16417 value: although software versions of the algorithms cannot fail
16418 any installed hardware versions can.
16422 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16423 this option is set, tolerate broken clients that send the negotiated
16424 protocol version number instead of the requested protocol version
16429 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16430 i.e. non-zero for export ciphersuites, zero otherwise.
16431 Previous versions had this flag inverted, inconsistent with
16432 rsa_tmp_cb (..._TMP_RSA_CB).
16434 *Bodo Moeller; problem reported by Amit Chopra*
16436 * Add missing DSA library text string. Work around for some IIS
16437 key files with invalid SEQUENCE encoding.
16441 * Add a document (doc/standards.txt) that list all kinds of standards
16442 and so on that are implemented in OpenSSL.
16446 * Enhance c_rehash script. Old version would mishandle certificates
16447 with the same subject name hash and wouldn't handle CRLs at all.
16448 Added -fingerprint option to crl utility, to support new c_rehash
16453 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16457 * Fix for SSL server purpose checking. Server checking was
16458 rejecting certificates which had extended key usage present
16459 but no ssl client purpose.
16461 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16463 * Make PKCS#12 code work with no password. The PKCS#12 spec
16464 is a little unclear about how a blank password is handled.
16465 Since the password in encoded as a BMPString with terminating
16466 double NULL a zero length password would end up as just the
16467 double NULL. However no password at all is different and is
16468 handled differently in the PKCS#12 key generation code. NS
16469 treats a blank password as zero length. MSIE treats it as no
16470 password on export: but it will try both on import. We now do
16471 the same: PKCS12_parse() tries zero length and no password if
16472 the password is set to "" or NULL (NULL is now a valid password:
16473 it wasn't before) as does the pkcs12 application.
16477 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16478 perror when PEM_read_bio_X509_REQ fails, the error message must
16479 be obtained from the error queue.
16483 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16484 it in ERR_remove_state if appropriate, and change ERR_get_state
16485 accordingly to avoid race conditions (this is necessary because
16486 thread_hash is no longer constant once set).
16490 * Bugfix for linux-elf makefile.one.
16494 * RSA_get_default_method() will now cause a default
16495 RSA_METHOD to be chosen if one doesn't exist already.
16496 Previously this was only set during a call to RSA_new()
16497 or RSA_new_method(NULL) meaning it was possible for
16498 RSA_get_default_method() to return NULL.
16502 * Added native name translation to the existing DSO code
16503 that will convert (if the flag to do so is set) filenames
16504 that are sufficiently small and have no path information
16505 into a canonical native form. Eg. "blah" converted to
16506 "libblah.so" or "blah.dll" etc.
16510 * New function ERR_error_string_n(e, buf, len) which is like
16511 ERR_error_string(e, buf), but writes at most 'len' bytes
16512 including the 0 terminator. For ERR_error_string_n, 'buf'
16515 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16517 * CONF library reworked to become more general. A new CONF
16518 configuration file reader "class" is implemented as well as a
16519 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16520 old `CONF_*` functions are still there, but are reimplemented to
16521 work in terms of the new functions. Also, a set of functions
16522 to handle the internal storage of the configuration data is
16523 provided to make it easier to write new configuration file
16524 reader "classes" (I can definitely see something reading a
16525 configuration file in XML format, for example), called `_CONF_*`,
16526 or "the configuration storage API"...
16528 The new configuration file reading functions are:
16530 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16531 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16533 NCONF_default, NCONF_WIN32
16535 NCONF_dump_fp, NCONF_dump_bio
16537 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16538 NCONF_new creates a new CONF object. This works in the same way
16539 as other interfaces in OpenSSL, like the BIO interface.
16540 `NCONF_dump_*` dump the internal storage of the configuration file,
16541 which is useful for debugging. All other functions take the same
16542 arguments as the old `CONF_*` functions with the exception of the
16543 first that must be a `CONF *` instead of a `LHASH *`.
16545 To make it easier to use the new classes with the old `CONF_*` functions,
16546 the function CONF_set_default_method is provided.
16550 * Add '-tls1' option to 'openssl ciphers', which was already
16551 mentioned in the documentation but had not been implemented.
16552 (This option is not yet really useful because even the additional
16553 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16557 * Initial DSO code added into libcrypto for letting OpenSSL (and
16558 OpenSSL-based applications) load shared libraries and bind to
16559 them in a portable way.
16561 *Geoff Thorpe, with contributions from Richard Levitte*
16563 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16565 * Make sure _lrotl and _lrotr are only used with MSVC.
16567 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16568 (the default implementation of RAND_status).
16570 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16571 to '-clrext' (= clear extensions), as intended and documented.
16572 *Bodo Moeller; inconsistency pointed out by Michael Attili
16573 <attili@amaxo.com>*
16575 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16576 was larger than the MD block size.
16578 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16580 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16581 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16582 using the passed key: if the passed key was a private key the result
16583 of X509_print(), for example, would be to print out all the private key
16588 * des_quad_cksum() byte order bug fix.
16589 *Ulf Möller, using the problem description in krb4-0.9.7, where
16590 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16592 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16595 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16597 * For easily testing in shell scripts whether some command
16598 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16599 returns with exit code 0 iff no command of the given name is available.
16600 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16601 the output goes to stdout and nothing is printed to stderr.
16602 Additional arguments are always ignored.
16604 Since for each cipher there is a command of the same name,
16605 the 'no-cipher' compilation switches can be tested this way.
16607 ('openssl no-XXX' is not able to detect pseudo-commands such
16608 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16612 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16616 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16617 is set; it will be thrown away anyway because each handshake creates
16619 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16620 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16621 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16622 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16626 * New s_client option -ign_eof: EOF at stdin is ignored, and
16627 'Q' and 'R' lose their special meanings (quit/renegotiate).
16628 This is part of what -quiet does; unlike -quiet, -ign_eof
16629 does not suppress any output.
16633 * Add compatibility options to the purpose and trust code. The
16634 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16635 accepts a certificate or CA, this was the previous behaviour,
16636 with all the associated security issues.
16638 X509_TRUST_COMPAT is the old trust behaviour: only and
16639 automatically trust self signed roots in certificate store. A
16640 new trust setting X509_TRUST_DEFAULT is used to specify that
16641 a purpose has no associated trust setting and it should instead
16642 use the value in the default purpose.
16646 * Fix the PKCS#8 DSA private key code so it decodes keys again
16647 and fix a memory leak.
16651 * In util/mkerr.pl (which implements 'make errors'), preserve
16652 reason strings from the previous version of the .c file, as
16653 the default to have only downcase letters (and digits) in
16654 automatically generated reasons codes is not always appropriate.
16658 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16659 using strerror. Previously, ERR_reason_error_string() returned
16660 library names as reason strings for SYSerr; but SYSerr is a special
16661 case where small numbers are errno values, not library numbers.
16665 * Add '-dsaparam' option to 'openssl dhparam' application. This
16666 converts DSA parameters into DH parameters. (When creating parameters,
16667 DSA_generate_parameters is used.)
16671 * Include 'length' (recommended exponent length) in C code generated
16672 by 'openssl dhparam -C'.
16676 * The second argument to set_label in perlasm was already being used
16677 so couldn't be used as a "file scope" flag. Moved to third argument
16682 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16683 instead of RAND_bytes for encryption IVs and salts.
16687 * Include RAND_status() into RAND_METHOD instead of implementing
16688 it only for md_rand.c Otherwise replacing the PRNG by calling
16689 RAND_set_rand_method would be impossible.
16693 * Don't let DSA_generate_key() enter an infinite loop if the random
16694 number generation fails.
16698 * New 'rand' application for creating pseudo-random output.
16702 * Added configuration support for Linux/IA64
16704 *Rolf Haberrecker <rolf@suse.de>*
16706 * Assembler module support for Mingw32.
16710 * Shared library support for HPUX (in shlib/).
16712 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16714 * Shared library support for Solaris gcc.
16716 *Lutz Behnke <behnke@trustcenter.de>*
16718 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16720 * PKCS7_encrypt() was adding text MIME headers twice because they
16721 were added manually and by SMIME_crlf_copy().
16725 * In bntest.c don't call BN_rand with zero bits argument.
16727 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16729 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16730 case was implemented. This caused BN_div_recp() to fail occasionally.
16734 * Add an optional second argument to the set_label() in the perl
16735 assembly language builder. If this argument exists and is set
16736 to 1 it signals that the assembler should use a symbol whose
16737 scope is the entire file, not just the current function. This
16738 is needed with MASM which uses the format label:: for this scope.
16740 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16742 * Change the ASN1 types so they are typedefs by default. Before
16743 almost all types were #define'd to ASN1_STRING which was causing
16744 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16749 * Change names of new functions to the new get1/get0 naming
16750 convention: After 'get1', the caller owns a reference count
16751 and has to call `..._free`; 'get0' returns a pointer to some
16752 data structure without incrementing reference counters.
16753 (Some of the existing 'get' functions increment a reference
16754 counter, some don't.)
16755 Similarly, 'set1' and 'add1' functions increase reference
16756 counters or duplicate objects.
16760 * Allow for the possibility of temp RSA key generation failure:
16761 the code used to assume it always worked and crashed on failure.
16765 * Fix potential buffer overrun problem in BIO_printf().
16766 *Ulf Möller, using public domain code by Patrick Powell; problem
16767 pointed out by David Sacerdote <das33@cornell.edu>*
16769 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16770 RAND_egd() and RAND_status(). In the command line application,
16771 the EGD socket can be specified like a seed file using RANDFILE
16776 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16777 Some CAs (e.g. Verisign) distribute certificates in this form.
16781 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16782 list to exclude them. This means that no special compilation option
16783 is needed to use anonymous DH: it just needs to be included in the
16788 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16789 EVP_MD_type. The old functionality is available in a new macro called
16790 EVP_MD_md(). Change code that uses it and update docs.
16794 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16795 where the `void *` argument is replaced by a function pointer argument.
16796 Previously `void *` was abused to point to functions, which works on
16797 many platforms, but is not correct. As these functions are usually
16798 called by macros defined in OpenSSL header files, most source code
16799 should work without changes.
16803 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16804 sections with information on -D... compiler switches used for
16805 compiling the library so that applications can see them. To enable
16806 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16807 must be defined. E.g.,
16808 #define OPENSSL_ALGORITHM_DEFINES
16809 #include <openssl/opensslconf.h>
16810 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16812 *Richard Levitte, Ulf and Bodo Möller*
16814 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16819 * Change the 'other' type in certificate aux info to a STACK_OF
16820 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16821 the required ASN1 format: arbitrary types determined by an OID.
16825 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16826 argument to 'req'. This is not because the function is newer or
16827 better than others it just uses the work 'NEW' in the certificate
16828 request header lines. Some software needs this.
16832 * Reorganise password command line arguments: now passwords can be
16833 obtained from various sources. Delete the PEM_cb function and make
16834 it the default behaviour: i.e. if the callback is NULL and the
16835 usrdata argument is not NULL interpret it as a null terminated pass
16836 phrase. If usrdata and the callback are NULL then the pass phrase
16837 is prompted for as usual.
16841 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16842 the support is automatically enabled. The resulting binaries will
16843 autodetect the card and use it if present.
16845 *Ben Laurie and Compaq Inc.*
16847 * Work around for Netscape hang bug. This sends certificate request
16848 and server done in one record. Since this is perfectly legal in the
16849 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16850 the bugs/SSLv3 entry for more info.
16854 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16858 * Add -rand argument to smime and pkcs12 applications and read/write
16863 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16867 * Add command line password options to the remaining applications.
16871 * Bug fix for BN_div_recp() for numerators with an even number of
16876 * More tests in bntest.c, and changed test_bn output.
16880 * ./config recognizes MacOS X now.
16884 * Bug fix for BN_div() when the first words of num and divisor are
16885 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16889 * Add support for various broken PKCS#8 formats, and command line
16890 options to produce them.
16894 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16895 get temporary BIGNUMs from a BN_CTX.
16899 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16904 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16905 include a #define from the old name to the new. The original intent
16906 was that statically linked binaries could for example just call
16907 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16908 link with digests. This never worked because SSLeay_add_all_digests()
16909 and SSLeay_add_all_ciphers() were in the same source file so calling
16910 one would link with the other. They are now in separate source files.
16914 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16918 * Use a less unusual form of the Miller-Rabin primality test (it used
16919 a binary algorithm for exponentiation integrated into the Miller-Rabin
16920 loop, our standard modexp algorithms are faster).
16924 * Support for the EBCDIC character set completed.
16926 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16928 * Source code cleanups: use const where appropriate, eliminate casts,
16929 use `void *` instead of `char *` in lhash.
16933 * Bugfix: ssl3_send_server_key_exchange was not restartable
16934 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16935 this the server could overwrite ephemeral keys that the client
16940 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16941 using 50 iterations of the Rabin-Miller test.
16943 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16944 iterations of the Rabin-Miller test as required by the appendix
16945 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16946 As BN_is_prime_fasttest includes trial division, DSA parameter
16947 generation becomes much faster.
16949 This implies a change for the callback functions in DSA_is_prime
16950 and DSA_generate_parameters: The callback function is called once
16951 for each positive witness in the Rabin-Miller test, not just
16952 occasionally in the inner loop; and the parameters to the
16953 callback function now provide an iteration count for the outer
16954 loop rather than for the current invocation of the inner loop.
16955 DSA_generate_parameters additionally can call the callback
16956 function with an 'iteration count' of -1, meaning that a
16957 candidate has passed the trial division test (when q is generated
16958 from an application-provided seed, trial division is skipped).
16962 * New function BN_is_prime_fasttest that optionally does trial
16963 division before starting the Rabin-Miller test and has
16964 an additional BN_CTX * argument (whereas BN_is_prime always
16965 has to allocate at least one BN_CTX).
16966 'callback(1, -1, cb_arg)' is called when a number has passed the
16967 trial division stage.
16971 * Fix for bug in CRL encoding. The validity dates weren't being handled
16976 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16980 * New function BN_pseudo_rand().
16984 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16985 bignum version of BN_from_montgomery() with the working code from
16986 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16991 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16992 made it impossible to use the same SSL_SESSION data structure in
16993 SSL2 clients in multiple threads.
16997 * The return value of RAND_load_file() no longer counts bytes obtained
16998 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16999 to seed the PRNG (previously an explicit byte count was required).
17001 *Ulf Möller, Bodo Möller*
17003 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
17004 used `char *` instead of `void *` and had casts all over the place.
17008 * Make BN_generate_prime() return NULL on error if ret!=NULL.
17012 * Retain source code compatibility for BN_prime_checks macro:
17013 BN_is_prime(..., BN_prime_checks, ...) now uses
17014 BN_prime_checks_for_size to determine the appropriate number of
17015 Rabin-Miller iterations.
17019 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
17020 DH_CHECK_P_NOT_SAFE_PRIME.
17021 (Check if this is true? OpenPGP calls them "strong".)
17025 * Merge the functionality of "dh" and "gendh" programs into a new program
17026 "dhparam". The old programs are retained for now but will handle DH keys
17027 (instead of parameters) in future.
17031 * Make the ciphers, s_server and s_client programs check the return values
17032 when a new cipher list is set.
17036 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17037 ciphers. Before when the 56bit ciphers were enabled the sorting was
17040 The syntax for the cipher sorting has been extended to support sorting by
17041 cipher-strength (using the strength_bits hard coded in the tables).
17042 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17044 Fix a bug in the cipher-command parser: when supplying a cipher command
17045 string with an "undefined" symbol (neither command nor alphanumeric
17046 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17047 an error is flagged.
17049 Due to the strength-sorting extension, the code of the
17050 ssl_create_cipher_list() function was completely rearranged. I hope that
17051 the readability was also increased :-)
17053 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17055 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17056 for the first serial number and places 2 in the serial number file. This
17057 avoids problems when the root CA is created with serial number zero and
17058 the first user certificate has the same issuer name and serial number
17063 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17064 the new code. Add documentation for this stuff.
17068 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17069 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17070 structures and behave in an analogous way to the X509v3 functions:
17071 they shouldn't be called directly but wrapper functions should be used
17074 So we also now have some wrapper functions that call the X509at functions
17075 when passed certificate requests. (TO DO: similar things can be done with
17076 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17077 things. Some of these need some d2i or i2d and print functionality
17078 because they handle more complex structures.)
17082 * Add missing #ifndefs that caused missing symbols when building libssl
17083 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17084 NO_RSA in `ssl/s2*.c`.
17086 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17088 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17089 has a return value which indicates the quality of the random data
17090 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17091 error queue. New function RAND_pseudo_bytes() generates output that is
17092 guaranteed to be unique but not unpredictable. RAND_add is like
17093 RAND_seed, but takes an extra argument for an entropy estimate
17094 (RAND_seed always assumes full entropy).
17098 * Do more iterations of Rabin-Miller probable prime test (specifically,
17099 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17100 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17101 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17102 false-positive rate of at most 2^-80 for random input.
17106 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17110 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17111 in the 0.9.5 release), this returns the chain
17112 from an X509_CTX structure with a dup of the stack and all
17113 the X509 reference counts upped: so the stack will exist
17114 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17117 Also make SSL_SESSION_print() print out the verify return
17122 * Add manpage for the pkcs12 command. Also change the default
17123 behaviour so MAC iteration counts are used unless the new
17124 -nomaciter option is used. This improves file security and
17125 only older versions of MSIE (4.0 for example) need it.
17129 * Honor the no-xxx Configure options when creating .DEF files.
17133 * Add PKCS#10 attributes to field table: challengePassword,
17134 unstructuredName and unstructuredAddress. These are taken from
17135 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17136 international characters are used.
17138 More changes to X509_ATTRIBUTE code: allow the setting of types
17139 based on strings. Remove the 'loc' parameter when adding
17140 attributes because these will be a SET OF encoding which is sorted
17145 * Initial changes to the 'req' utility to allow request generation
17146 automation. This will allow an application to just generate a template
17147 file containing all the field values and have req construct the
17150 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17151 used all over the place including certificate requests and PKCS#7
17152 structures. They are currently handled manually where necessary with
17153 some primitive wrappers for PKCS#7. The new functions behave in a
17154 manner analogous to the X509 extension functions: they allow
17155 attributes to be looked up by NID and added.
17157 Later something similar to the X509V3 code would be desirable to
17158 automatically handle the encoding, decoding and printing of the
17159 more complex types. The string types like challengePassword can
17160 be handled by the string table functions.
17162 Also modified the multi byte string table handling. Now there is
17163 a 'global mask' which masks out certain types. The table itself
17164 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17165 is useful when for example there is only one permissible type
17166 (as in countryName) and using the mask might result in no valid
17171 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17172 SSL_get_peer_finished to allow applications to obtain the latest
17173 Finished messages sent to the peer or expected from the peer,
17174 respectively. (SSL_get_peer_finished is usually the Finished message
17175 actually received from the peer, otherwise the protocol will be aborted.)
17177 As the Finished message are message digests of the complete handshake
17178 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17179 be used for external authentication procedures when the authentication
17180 provided by SSL/TLS is not desired or is not enough.
17184 * Enhanced support for Alpha Linux is added. Now ./config checks if
17185 the host supports BWX extension and if Compaq C is present on the
17186 $PATH. Just exploiting of the BWX extension results in 20-30%
17187 performance kick for some algorithms, e.g. DES and RC4 to mention
17188 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17193 * Add support for MS "fast SGC". This is arguably a violation of the
17194 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17195 weak crypto and after checking the certificate is SGC a second one
17196 with strong crypto. MS SGC stops the first handshake after receiving
17197 the server certificate message and sends a second client hello. Since
17198 a server will typically do all the time consuming operations before
17199 expecting any further messages from the client (server key exchange
17200 is the most expensive) there is little difference between the two.
17202 To get OpenSSL to support MS SGC we have to permit a second client
17203 hello message after we have sent server done. In addition we have to
17204 reset the MAC if we do get this second client hello.
17208 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17209 if a DER encoded private key is RSA or DSA traditional format. Changed
17210 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17211 format DER encoded private key. Newer code should use PKCS#8 format which
17212 has the key type encoded in the ASN1 structure. Added DER private key
17213 support to pkcs8 application.
17217 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17218 ciphersuites has been selected (as required by the SSL 3/TLS 1
17219 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17220 is set, we interpret this as a request to violate the specification
17221 (the worst that can happen is a handshake failure, and 'correct'
17222 behaviour would result in a handshake failure anyway).
17226 * In SSL_CTX_add_session, take into account that there might be multiple
17227 SSL_SESSION structures with the same session ID (e.g. when two threads
17228 concurrently obtain them from an external cache).
17229 The internal cache can handle only one SSL_SESSION with a given ID,
17230 so if there's a conflict, we now throw out the old one to achieve
17235 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17236 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17237 some routines that use cipher OIDs: some ciphers do not have OIDs
17238 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17243 * Simplify the trust setting structure and code. Now we just have
17244 two sequences of OIDs for trusted and rejected settings. These will
17245 typically have values the same as the extended key usage extension
17246 and any application specific purposes.
17248 The trust checking code now has a default behaviour: it will just
17249 check for an object with the same NID as the passed id. Functions can
17250 be provided to override either the default behaviour or the behaviour
17251 for a given id. SSL client, server and email already have functions
17252 in place for compatibility: they check the NID and also return "trusted"
17253 if the certificate is self signed.
17257 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17258 traditional format into an EVP_PKEY structure.
17262 * Add a password callback function PEM_cb() which either prompts for
17263 a password if usr_data is NULL or otherwise assumes it is a null
17264 terminated password. Allow passwords to be passed on command line
17265 environment or config files in a few more utilities.
17269 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17270 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17271 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17272 Update documentation.
17276 * Support for ASN1 "NULL" type. This could be handled before by using
17277 ASN1_TYPE but there wasn't any function that would try to read a NULL
17278 and produce an error if it couldn't. For compatibility we also have
17279 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17280 don't allocate anything because they don't need to.
17284 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17287 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17289 * Rebuild of the memory allocation routines used by OpenSSL code and
17290 possibly others as well. The purpose is to make an interface that
17291 provide hooks so anyone can build a separate set of allocation and
17292 deallocation routines to be used by OpenSSL, for example memory
17293 pool implementations, or something else, which was previously hard
17294 since Malloc(), Realloc() and Free() were defined as macros having
17295 the values malloc, realloc and free, respectively (except for Win32
17296 compilations). The same is provided for memory debugging code.
17297 OpenSSL already comes with functionality to find memory leaks, but
17298 this gives people a chance to debug other memory problems.
17300 With these changes, a new set of functions and macros have appeared:
17302 CRYPTO_set_mem_debug_functions() [F]
17303 CRYPTO_get_mem_debug_functions() [F]
17304 CRYPTO_dbg_set_options() [F]
17305 CRYPTO_dbg_get_options() [F]
17306 CRYPTO_malloc_debug_init() [M]
17308 The memory debug functions are NULL by default, unless the library
17309 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17310 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17311 gives the standard debugging functions that come with OpenSSL) or
17312 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17313 provided by the library user) must be used. When the standard
17314 debugging functions are used, CRYPTO_dbg_set_options can be used to
17315 request additional information:
17316 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17317 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17319 Also, things like CRYPTO_set_mem_functions will always give the
17320 expected result (the new set of functions is used for allocation
17321 and deallocation) at all times, regardless of platform and compiler
17324 To finish it up, some functions that were never use in any other
17325 way than through macros have a new API and new semantic:
17327 CRYPTO_dbg_malloc()
17328 CRYPTO_dbg_realloc()
17331 All macros of value have retained their old syntax.
17333 *Richard Levitte and Bodo Moeller*
17335 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17336 ordering of SMIMECapabilities wasn't in "strength order" and there
17337 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17342 * Some ASN1 types with illegal zero length encoding (INTEGER,
17343 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17345 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17347 * Merge in my S/MIME library for OpenSSL. This provides a simple
17348 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17349 functionality to handle multipart/signed properly) and a utility
17350 called 'smime' to call all this stuff. This is based on code I
17351 originally wrote for Celo who have kindly allowed it to be
17352 included in OpenSSL.
17356 * Add variants des_set_key_checked and des_set_key_unchecked of
17357 des_set_key (aka des_key_sched). Global variable des_check_key
17358 decides which of these is called by des_set_key; this way
17359 des_check_key behaves as it always did, but applications and
17360 the library itself, which was buggy for des_check_key == 1,
17361 have a cleaner way to pick the version they need.
17365 * New function PKCS12_newpass() which changes the password of a
17370 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17371 dynamic mix. In both cases the ids can be used as an index into the
17372 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17373 functions so they accept a list of the field values and the
17374 application doesn't need to directly manipulate the X509_TRUST
17379 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17384 * Modify the way the V3 extension code looks up extensions. This now
17385 works in a similar way to the object code: we have some "standard"
17386 extensions in a static table which is searched with OBJ_bsearch()
17387 and the application can add dynamic ones if needed. The file
17388 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17389 updated whenever a new extension is added to the core code and kept
17390 in ext_nid order. There is a simple program 'tabtest.c' which checks
17391 this. New extensions are not added too often so this file can readily
17392 be maintained manually.
17394 There are two big advantages in doing things this way. The extensions
17395 can be looked up immediately and no longer need to be "added" using
17396 X509V3_add_standard_extensions(): this function now does nothing.
17397 Side note: I get *lots* of email saying the extension code doesn't
17398 work because people forget to call this function.
17399 Also no dynamic allocation is done unless new extensions are added:
17400 so if we don't add custom extensions there is no need to call
17401 X509V3_EXT_cleanup().
17405 * Modify enc utility's salting as follows: make salting the default. Add a
17406 magic header, so unsalted files fail gracefully instead of just decrypting
17407 to garbage. This is because not salting is a big security hole, so people
17408 should be discouraged from doing it.
17412 * Fixes and enhancements to the 'x509' utility. It allowed a message
17413 digest to be passed on the command line but it only used this
17414 parameter when signing a certificate. Modified so all relevant
17415 operations are affected by the digest parameter including the
17416 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17417 DSA key was used because it didn't fix the digest.
17421 * Initial certificate chain verify code. Currently tests the untrusted
17422 certificates for consistency with the verify purpose (which is set
17423 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17425 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17426 this is because it will reject chains with invalid extensions whereas
17427 every previous version of OpenSSL and SSLeay made no checks at all.
17429 Trust code: checks the root CA for the relevant trust settings. Trust
17430 settings have an initial value consistent with the verify purpose: e.g.
17431 if the verify purpose is for SSL client use it expects the CA to be
17432 trusted for SSL client use. However the default value can be changed to
17433 permit custom trust settings: one example of this would be to only trust
17434 certificates from a specific "secure" set of CAs.
17436 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17437 which should be used for version portability: especially since the
17438 verify structure is likely to change more often now.
17440 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17441 to set them. If not set then assume SSL clients will verify SSL servers
17444 Two new options to the verify program: -untrusted allows a set of
17445 untrusted certificates to be passed in and -purpose which sets the
17446 intended purpose of the certificate. If a purpose is set then the
17447 new chain verify code is used to check extension consistency.
17451 * Support for the authority information access extension.
17455 * Modify RSA and DSA PEM read routines to transparently handle
17456 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17457 public keys in a format compatible with certificate
17458 SubjectPublicKeyInfo structures. Unfortunately there were already
17459 functions called *_PublicKey_* which used various odd formats so
17460 these are retained for compatibility: however the DSA variants were
17461 never in a public release so they have been deleted. Changed dsa/rsa
17462 utilities to handle the new format: note no releases ever handled public
17463 keys so we should be OK.
17465 The primary motivation for this change is to avoid the same fiasco
17466 that dogs private keys: there are several incompatible private key
17467 formats some of which are standard and some OpenSSL specific and
17468 require various evil hacks to allow partial transparent handling and
17469 even then it doesn't work with DER formats. Given the option anything
17470 other than PKCS#8 should be dumped: but the other formats have to
17471 stay in the name of compatibility.
17473 With public keys and the benefit of hindsight one standard format
17474 is used which works with EVP_PKEY, RSA or DSA structures: though
17475 it clearly returns an error if you try to read the wrong kind of key.
17477 Added a -pubkey option to the 'x509' utility to output the public key.
17478 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17479 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17480 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17481 that do the same as the `EVP_PKEY_assign_*()` except they up the
17482 reference count of the added key (they don't "swallow" the
17487 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17488 CRLs would fail if the file contained no certificates or no CRLs:
17489 added a new function to read in both types and return the number
17490 read: this means that if none are read it will be an error. The
17491 DER versions of the certificate and CRL reader would always fail
17492 because it isn't possible to mix certificates and CRLs in DER format
17493 without choking one or the other routine. Changed this to just read
17494 a certificate: this is the best we can do. Also modified the code
17495 in `apps/verify.c` to take notice of return codes: it was previously
17496 attempting to read in certificates from NULL pointers and ignoring
17497 any errors: this is one reason why the cert and CRL reader seemed
17498 to work. It doesn't check return codes from the default certificate
17499 routines: these may well fail if the certificates aren't installed.
17503 * Code to support otherName option in GeneralName.
17507 * First update to verify code. Change the verify utility
17508 so it warns if it is passed a self signed certificate:
17509 for consistency with the normal behaviour. X509_verify
17510 has been modified to it will now verify a self signed
17511 certificate if *exactly* the same certificate appears
17512 in the store: it was previously impossible to trust a
17513 single self signed certificate. This means that:
17514 openssl verify ss.pem
17515 now gives a warning about a self signed certificate but
17516 openssl verify -CAfile ss.pem ss.pem
17521 * For servers, store verify_result in SSL_SESSION data structure
17522 (and add it to external session representation).
17523 This is needed when client certificate verifications fails,
17524 but an application-provided verification callback (set by
17525 SSL_CTX_set_cert_verify_callback) allows accepting the session
17526 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17527 but returns 1): When the session is reused, we have to set
17528 ssl->verify_result to the appropriate error code to avoid
17531 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17533 * Fix a bug in the new PKCS#7 code: it didn't consider the
17534 case in PKCS7_dataInit() where the signed PKCS7 structure
17535 didn't contain any existing data because it was being created.
17537 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17539 * Add a salt to the key derivation routines in enc.c. This
17540 forms the first 8 bytes of the encrypted file. Also add a
17541 -S option to allow a salt to be input on the command line.
17545 * New function X509_cmp(). Oddly enough there wasn't a function
17546 to compare two certificates. We do this by working out the SHA1
17547 hash and comparing that. X509_cmp() will be needed by the trust
17552 * SSL_get1_session() is like SSL_get_session(), but increments
17553 the reference count in the SSL_SESSION returned.
17555 *Geoff Thorpe <geoff@eu.c2.net>*
17557 * Fix for 'req': it was adding a null to request attributes.
17558 Also change the X509_LOOKUP and X509_INFO code to handle
17559 certificate auxiliary information.
17563 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17568 * Add the possibility to add extra information to the memory leak
17569 detecting output, to form tracebacks, showing from where each
17570 allocation was originated: CRYPTO_push_info("constant string") adds
17571 the string plus current file name and line number to a per-thread
17572 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17573 is like calling CYRPTO_pop_info() until the stack is empty.
17574 Also updated memory leak detection code to be multi-thread-safe.
17578 * Add options -text and -noout to pkcs7 utility and delete the
17579 encryption options which never did anything. Update docs.
17583 * Add options to some of the utilities to allow the pass phrase
17584 to be included on either the command line (not recommended on
17585 OSes like Unix) or read from the environment. Update the
17586 manpages and fix a few bugs.
17590 * Add a few manpages for some of the openssl commands.
17594 * Fix the -revoke option in ca. It was freeing up memory twice,
17595 leaking and not finding already revoked certificates.
17599 * Extensive changes to support certificate auxiliary information.
17600 This involves the use of X509_CERT_AUX structure and X509_AUX
17601 functions. An X509_AUX function such as PEM_read_X509_AUX()
17602 can still read in a certificate file in the usual way but it
17603 will also read in any additional "auxiliary information". By
17604 doing things this way a fair degree of compatibility can be
17605 retained: existing certificates can have this information added
17606 using the new 'x509' options.
17608 Current auxiliary information includes an "alias" and some trust
17609 settings. The trust settings will ultimately be used in enhanced
17610 certificate chain verification routines: currently a certificate
17611 can only be trusted if it is self signed and then it is trusted
17616 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17617 The problem was that one of the replacement routines had not been working
17618 since SSLeay releases. For now the offending routine has been replaced
17619 with non-optimised assembler. Even so, this now gives around 95%
17620 performance improvement for 1024 bit RSA signs.
17624 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17625 handling. Most clients have the effective key size in bits equal to
17626 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17627 A few however don't do this and instead use the size of the decrypted key
17628 to determine the RC2 key length and the AlgorithmIdentifier to determine
17629 the effective key length. In this case the effective key length can still
17630 be 40 bits but the key length can be 168 bits for example. This is fixed
17631 by manually forcing an RC2 key into the EVP_PKEY structure because the
17632 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17633 the key length and effective key length are equal.
17637 * Add a bunch of functions that should simplify the creation of
17638 X509_NAME structures. Now you should be able to do:
17639 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17640 and have it automatically work out the correct field type and fill in
17641 the structures. The more adventurous can try:
17642 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17643 and it will (hopefully) work out the correct multibyte encoding.
17647 * Change the 'req' utility to use the new field handling and multibyte
17648 copy routines. Before the DN field creation was handled in an ad hoc
17649 way in req, ca, and x509 which was rather broken and didn't support
17650 BMPStrings or UTF8Strings. Since some software doesn't implement
17651 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17652 using the dirstring_type option. See the new comment in the default
17653 openssl.cnf for more info.
17657 * Make crypto/rand/md_rand.c more robust:
17658 - Assure unique random numbers after fork().
17659 - Make sure that concurrent threads access the global counter and
17660 md serializably so that we never lose entropy in them
17661 or use exactly the same state in multiple threads.
17662 Access to the large state is not always serializable because
17663 the additional locking could be a performance killer, and
17664 md should be large enough anyway.
17668 * New file `apps/app_rand.c` with commonly needed functionality
17669 for handling the random seed file.
17671 Use the random seed file in some applications that previously did not:
17673 dsaparam -genkey (which also ignored its '-rand' option),
17676 x509 (when signing).
17677 Except on systems with /dev/urandom, it is crucial to have a random
17678 seed file at least for key creation, DSA signing, and for DH exchanges;
17679 for RSA signatures we could do without one.
17681 gendh and gendsa (unlike genrsa) used to read only the first byte
17682 of each file listed in the '-rand' option. The function as previously
17683 found in genrsa is now in app_rand.c and is used by all programs
17684 that support '-rand'.
17688 * In RAND_write_file, use mode 0600 for creating files;
17689 don't just chmod when it may be too late.
17693 * Report an error from X509_STORE_load_locations
17694 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17698 * New function ASN1_mbstring_copy() this copies a string in either
17699 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17700 into an ASN1_STRING type. A mask of permissible types is passed
17701 and it chooses the "minimal" type to use or an error if not type
17706 * Add function equivalents to the various macros in asn1.h. The old
17707 macros are retained with an `M_` prefix. Code inside the library can
17708 use the `M_` macros. External code (including the openssl utility)
17709 should *NOT* in order to be "shared library friendly".
17713 * Add various functions that can check a certificate's extensions
17714 to see if it usable for various purposes such as SSL client,
17715 server or S/MIME and CAs of these types. This is currently
17716 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17717 verification. Also added a -purpose flag to x509 utility to
17718 print out all the purposes.
17722 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17727 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17728 for, obtain and decode and extension and obtain its critical flag.
17729 This allows all the necessary extension code to be handled in a
17730 single function call.
17734 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17735 platforms. See crypto/rc4/rc4_enc.c for further details.
17739 * New -noout option to asn1parse. This causes no output to be produced
17740 its main use is when combined with -strparse and -out to extract data
17741 from a file (which may not be in ASN.1 format).
17745 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17746 when producing the local key id.
17748 *Richard Levitte <levitte@stacken.kth.se>*
17750 * New option -dhparam in s_server. This allows a DH parameter file to be
17751 stated explicitly. If it is not stated then it tries the first server
17752 certificate file. The previous behaviour hard coded the filename
17757 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17758 a public key to be input or output. For example:
17759 openssl rsa -in key.pem -pubout -out pubkey.pem
17760 Also added necessary DSA public key functions to handle this.
17764 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17765 in the message. This was handled by allowing
17766 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17768 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17770 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17771 to the end of the strings whereas this didn't. This would cause problems
17772 if strings read with d2i_ASN1_bytes() were later modified.
17774 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17776 * Fix for base64 decode bug. When a base64 bio reads only one line of
17777 data and it contains EOF it will end up returning an error. This is
17778 caused by input 46 bytes long. The cause is due to the way base64
17779 BIOs find the start of base64 encoded data. They do this by trying a
17780 trial decode on each line until they find one that works. When they
17781 do a flag is set and it starts again knowing it can pass all the
17782 data directly through the decoder. Unfortunately it doesn't reset
17783 the context it uses. This means that if EOF is reached an attempt
17784 is made to pass two EOFs through the context and this causes the
17785 resulting error. This can also cause other problems as well. As is
17786 usual with these problems it takes *ages* to find and the fix is
17787 trivial: move one line.
17789 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17791 * Ugly workaround to get s_client and s_server working under Windows. The
17792 old code wouldn't work because it needed to select() on sockets and the
17793 tty (for keypresses and to see if data could be written). Win32 only
17794 supports select() on sockets so we select() with a 1s timeout on the
17795 sockets and then see if any characters are waiting to be read, if none
17796 are present then we retry, we also assume we can always write data to
17797 the tty. This isn't nice because the code then blocks until we've
17798 received a complete line of data and it is effectively polling the
17799 keyboard at 1s intervals: however it's quite a bit better than not
17800 working at all :-) A dedicated Windows application might handle this
17801 with an event loop for example.
17805 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17806 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17807 will be called when RSA_sign() and RSA_verify() are used. This is useful
17808 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17809 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17810 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17811 This necessitated the support of an extra signature type NID_md5_sha1
17812 for SSL signatures and modifications to the SSL library to use it instead
17813 of calling RSA_public_decrypt() and RSA_private_encrypt().
17817 * Add new -verify -CAfile and -CApath options to the crl program, these
17818 will lookup a CRL issuers certificate and verify the signature in a
17819 similar way to the verify program. Tidy up the crl program so it
17820 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17821 less strict. It will now permit CRL extensions even if it is not
17822 a V2 CRL: this will allow it to tolerate some broken CRLs.
17826 * Initialize all non-automatic variables each time one of the openssl
17827 sub-programs is started (this is necessary as they may be started
17828 multiple times from the "OpenSSL>" prompt).
17830 *Lennart Bang, Bodo Moeller*
17832 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17833 removing all other RSA functionality (this is what NO_RSA does). This
17834 is so (for example) those in the US can disable those operations covered
17835 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17840 * Non-copying interface to BIO pairs.
17841 (still largely untested)
17845 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17846 ASCII string. This was handled independently in various places before.
17850 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17851 UTF8 strings a character at a time.
17855 * Use client_version from client hello to select the protocol
17856 (s23_srvr.c) and for RSA client key exchange verification
17857 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17861 * Add various utility functions to handle SPKACs, these were previously
17862 handled by poking round in the structure internals. Added new function
17863 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17864 print, verify and generate SPKACs. Based on an original idea from
17865 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17869 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17873 * Allow the config file extension section to be overwritten on the
17874 command line. Based on an original idea from Massimiliano Pala
17875 <madwolf@comune.modena.it>. The new option is called -extensions
17876 and can be applied to ca, req and x509. Also -reqexts to override
17877 the request extensions in req and -crlexts to override the crl extensions
17882 * Add new feature to the SPKAC handling in ca. Now you can include
17883 the same field multiple times by preceding it by "XXXX." for example:
17886 this is the same syntax as used in the req config file.
17890 * Allow certificate extensions to be added to certificate requests. These
17891 are specified in a 'req_extensions' option of the req section of the
17892 config file. They can be printed out with the -text option to req but
17893 are otherwise ignored at present.
17897 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17898 data read consists of only the final block it would not decrypted because
17899 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17900 A misplaced 'break' also meant the decrypted final block might not be
17901 copied until the next read.
17905 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17906 a few extra parameters to the DH structure: these will be useful if
17907 for example we want the value of 'q' or implement X9.42 DH.
17911 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17912 provides hooks that allow the default DSA functions or functions on a
17913 "per key" basis to be replaced. This allows hardware acceleration and
17914 hardware key storage to be handled without major modification to the
17915 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17916 associated functions.
17920 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17921 as "read only": it can't be written to and the buffer it points to will
17922 not be freed. Reading from a read only BIO is much more efficient than
17923 a normal memory BIO. This was added because there are several times when
17924 an area of memory needs to be read from a BIO. The previous method was
17925 to create a memory BIO and write the data to it, this results in two
17926 copies of the data and an O(n^2) reading algorithm. There is a new
17927 function BIO_new_mem_buf() which creates a read only memory BIO from
17928 an area of memory. Also modified the PKCS#7 routines to use read only
17933 * Bugfix: ssl23_get_client_hello did not work properly when called in
17934 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17935 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17936 but a retry condition occurred while trying to read the rest.
17940 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17941 NID_pkcs7_encrypted by default: this was wrong since this should almost
17942 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17943 the encrypted data type: this is a more sensible place to put it and it
17944 allows the PKCS#12 code to be tidied up that duplicated this
17949 * Changed obj_dat.pl script so it takes its input and output files on
17950 the command line. This should avoid shell escape redirection problems
17955 * Initial support for certificate extension requests, these are included
17956 in things like Xenroll certificate requests. Included functions to allow
17957 extensions to be obtained and added.
17961 * -crlf option to s_client and s_server for sending newlines as
17962 CRLF (as required by many protocols).
17966 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17968 * Install libRSAglue.a when OpenSSL is built with RSAref.
17970 *Ralf S. Engelschall*
17972 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17974 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17976 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17981 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17982 DH parameters/keys (q is lost during that conversion, but the resulting
17983 DH parameters contain its length).
17985 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17986 much faster than DH_generate_parameters (which creates parameters
17987 where `p = 2*q + 1`), and also the smaller q makes DH computations
17988 much more efficient (160-bit exponentiation instead of 1024-bit
17989 exponentiation); so this provides a convenient way to support DHE
17990 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17991 utter importance to use
17992 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17994 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17995 when such DH parameters are used, because otherwise small subgroup
17996 attacks may become possible!
18000 * Avoid memory leak in i2d_DHparams.
18004 * Allow the -k option to be used more than once in the enc program:
18005 this allows the same encrypted message to be read by multiple recipients.
18009 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
18010 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
18011 it will always use the numerical form of the OID, even if it has a short
18016 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
18017 method only got called if p,q,dmp1,dmq1,iqmp components were present,
18018 otherwise bn_mod_exp was called. In the case of hardware keys for example
18019 no private key components need be present and it might store extra data
18020 in the RSA structure, which cannot be accessed from bn_mod_exp.
18021 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
18022 private key operations.
18026 * Added support for SPARC Linux.
18030 * pem_password_cb function type incompatibly changed from
18031 typedef int pem_password_cb(char *buf, int size, int rwflag);
18033 ....(char *buf, int size, int rwflag, void *userdata);
18034 so that applications can pass data to their callbacks:
18035 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18036 additional void * argument, which is just handed through whenever
18037 the password callback is called.
18039 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18041 New function SSL_CTX_set_default_passwd_cb_userdata.
18043 Compatibility note: As many C implementations push function arguments
18044 onto the stack in reverse order, the new library version is likely to
18045 interoperate with programs that have been compiled with the old
18046 pem_password_cb definition (PEM_whatever takes some data that
18047 happens to be on the stack as its last argument, and the callback
18048 just ignores this garbage); but there is no guarantee whatsoever that
18051 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18052 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18053 problems not only on Windows, but also on some Unix platforms.
18054 To avoid problematic command lines, these definitions are now in an
18055 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18056 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18060 * MIPS III/IV assembler module is reimplemented.
18064 * More DES library cleanups: remove references to srand/rand and
18065 delete an unused file.
18069 * Add support for the free Netwide assembler (NASM) under Win32,
18070 since not many people have MASM (ml) and it can be hard to obtain.
18071 This is currently experimental but it seems to work OK and pass all
18072 the tests. Check out INSTALL.W32 for info.
18076 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18077 without temporary keys kept an extra copy of the server key,
18078 and connections with temporary keys did not free everything in case
18083 * New function RSA_check_key and new openssl rsa option -check
18084 for verifying the consistency of RSA keys.
18086 *Ulf Moeller, Bodo Moeller*
18088 * Various changes to make Win32 compile work:
18089 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18090 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18091 comparison" warnings.
18092 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18096 * Add a debugging option to PKCS#5 v2 key generation function: when
18097 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18098 derived keys are printed to stderr.
18102 * Copy the flags in ASN1_STRING_dup().
18104 *Roman E. Pavlov <pre@mo.msk.ru>*
18106 * The x509 application mishandled signing requests containing DSA
18107 keys when the signing key was also DSA and the parameters didn't match.
18109 It was supposed to omit the parameters when they matched the signing key:
18110 the verifying software was then supposed to automatically use the CA's
18111 parameters if they were absent from the end user certificate.
18113 Omitting parameters is no longer recommended. The test was also
18114 the wrong way round! This was probably due to unusual behaviour in
18115 EVP_cmp_parameters() which returns 1 if the parameters match.
18116 This meant that parameters were omitted when they *didn't* match and
18117 the certificate was useless. Certificates signed with 'ca' didn't have
18120 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18122 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18123 The interface is as follows:
18124 Applications can use
18125 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18126 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18127 "off" is now the default.
18128 The library internally uses
18129 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18130 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18131 to disable memory-checking temporarily.
18133 Some inconsistent states that previously were possible (and were
18134 even the default) are now avoided.
18136 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18137 with each memory chunk allocated; this is occasionally more helpful
18138 than just having a counter.
18140 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18142 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18147 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18148 which largely parallels "options", but is for changing API behaviour,
18149 whereas "options" are about protocol behaviour.
18150 Initial "mode" flags are:
18152 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18153 a single record has been written.
18154 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18155 retries use the same buffer location.
18156 (But all of the contents must be
18161 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18164 * Fix problems with no-hmac etc.
18166 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18168 * New functions RSA_get_default_method(), RSA_set_method() and
18169 RSA_get_method(). These allows replacement of RSA_METHODs without having
18170 to mess around with the internals of an RSA structure.
18174 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18175 Also really enable memory leak checks in openssl.c and in some
18178 *Chad C. Mulligan, Bodo Moeller*
18180 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18181 up the length of negative integers. This has now been simplified to just
18182 store the length when it is first determined and use it later, rather
18183 than trying to keep track of where data is copied and updating it to
18185 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18187 * Add a new function PKCS7_signatureVerify. This allows the verification
18188 of a PKCS#7 signature but with the signing certificate passed to the
18189 function itself. This contrasts with PKCS7_dataVerify which assumes the
18190 certificate is present in the PKCS#7 structure. This isn't always the
18191 case: certificates can be omitted from a PKCS#7 structure and be
18192 distributed by "out of band" means (such as a certificate database).
18196 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18197 function prototypes in pem.h, also change util/mkdef.pl to add the
18198 necessary function names.
18202 * mk1mf.pl (used by Windows builds) did not properly read the
18203 options set by Configure in the top level Makefile, and Configure
18204 was not even able to write more than one option correctly.
18205 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18209 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18210 file to be loaded from a BIO or FILE pointer. The BIO version will
18211 for example allow memory BIOs to contain config info.
18215 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18216 Whoever hopes to achieve shared-library compatibility across versions
18217 must use this, not the compile-time macro.
18218 (Exercise 0.9.4: Which is the minimum library version required by
18220 Note: All this applies only to multi-threaded programs, others don't
18225 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18226 through a BIO pair triggered the default case, i.e.
18227 SSLerr(...,SSL_R_UNKNOWN_STATE).
18231 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18232 can use the SSL library even if none of the specific BIOs is
18237 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18238 for the encoded length.
18240 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18242 * Add initial documentation of the X509V3 functions.
18246 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18247 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18248 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18249 secure PKCS#8 private key format with a high iteration count.
18253 * Fix determination of Perl interpreter: A perl or perl5
18254 *directory* in $PATH was also accepted as the interpreter.
18256 *Ralf S. Engelschall*
18258 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18259 wrong with it but it was very old and did things like calling
18260 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18261 unusual formatting.
18265 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18266 to use the new extension code.
18270 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18271 with macros. This should make it easier to change their form, add extra
18272 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18277 * Add to configuration table a new entry that can specify an alternative
18278 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18279 according to Mark Crispin <MRC@Panda.COM>.
18283 * DES CBC did not update the IV. Weird.
18287 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18288 Changing the behaviour of the former might break existing programs --
18289 where IV updating is needed, des_ncbc_encrypt can be used.
18292 * When bntest is run from "make test" it drives bc to check its
18293 calculations, as well as internally checking them. If an internal check
18294 fails, it needs to cause bc to give a non-zero result or make test carries
18295 on without noticing the failure. Fixed.
18299 * DES library cleanups.
18303 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18304 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18305 ciphers. NOTE: although the key derivation function has been verified
18306 against some published test vectors it has not been extensively tested
18307 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18312 * Instead of "mkdir -p", which is not fully portable, use new
18313 Perl script "util/mkdir-p.pl".
18317 * Rewrite the way password based encryption (PBE) is handled. It used to
18318 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18319 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18320 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18321 the 'parameter' field of the AlgorithmIdentifier is passed to the
18322 underlying key generation function so it must do its own ASN1 parsing.
18323 This has also changed the EVP_PBE_CipherInit() function which now has a
18324 'parameter' argument instead of literal salt and iteration count values
18325 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18329 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18330 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18331 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18332 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18333 value was just used as a "magic string" and not used directly its
18334 value doesn't matter.
18338 * Introduce some semblance of const correctness to BN. Shame C doesn't
18343 * "linux-sparc64" configuration (ultrapenguin).
18345 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18346 "linux-sparc" configuration.
18348 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18350 * config now generates no-xxx options for missing ciphers.
18354 * Support the EBCDIC character set (work in progress).
18355 File ebcdic.c not yet included because it has a different license.
18357 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18359 * Support BS2000/OSD-POSIX.
18361 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18363 * Make callbacks for key generation use `void *` instead of `char *`.
18367 * Make S/MIME samples compile (not yet tested).
18371 * Additional typesafe stacks.
18375 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18379 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18381 * New configuration variant "sco5-gcc".
18383 * Updated some demos.
18385 *Sean O Riordain, Wade Scholine*
18387 * Add missing BIO_free at exit of pkcs12 application.
18391 * Fix memory leak in conf.c.
18395 * Updates for Win32 to assembler version of MD5.
18399 * Set #! path to perl in `apps/der_chop` to where we found it
18400 instead of using a fixed path.
18404 * SHA library changes for irix64-mips4-cc.
18408 * Improvements for VMS support.
18412 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18414 * Bignum library bug fix. IRIX 6 passes "make test" now!
18415 This also avoids the problems with SC4.2 and unpatched SC5.
18417 *Andy Polyakov <appro@fy.chalmers.se>*
18419 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18420 These are required because of the typesafe stack would otherwise break
18421 existing code. If old code used a structure member which used to be STACK
18422 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18423 sk_num or sk_value it would produce an error because the num, data members
18424 are not present in STACK_OF. Now it just produces a warning. sk_set
18425 replaces the old method of assigning a value to sk_value
18426 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18427 that does this will no longer work (and should use sk_set instead) but
18428 this could be regarded as a "questionable" behaviour anyway.
18432 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18433 correctly handle encrypted S/MIME data.
18437 * Change type of various DES function arguments from des_cblock
18438 (which means, in function argument declarations, pointer to char)
18439 to des_cblock * (meaning pointer to array with 8 char elements),
18440 which allows the compiler to do more typechecking; it was like
18441 that back in SSLeay, but with lots of ugly casts.
18443 Introduce new type const_des_cblock.
18447 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18448 problems: find RecipientInfo structure that matches recipient certificate
18449 and initialise the ASN1 structures properly based on passed cipher.
18453 * Belatedly make the BN tests actually check the results.
18457 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18458 to and from BNs: it was completely broken. New compilation option
18459 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18460 key elements as negative integers.
18464 * Reorganize and speed up MD5.
18466 *Andy Polyakov <appro@fy.chalmers.se>*
18470 *Richard Levitte <richard@levitte.org>*
18472 * New option -out to asn1parse to allow the parsed structure to be
18473 output to a file. This is most useful when combined with the -strparse
18474 option to examine the output of things like OCTET STRINGS.
18478 * Make SSL library a little more fool-proof by not requiring any longer
18479 that `SSL_set_{accept,connect}_state` be called before
18480 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18481 in many applications because usually everything *appeared* to work as
18482 intended anyway -- now it really works as intended).
18486 * Move openssl.cnf out of lib/.
18490 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18491 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18492 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18494 *Ralf S. Engelschall*
18496 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18497 handle PKCS#7 enveloped data properly.
18499 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18501 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18502 copying pointers. The cert_st handling is changed by this in
18503 various ways (and thus what used to be known as ctx->default_cert
18504 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18505 any longer when s->cert does not give us what we need).
18506 ssl_cert_instantiate becomes obsolete by this change.
18507 As soon as we've got the new code right (possibly it already is?),
18508 we have solved a couple of bugs of the earlier code where s->cert
18509 was used as if it could not have been shared with other SSL structures.
18511 Note that using the SSL API in certain dirty ways now will result
18512 in different behaviour than observed with earlier library versions:
18513 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18514 does not influence s as it used to.
18516 In order to clean up things more thoroughly, inside SSL_SESSION
18517 we don't use CERT any longer, but a new structure SESS_CERT
18518 that holds per-session data (if available); currently, this is
18519 the peer's certificate chain and, for clients, the server's certificate
18520 and temporary key. CERT holds only those values that can have
18521 meaningful defaults in an SSL_CTX.
18525 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18526 from the internal representation. Various PKCS#7 fixes: remove some
18527 evil casts and set the enc_dig_alg field properly based on the signing
18532 * Allow PKCS#12 password to be set from the command line or the
18533 environment. Let 'ca' get its config file name from the environment
18534 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18539 * Allow certificate policies extension to use an IA5STRING for the
18540 organization field. This is contrary to the PKIX definition but
18541 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18546 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18547 without disallowing inline assembler and the like for non-pedantic builds.
18551 * Support Borland C++ builder.
18553 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18559 * SHA-1 cleanups and performance enhancements.
18561 *Andy Polyakov <appro@fy.chalmers.se>*
18563 * Sparc v8plus assembler for the bignum library.
18565 *Andy Polyakov <appro@fy.chalmers.se>*
18567 * Accept any -xxx and +xxx compiler options in Configure.
18571 * Update HPUX configuration.
18575 * Add missing `sk_<type>_unshift()` function to safestack.h
18577 *Ralf S. Engelschall*
18579 * New function SSL_CTX_use_certificate_chain_file that sets the
18580 "extra_cert"s in addition to the certificate. (This makes sense
18581 only for "PEM" format files, as chains as a whole are not
18586 * Support verify_depth from the SSL API.
18587 x509_vfy.c had what can be considered an off-by-one-error:
18588 Its depth (which was not part of the external interface)
18589 was actually counting the number of certificates in a chain;
18590 now it really counts the depth.
18594 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18595 instead of X509err, which often resulted in confusing error
18596 messages since the error codes are not globally unique
18597 (e.g. an alleged error in ssl3_accept when a certificate
18598 didn't match the private key).
18600 * New function SSL_CTX_set_session_id_context that allows to set a default
18601 value (so that you don't need SSL_set_session_id_context for each
18602 connection using the SSL_CTX).
18606 * OAEP decoding bug fix.
18610 * Support INSTALL_PREFIX for package builders, as proposed by
18615 * New Configure options "threads" and "no-threads". For systems
18616 where the proper compiler options are known (currently Solaris
18617 and Linux), "threads" is the default.
18621 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18625 * Install various scripts to $(OPENSSLDIR)/misc, not to
18626 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18627 such as /usr/local/bin.
18631 * "make linux-shared" to build shared libraries.
18633 *Niels Poppe <niels@netbox.org>*
18635 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18639 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18640 extension adding in x509 utility.
18644 * Remove NOPROTO sections and error code comments.
18648 * Partial rewrite of the DEF file generator to now parse the ANSI
18653 * New Configure options --prefix=DIR and --openssldir=DIR.
18657 * Complete rewrite of the error code script(s). It is all now handled
18658 by one script at the top level which handles error code gathering,
18659 header rewriting and C source file generation. It should be much better
18660 than the old method: it now uses a modified version of Ulf's parser to
18661 read the ANSI prototypes in all header files (thus the old K&R definitions
18662 aren't needed for error creation any more) and do a better job of
18663 translating function codes into names. The old 'ASN1 error code embedded
18664 in a comment' is no longer necessary and it doesn't use .err files which
18665 have now been deleted. Also the error code call doesn't have to appear all
18666 on one line (which resulted in some large lines...).
18670 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18674 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18675 0 (which usually indicates a closed connection), but continue reading.
18679 * Fix some race conditions.
18683 * Add support for CRL distribution points extension. Add Certificate
18684 Policies and CRL distribution points documentation.
18688 * Move the autogenerated header file parts to crypto/opensslconf.h.
18692 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18693 8 of keying material. Merlin has also confirmed interop with this fix
18694 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18696 *Merlin Hughes <merlin@baltimore.ie>*
18698 * Fix lots of warnings.
18700 *Richard Levitte <levitte@stacken.kth.se>*
18702 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18703 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18705 *Richard Levitte <levitte@stacken.kth.se>*
18707 * Fix problems with sizeof(long) == 8.
18709 *Andy Polyakov <appro@fy.chalmers.se>*
18711 * Change functions to ANSI C.
18715 * Fix typos in error codes.
18717 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18719 * Remove defunct assembler files from Configure.
18723 * SPARC v8 assembler BIGNUM implementation.
18725 *Andy Polyakov <appro@fy.chalmers.se>*
18727 * Support for Certificate Policies extension: both print and set.
18728 Various additions to support the r2i method this uses.
18732 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18733 return a const string when you are expecting an allocated buffer.
18737 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18738 types DirectoryString and DisplayText.
18742 * Add code to allow r2i extensions to access the configuration database,
18743 add an LHASH database driver and add several ctx helper functions.
18747 * Fix an evil bug in bn_expand2() which caused various BN functions to
18748 fail when they extended the size of a BIGNUM.
18752 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18753 support typesafe stack.
18757 * Fix typo in SSL_[gs]et_options().
18759 *Nils Frostberg <nils@medcom.se>*
18761 * Delete various functions and files that belonged to the (now obsolete)
18762 old X509V3 handling code.
18766 * New Configure option "rsaref".
18770 * Don't auto-generate pem.h.
18774 * Introduce type-safe ASN.1 SETs.
18778 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18780 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18782 * Introduce type-safe STACKs. This will almost certainly break lots of code
18783 that links with OpenSSL (well at least cause lots of warnings), but fear
18784 not: the conversion is trivial, and it eliminates loads of evil casts. A
18785 few STACKed things have been converted already. Feel free to convert more.
18786 In the fullness of time, I'll do away with the STACK type altogether.
18790 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18791 specified in `<certfile>` by updating the entry in the index.txt file.
18792 This way one no longer has to edit the index.txt file manually for
18793 revoking a certificate. The -revoke option does the gory details now.
18795 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18797 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18798 `-text` option at all and this way the `-noout -text` combination was
18799 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18801 *Ralf S. Engelschall*
18803 * Make sure a corresponding plain text error message exists for the
18804 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18805 verify callback function determined that a certificate was revoked.
18807 *Ralf S. Engelschall*
18809 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18810 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18811 all available ciphers including rc5, which was forgotten until now.
18812 In order to let the testing shell script know which algorithms
18813 are available, a new (up to now undocumented) command
18814 `openssl list-cipher-commands` is used.
18818 * Bugfix: s_client occasionally would sleep in select() when
18819 it should have checked SSL_pending() first.
18823 * New functions DSA_do_sign and DSA_do_verify to provide access to
18824 the raw DSA values prior to ASN.1 encoding.
18828 * Tweaks to Configure
18830 *Niels Poppe <niels@netbox.org>*
18832 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18837 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18841 * New config option to avoid instructions that are illegal on the 80386.
18842 The default code is faster, but requires at least a 486.
18846 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18847 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18848 same as SSL2_VERSION anyway.
18852 * New "-showcerts" option for s_client.
18856 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18857 application. Various cleanups and fixes.
18861 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18862 modify error routines to work internally. Add error codes and PBE init
18863 to library startup routines.
18867 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18868 packing functions to asn1 and evp. Changed function names and error
18869 codes along the way.
18873 * PKCS12 integration: and so it begins... First of several patches to
18874 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18875 objects to objects.h
18879 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18880 and display support for Thawte strong extranet extension.
18884 * Add LinuxPPC support.
18886 *Jeff Dubrule <igor@pobox.org>*
18888 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18889 bn_div_words in alpha.s.
18891 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18893 * Make sure the RSA OAEP test is skipped under -DRSAref because
18894 OAEP isn't supported when OpenSSL is built with RSAref.
18896 *Ulf Moeller <ulf@fitug.de>*
18898 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18899 so they no longer are missing under -DNOPROTO.
18901 *Soren S. Jorvang <soren@t.dk>*
18903 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18905 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18906 doesn't work when the session is reused. Coming soon!
18910 * Fix a security hole, that allows sessions to be reused in the wrong
18911 context thus bypassing client cert protection! All software that uses
18912 client certs and session caches in multiple contexts NEEDS PATCHING to
18913 allow session reuse! A fuller solution is in the works.
18915 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18917 * Some more source tree cleanups (removed obsolete files
18918 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18919 permission on "config" script to be executable) and a fix for the INSTALL
18922 *Ulf Moeller <ulf@fitug.de>*
18924 * Remove some legacy and erroneous uses of malloc, free instead of
18927 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18929 * Make rsa_oaep_test return non-zero on error.
18931 *Ulf Moeller <ulf@fitug.de>*
18933 * Add support for native Solaris shared libraries. Configure
18934 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18935 if someone would make that last step automatic.
18937 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18939 * ctx_size was not built with the right compiler during "make links". Fixed.
18943 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18944 except NULL ciphers". This means the default cipher list will no longer
18945 enable NULL ciphers. They need to be specifically enabled e.g. with
18946 the string "DEFAULT:eNULL".
18950 * Fix to RSA private encryption routines: if p < q then it would
18951 occasionally produce an invalid result. This will only happen with
18952 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18956 * Be less restrictive and allow also `perl util/perlpath.pl
18957 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18958 because this way one can also use an interpreter named `perl5` (which is
18959 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18960 installed as `perl`).
18962 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18964 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18966 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18968 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18969 advapi32.lib to Win32 build and change the pem test comparison
18970 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18971 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18972 and crypto/des/ede_cbcm_enc.c.
18976 * DES quad checksum was broken on big-endian architectures. Fixed.
18980 * Comment out two functions in bio.h that aren't implemented. Fix up the
18981 Win32 test batch file so it (might) work again. The Win32 test batch file
18982 is horrible: I feel ill....
18986 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18987 in e_os.h. Audit of header files to check ANSI and non ANSI
18988 sections: 10 functions were absent from non ANSI section and not exported
18989 from Windows DLLs. Fixed up libeay.num for new functions.
18993 * Make `openssl version` output lines consistent.
18995 *Ralf S. Engelschall*
18997 * Fix Win32 symbol export lists for BIO functions: Added
18998 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18999 to ms/libeay{16,32}.def.
19001 *Ralf S. Engelschall*
19003 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
19004 fine under Unix and passes some trivial tests I've now added. But the
19005 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
19006 added to make sure no one expects that this stuff really works in the
19007 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
19008 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
19011 *Ralf S. Engelschall*
19013 * Fix the generation of two part addresses in perl.
19015 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
19017 * Add config entry for Linux on MIPS.
19019 *John Tobey <jtobey@channel1.com>*
19021 * Make links whenever Configure is run, unless we are on Windoze.
19025 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
19026 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19031 * Add a useful kludge to allow package maintainers to specify compiler and
19032 other platforms details on the command line without having to patch the
19033 Configure script every time: One now can use
19034 `perl Configure <id>:<details>`,
19035 i.e. platform ids are allowed to have details appended
19036 to them (separated by colons). This is treated as there would be a static
19037 pre-configured entry in Configure's %table under key `<id>` with value
19038 `<details>` and `perl Configure <id>` is called. So, when you want to
19039 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19040 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19041 now, which overrides the FreeBSD-elf entry on-the-fly.
19043 *Ralf S. Engelschall*
19045 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19049 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19050 on the `perl Configure ...` command line. This way one can compile
19051 OpenSSL libraries with Position Independent Code (PIC) which is needed
19052 for linking it into DSOs.
19054 *Ralf S. Engelschall*
19056 * Remarkably, export ciphers were totally broken and no-one had noticed!
19061 * Cleaned up the LICENSE document: The official contact for any license
19062 questions now is the OpenSSL core team under openssl-core@openssl.org.
19063 And add a paragraph about the dual-license situation to make sure people
19064 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19065 to the OpenSSL toolkit.
19067 *Ralf S. Engelschall*
19069 * General source tree makefile cleanups: Made `making xxx in yyy...`
19070 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19071 Additionally cleaned up the `make links` target: Remove unnecessary
19072 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19073 to speed processing and no longer clutter the display with confusing
19074 stuff. Instead only the actually done links are displayed.
19076 *Ralf S. Engelschall*
19078 * Permit null encryption ciphersuites, used for authentication only. It used
19079 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19080 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19085 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19086 signed attributes when verifying signatures (this would break them),
19087 the detached data encoding was wrong and public keys obtained using
19088 X509_get_pubkey() weren't freed.
19092 * Add text documentation for the BUFFER functions. Also added a work around
19093 to a Win95 console bug. This was triggered by the password read stuff: the
19094 last character typed gets carried over to the next fread(). If you were
19095 generating a new cert request using 'req' for example then the last
19096 character of the passphrase would be CR which would then enter the first
19101 * Added the new 'Includes OpenSSL Cryptography Software' button as
19102 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19103 button and can be used by applications based on OpenSSL to show the
19104 relationship to the OpenSSL project.
19106 *Ralf S. Engelschall*
19108 * Remove confusing variables in function signatures in files
19109 ssl/ssl_lib.c and ssl/ssl.h.
19111 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19113 * Don't install bss_file.c under PREFIX/include/
19115 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19117 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19118 functions that return function pointers and has support for NT specific
19119 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19120 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19121 unsigned to signed types: this was killing the Win32 compile.
19125 * Add new certificate file to stack functions,
19126 SSL_add_dir_cert_subjects_to_stack() and
19127 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19128 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19129 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19130 This means that Apache-SSL and similar packages don't have to mess around
19131 to add as many CAs as they want to the preferred list.
19135 * Experiment with doxygen documentation. Currently only partially applied to
19137 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19138 openssl.doxy as the configuration file.
19142 * Get rid of remaining C++-style comments which strict C compilers hate.
19144 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19146 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19147 compiled in by default: it has problems with large keys.
19151 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19152 DH private keys and/or callback functions which directly correspond to
19153 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19154 is needed for applications which have to configure certificates on a
19155 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19157 For the RSA certificate situation is makes no difference, but
19158 for the DSA certificate situation this fixes the "no shared cipher"
19159 problem where the OpenSSL cipher selection procedure failed because the
19160 temporary keys were not overtaken from the context and the API provided
19161 no way to reconfigure them.
19162 The new functions now let applications reconfigure the stuff and they
19163 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19164 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19165 non-public-API function ssl_cert_instantiate() is used as a helper
19166 function and also to reduce code redundancy inside ssl_rsa.c.
19168 *Ralf S. Engelschall*
19170 * Move s_server -dcert and -dkey options out of the undocumented feature
19171 area because they are useful for the DSA situation and should be
19172 recognized by the users.
19174 *Ralf S. Engelschall*
19176 * Fix the cipher decision scheme for export ciphers: the export bits are
19177 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19178 SSL_EXP_MASK. So, the original variable has to be used instead of the
19179 already masked variable.
19181 *Richard Levitte <levitte@stacken.kth.se>*
19183 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19185 *Richard Levitte <levitte@stacken.kth.se>*
19187 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19188 from `int` to `unsigned int` because it is a length and initialized by
19189 EVP_DigestFinal() which expects an `unsigned int *`.
19191 *Richard Levitte <levitte@stacken.kth.se>*
19193 * Don't hard-code path to Perl interpreter on shebang line of Configure
19194 script. Instead use the usual Shell->Perl transition trick.
19196 *Ralf S. Engelschall*
19198 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19199 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19200 -noout -modulus` as it's already the case for `openssl rsa -noout
19201 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19202 currently the public key is printed (a decision which was already done by
19203 `openssl dsa -modulus` in the past) which serves a similar purpose.
19204 Additionally the NO_RSA no longer completely removes the whole -modulus
19205 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19208 *Ralf S. Engelschall*
19210 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19211 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19213 *Arne Ansper <arne@ats.cyber.ee>*
19215 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19216 to be added. Now both 'req' and 'ca' can use new objects defined in the
19221 * Add cool BIO that does syslog (or event log on NT).
19223 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19225 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19226 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19227 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19228 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19232 * Add preliminary config info for new extension code.
19236 * Make RSA_NO_PADDING really use no padding.
19238 *Ulf Moeller <ulf@fitug.de>*
19240 * Generate errors when private/public key check is done.
19244 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19245 for some CRL extensions and new objects added.
19249 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19250 key usage extension and fuller support for authority key id.
19254 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19255 padding method for RSA, which is recommended for new applications in PKCS
19256 #1 v2.0 (RFC 2437, October 1998).
19257 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19258 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19259 against Bleichbacher's attack on RSA.
19260 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19263 * Updates to the new SSL compression code
19265 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19267 * Fix so that the version number in the master secret, when passed
19268 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19269 (because the server will not accept higher), that the version number
19270 is 0x03,0x01, not 0x03,0x00
19272 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19274 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19275 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19276 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19280 * Support for RAW extensions where an arbitrary extension can be
19281 created by including its DER encoding. See `apps/openssl.cnf` for
19286 * Make sure latest Perl versions don't interpret some generated C array
19287 code as Perl array code in the crypto/err/err_genc.pl script.
19289 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19291 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19292 not many people have the assembler. Various Win32 compilation fixes and
19293 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19294 build instructions.
19298 * Modify configure script 'Configure' to automatically create crypto/date.h
19299 file under Win32 and also build pem.h from pem.org. New script
19300 util/mkfiles.pl to create the MINFO file on environments that can't do a
19301 'make files': perl util/mkfiles.pl >MINFO should work.
19305 * Major rework of DES function declarations, in the pursuit of correctness
19306 and purity. As a result, many evil casts evaporated, and some weirdness,
19307 too. You may find this causes warnings in your code. Zapping your evil
19308 casts will probably fix them. Mostly.
19312 * Fix for a typo in asn1.h. Bug fix to object creation script
19313 obj_dat.pl. It considered a zero in an object definition to mean
19314 "end of object": none of the objects in objects.h have any zeros
19315 so it wasn't spotted.
19317 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19319 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19320 Masking (CBCM). In the absence of test vectors, the best I have been able
19321 to do is check that the decrypt undoes the encrypt, so far. Send me test
19322 vectors if you have them.
19326 * Correct calculation of key length for export ciphers (too much space was
19327 allocated for null ciphers). This has not been tested!
19331 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19332 message is now correct (it understands "crypto" and "ssl" on its
19333 command line). There is also now an "update" option. This will update
19334 the util/ssleay.num and util/libeay.num files with any new functions.
19336 perl util/mkdef.pl crypto ssl update
19337 it will update them.
19341 * Overhauled the Perl interface:
19342 - ported BN stuff to OpenSSL's different BN library
19343 - made the perl/ source tree CVS-aware
19344 - renamed the package from SSLeay to OpenSSL (the files still contain
19345 their history because I've copied them in the repository)
19346 - removed obsolete files (the test scripts will be replaced
19347 by better Test::Harness variants in the future)
19349 *Ralf S. Engelschall*
19351 * First cut for a very conservative source tree cleanup:
19352 1. merge various obsolete readme texts into doc/ssleay.txt
19353 where we collect the old documents and readme texts.
19354 2. remove the first part of files where I'm already sure that we no
19355 longer need them because of three reasons: either they are just temporary
19356 files which were left by Eric or they are preserved original files where
19357 I've verified that the diff is also available in the CVS via "cvs diff
19358 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19359 the crypto/md/ stuff).
19361 *Ralf S. Engelschall*
19363 * More extension code. Incomplete support for subject and issuer alt
19364 name, issuer and authority key id. Change the i2v function parameters
19365 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19366 what that's for :-) Fix to ASN1 macro which messed up
19367 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19371 * Preliminary support for ENUMERATED type. This is largely copied from the
19376 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19378 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19380 * Make sure `make rehash` target really finds the `openssl` program.
19382 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19384 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19385 like to hear about it if this slows down other processors.
19389 * Add CygWin32 platform information to Configure script.
19391 *Alan Batie <batie@aahz.jf.intel.com>*
19393 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19395 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19397 * New program nseq to manipulate netscape certificate sequences
19401 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19406 * Fixes to BN code. Previously the default was to define BN_RECURSION
19407 but the BN code had some problems that would cause failures when
19408 doing certificate verification and some other functions.
19410 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19412 * Add ASN1 and PEM code to support netscape certificate sequences.
19416 * Add ASN1 and PEM code to support netscape certificate sequences.
19420 * Add several PKIX and private extended key usage OIDs.
19424 * Modify the 'ca' program to handle the new extension code. Modify
19425 openssl.cnf for new extension format, add comments.
19429 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19430 and add a sample to openssl.cnf so req -x509 now adds appropriate
19435 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19436 error code, add initial support to X509_print() and x509 application.
19440 * Takes a deep breath and start adding X509 V3 extension support code. Add
19441 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19442 stuff is currently isolated and isn't even compiled yet.
19446 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19447 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19448 Removed the versions check from X509 routines when loading extensions:
19449 this allows certain broken certificates that don't set the version
19450 properly to be processed.
19454 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19455 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19456 can still be regenerated with "make depend".
19460 * Spelling mistake in C version of CAST-128.
19462 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19464 * Changes to the error generation code. The perl script err-code.pl
19465 now reads in the old error codes and retains the old numbers, only
19466 adding new ones if necessary. It also only changes the .err files if new
19467 codes are added. The makefiles have been modified to only insert errors
19468 when needed (to avoid needlessly modifying header files). This is done
19469 by only inserting errors if the .err file is newer than the auto generated
19470 C file. To rebuild all the error codes from scratch (the old behaviour)
19471 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19472 or delete all the .err files.
19476 * CAST-128 was incorrectly implemented for short keys. The C version has
19477 been fixed, but is untested. The assembler versions are also fixed, but
19478 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19479 to regenerate it if needed.
19480 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19481 Hagino <itojun@kame.net>*
19483 * File was opened incorrectly in randfile.c.
19485 *Ulf Möller <ulf@fitug.de>*
19487 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19488 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19489 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19490 al: it's just almost always a UTCTime. Note this patch adds new error
19491 codes so do a "make errors" if there are problems.
19495 * Correct Linux 1 recognition in config.
19497 *Ulf Möller <ulf@fitug.de>*
19499 * Remove pointless MD5 hash when using DSA keys in ca.
19501 *Anonymous <nobody@replay.com>*
19503 * Generate an error if given an empty string as a cert directory. Also
19504 generate an error if handed NULL (previously returned 0 to indicate an
19505 error, but didn't set one).
19507 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19509 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19513 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19514 parameters. This was causing a warning which killed off the Win32 compile.
19518 * Remove C++ style comments from crypto/bn/bn_local.h.
19520 *Neil Costigan <neil.costigan@celocom.com>*
19522 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19523 based on a text string, looking up short and long names and finally
19524 "dot" format. The "dot" format stuff didn't work. Added new function
19525 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19526 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19527 OID is not part of the table.
19531 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19532 X509_LOOKUP_by_alias().
19536 * Sort openssl functions by name.
19540 * Get the `gendsa` command working and add it to the `list` command. Remove
19541 encryption from sample DSA keys (in case anyone is interested the password
19546 * Make *all* `*_free` functions accept a NULL pointer.
19548 *Frans Heymans <fheymans@isaserver.be>*
19550 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19553 *Anonymous <nobody@replay.com>*
19555 * s_server should send the CAfile as acceptable CAs, not its own cert.
19557 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19559 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19561 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19563 * Temp key "for export" tests were wrong in s3_srvr.c.
19565 *Anonymous <nobody@replay.com>*
19567 * Add prototype for temp key callback functions
19568 SSL_CTX_set_tmp_{rsa,dh}_callback().
19572 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19573 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19577 * X509_name_add_entry() freed the wrong thing after an error.
19579 *Arne Ansper <arne@ats.cyber.ee>*
19581 * rsa_eay.c would attempt to free a NULL context.
19583 *Arne Ansper <arne@ats.cyber.ee>*
19585 * BIO_s_socket() had a broken should_retry() on Windoze.
19587 *Arne Ansper <arne@ats.cyber.ee>*
19589 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19591 *Arne Ansper <arne@ats.cyber.ee>*
19593 * Make sure the already existing X509_STORE->depth variable is initialized
19594 in X509_STORE_new(), but document the fact that this variable is still
19595 unused in the certificate verification process.
19597 *Ralf S. Engelschall*
19599 * Fix the various library and `apps/` files to free up pkeys obtained from
19600 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19604 * Fix reference counting in X509_PUBKEY_get(). This makes
19605 demos/maurice/example2.c work, amongst others, probably.
19607 *Steve Henson and Ben Laurie*
19609 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19610 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19611 are no longer created. This way we have a single and consistent command
19612 line interface `openssl <command>`, similar to `cvs <command>`.
19614 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19616 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19617 BIT STRING wrapper always have zero unused bits.
19621 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19625 * Make the top-level INSTALL documentation easier to understand.
19629 * Makefiles updated to exit if an error occurs in a sub-directory
19630 make (including if user presses ^C) [Paul Sutton]
19632 * Make Montgomery context stuff explicit in RSA data structure.
19636 * Fix build order of pem and err to allow for generated pem.h.
19640 * Fix renumbering bug in X509_NAME_delete_entry().
19644 * Enhanced the err-ins.pl script so it makes the error library number
19645 global and can add a library name. This is needed for external ASN1 and
19646 other error libraries.
19650 * Fixed sk_insert which never worked properly.
19654 * Fix ASN1 macros so they can handle indefinite length constructed
19655 EXPLICIT tags. Some non standard certificates use these: they can now
19660 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19661 into a single doc/ssleay.txt bundle. This way the information is still
19662 preserved but no longer messes up this directory. Now it's new room for
19663 the new set of documentation files.
19665 *Ralf S. Engelschall*
19667 * SETs were incorrectly DER encoded. This was a major pain, because they
19668 shared code with SEQUENCEs, which aren't coded the same. This means that
19669 almost everything to do with SETs or SEQUENCEs has either changed name or
19670 number of arguments.
19672 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19674 * Fix test data to work with the above.
19678 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19679 was already fixed by Eric for 0.9.1 it seems.
19681 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19683 * Autodetect FreeBSD3.
19687 * Fix various bugs in Configure. This affects the following platforms:
19691 unixware-2.0-pentium
19696 * Eliminate generated files from CVS. Reorder tests to regenerate files
19697 before they are needed.
19701 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19705 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19707 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19708 changed SSLeay to OpenSSL in version strings.
19710 *Ralf S. Engelschall*
19712 * Some fixups to the top-level documents.
19716 * Fixed the nasty bug where rsaref.h was not found under compile-time
19717 because the symlink to include/ was missing.
19719 *Ralf S. Engelschall*
19721 * Incorporated the popular no-RSA/DSA-only patches
19722 which allow to compile an RSA-free SSLeay.
19724 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19726 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19727 when "ssleay" is still not found.
19729 *Ralf S. Engelschall*
19731 * Added more platforms to Configure: Cray T3E, HPUX 11,
19733 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19735 * Updated the README file.
19737 *Ralf S. Engelschall*
19739 * Added various .cvsignore files in the CVS repository subdirs
19740 to make a "cvs update" really silent.
19742 *Ralf S. Engelschall*
19744 * Recompiled the error-definition header files and added
19745 missing symbols to the Win32 linker tables.
19747 *Ralf S. Engelschall*
19749 * Cleaned up the top-level documents;
19750 o new files: CHANGES and LICENSE
19751 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19752 o merged COPYRIGHT into LICENSE
19753 o removed obsolete TODO file
19754 o renamed MICROSOFT to INSTALL.W32
19756 *Ralf S. Engelschall*
19758 * Removed dummy files from the 0.9.1b source tree:
19759 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19760 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19761 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19762 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19763 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19765 *Ralf S. Engelschall*
19767 * Added various platform portability fixes.
19771 * The Genesis of the OpenSSL rpject:
19772 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19773 Young and Tim J. Hudson created while they were working for C2Net until
19776 *The OpenSSL Project*
19778 ### Changes between 0.9.0b and 0.9.1b [not released]
19780 * Updated a few CA certificates under certs/
19784 * Changed some BIGNUM api stuff.
19788 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19789 DGUX x86, Linux Alpha, etc.
19793 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19794 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19799 * Add -strparse option to asn1pars program which parses nested
19802 *Dr Stephen Henson <shenson@bigfoot.com>*
19804 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19808 * DSA fix for "ca" program.
19812 * Added "-genkey" option to "dsaparam" program.
19816 * Added RIPE MD160 (rmd160) message digest.
19820 * Added -a (all) option to "ssleay version" command.
19824 * Added PLATFORM define which is the id given to Configure.
19828 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19832 * Extended the ASN.1 parser routines.
19836 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19840 * Added a BN_CTX to the BN library.
19844 * Fixed the weak key values in DES library
19848 * Changed API in EVP library for cipher aliases.
19852 * Added support for RC2/64bit cipher.
19856 * Converted the lhash library to the crypto/mem.c functions.
19860 * Added more recognized ASN.1 object ids.
19864 * Added more RSA padding checks for SSL/TLS.
19868 * Added BIO proxy/filter functionality.
19872 * Added extra_certs to SSL_CTX which can be used
19873 send extra CA certificates to the client in the CA cert chain sending
19874 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19878 * Now Fortezza is denied in the authentication phase because
19879 this is key exchange mechanism is not supported by SSLeay at all.
19883 * Additional PKCS1 checks.
19887 * Support the string "TLSv1" for all TLS v1 ciphers.
19891 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19892 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19896 * Fixed a few memory leaks.
19900 * Fixed various code and comment typos.
19904 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19905 bytes sent in the client random.
19907 *Edward Bishop <ebishop@spyglass.com>*
19911 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
19912 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
19913 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
19914 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
19915 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
19916 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
19917 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
19918 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
19919 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
19920 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
19921 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
19922 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
19923 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19924 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
19925 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19926 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19927 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19928 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19929 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19930 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19931 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19932 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19933 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19934 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19935 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19936 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19937 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19938 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19939 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19940 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19941 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19942 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19943 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19944 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19945 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19946 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19947 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19948 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19949 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19950 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19951 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19952 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19953 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19954 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19955 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19956 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19957 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19958 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19959 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19960 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19961 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19962 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19963 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19964 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19965 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19966 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19967 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19968 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19969 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19970 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19971 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19972 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19973 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19974 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19975 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19976 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19977 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19978 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19979 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19980 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19981 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19982 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19983 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19984 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19985 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19986 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19987 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19988 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19989 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19990 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19991 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19992 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19993 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19994 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19995 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19996 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19997 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19998 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19999 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
20000 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
20001 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
20002 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
20003 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
20004 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
20005 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
20006 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
20007 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
20008 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
20009 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
20010 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
20011 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
20012 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
20013 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
20014 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
20015 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
20016 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
20017 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
20018 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20019 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20020 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20021 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20022 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20023 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20024 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20025 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20026 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20027 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20028 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20029 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20030 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20031 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20032 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20033 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20034 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20035 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20036 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20037 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20038 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20039 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20040 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20041 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20042 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20043 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20044 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20045 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20046 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20047 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20048 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20049 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20050 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20051 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20052 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20053 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20054 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20055 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20056 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20057 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20058 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20059 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20060 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20061 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20062 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20063 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20064 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20065 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20066 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20067 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20068 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20069 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20070 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20071 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20072 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20073 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20074 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20075 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20076 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20077 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20078 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20079 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20080 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20081 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20082 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20083 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20084 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20085 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20086 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655