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 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
29 The previous fix for this timing side channel turned out to cause
30 a severe 2-3x performance regression in the typical use case
31 compared to 3.0.7. The new fix uses existing constant time
32 code paths, and restores the previous performance level while
33 fully eliminating all existing timing side channels.
34 The fix was developed by Bernd Edlinger with testing support
39 * Added an "advanced" command mode to s_client. Use this with the "-adv"
40 option. The old "basic" command mode recognises certain letters that must
41 always appear at the start of a line and cannot be escaped. The advanced
42 command mode enables commands to be entered anywhere and there is an
43 escaping mechanism. After starting s_client with "-adv" type "{help}"
44 to show a list of available commands.
48 * Add Raw Public Key (RFC7250) support. Authentication is supported
49 by matching keys against either local policy (TLSA records synthesised
50 from the expected keys) or DANE (TLSA records obtained by the
51 application from DNS). TLSA records will also match the same key in
52 the server certificate, should RPK use not happen to be negotiated.
56 * Added EC_GROUP_to_params which creates an OSSL_PARAM array
57 from a given EC_GROUP.
61 * Added support for pluggable (provider-based) TLS signature algorithms.
62 This enables TLS 1.3 authentication operations with algorithms embedded
63 in providers not included by default in OpenSSL. In combination with
64 the already available pluggable KEM and X.509 support, this enables
65 for example suitable providers to deliver post-quantum or quantum-safe
66 cryptography to OpenSSL users.
70 * Added support for Hybrid Public Key Encryption (HPKE) as defined
71 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
72 Message Layer Security (MLS) and other IETF specifications.
73 HPKE can also be used by other applications that require
74 encrypting "to" an ECDH public key. External APIs are defined in
75 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
79 * Add support for certificate compression (RFC8879), including
80 library support for Brotli and Zstandard compression.
84 * Add the ability to add custom attributes to PKCS12 files. Add a new API
85 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
86 for a user specified callback and optional argument.
87 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
88 added to the existing STACK_OF attrs.
92 * Major refactor of the libssl record layer
96 * Added a new BIO_s_dgram_mem() to read/write datagrams to memory
100 * Add a mac salt length option for the pkcs12 command.
104 * Add more SRTP protection profiles from RFC8723 and RFC8269.
108 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
110 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
112 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
113 supported and enabled.
117 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
118 to the list of ciphersuites providing Perfect Forward Secrecy as
119 required by SECLEVEL >= 3.
121 *Dmitry Belyavskiy, Nicola Tuveri*
123 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
124 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
125 SSL_get0_iana_groups() function-like macro, retrieves the list of
126 supported groups sent by the peer.
127 The function SSL_client_hello_get_extension_order() populates
128 a caller-supplied array with the list of extension types present in the
129 ClientHello, in order of appearance.
133 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
134 to make it possible to use empty passphrase strings.
138 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
139 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
140 of 160 bits and above and less than 224 bits were previously accepted by
141 default but are now no longer allowed. By default TLS compression was
142 already disabled in previous OpenSSL versions. At security level 2 it cannot
147 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
152 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
153 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
154 will need to load the legacy crypto provider.
157 * CCM8 cipher suites in TLS have been downgraded to security level zero
158 because they use a short authentication tag which lowers their strength.
162 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
167 * Add X.509 certificate codeSigning purpose and related checks on key usage and
168 extended key usage of the leaf certificate according to the CA/Browser Forum.
172 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
173 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
174 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
175 X.509 version 3 if the certificate information includes X.509 extensions.
179 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
180 such as adding a trace facility for debugging certificate chain building.
184 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
185 in particular supporting requests for central key generation, generalized
186 polling, and various types of genm/genp exchanges defined in CMP Updates.
190 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
191 like correcting the TLS and proxy support and adding tracing for debugging.
195 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
199 * `CMS_add0_cert()` and `CMS_add1_cert()` no more throw an error
200 if a certificate to be added is already present.
201 * `CMS_sign_ex()` and `CMS_sign()` now ignore any duplicate certificates
202 in their `certs` argument and no longer throw an error for them.
206 * Fixed and extended `util/check-format.pl` for checking adherence to the
207 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
208 The checks are meanwhile more complete and yield fewer false positives.
212 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
213 sending and receiving multiple messages in a single call. An implementation
214 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
218 * The `SSL_CERT_PATH` and `SSL_CERT_URI` environment variables are introduced.
219 `SSL_CERT_URI` can be used to specify a URI for a root certificate store. The
220 `SSL_CERT_PATH` environment variable specifies a delimiter-separated list of
221 paths which are searched for root certificates.
223 The existing `SSL_CERT_DIR` environment variable is deprecated.
224 `SSL_CERT_DIR` was previously used to specify either a delimiter-separated
225 list of paths or an URI, which is ambiguous. Setting `SSL_CERT_PATH` causes
226 `SSL_CERT_DIR` to be ignored for the purposes of determining root certificate
227 directories, and setting `SSL_CERT_URI` causes `SSL_CERT_DIR` to be ignored
228 for the purposes of determining root certificate stores.
232 * Support for loading root certificates from the Windows certificate store
233 has been added. The support is in the form of a store which recognises the
234 URI string of `org.openssl.winstore://`. This store is enabled by default and
235 can be disabled using the new compile-time option `no-winstore`.
239 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
240 kernel versions that support KTLS have a known bug in CCM processing. That
241 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
242 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
247 * Zerocopy KTLS sendfile() support on Linux.
251 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
252 decryption as a protection against Bleichenbacher-like attacks.
253 The RSA decryption API will now return a randomly generated deterministic
254 message instead of an error in case it detects an error when checking
255 padding during PKCS#1 v1.5 decryption. This is a general protection against
256 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
258 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
259 on the RSA decryption context.
263 * Support for Argon2d, Argon2i, Argon2id KDFs has been added along with
264 basic thread pool implementation for select platforms.
271 ### Changes between 3.1.0 and 3.1.1 [xx XXX xxxx]
273 * Add FIPS provider configuration option to disallow the use of
274 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
275 The option '-no_drbg_truncated_digests' can optionally be
276 supplied to 'openssl fipsinstall'.
280 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
281 that it does not enable policy checking. Thanks to David Benjamin for
282 discovering this issue.
287 * Fixed an issue where invalid certificate policies in leaf certificates are
288 silently ignored by OpenSSL and other certificate policy checks are skipped
289 for that certificate. A malicious CA could use this to deliberately assert
290 invalid certificate policies in order to circumvent policy checking on the
291 certificate altogether.
296 * Limited the number of nodes created in a policy tree to mitigate
297 against CVE-2023-0464. The default limit is set to 1000 nodes, which
298 should be sufficient for most installations. If required, the limit
299 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
300 time define to a desired maximum number of nodes or zero to allow
306 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
308 * Add FIPS provider configuration option to enforce the
309 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
310 The option '-ems-check' can optionally be supplied to
311 'openssl fipsinstall'.
315 * The FIPS provider includes a few non-approved algorithms for
316 backward compatibility purposes and the "fips=yes" property query
317 must be used for all algorithm fetches to ensure FIPS compliance.
319 The algorithms that are included but not approved are Triple DES ECB,
320 Triple DES CBC and EdDSA.
324 * Added support for KMAC in KBKDF.
328 * RNDR and RNDRRS support in provider functions to provide
329 random number generation for Arm CPUs (aarch64).
333 * s_client and s_server apps now explicitly say when the TLS version
334 does not include the renegotiation mechanism. This avoids confusion
335 between that scenario versus when the TLS version includes secure
336 renegotiation but the peer lacks support for it.
340 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
342 *Tomasz Kantecki, Andrey Matyukov*
344 * The various OBJ_* functions have been made thread safe.
348 * Parallel dual-prime 1536/2048-bit modular exponentiation for
349 AVX512_IFMA capable processors.
351 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
353 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
354 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
355 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
356 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
357 `OPENSSL_NO_DEPRECATED_3_1`.
359 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
360 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
361 definitions for these functions regardless of whether
362 `OPENSSL_NO_DEPRECATED_3_1` is defined.
364 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
365 functions regardless of whether they are using them. It is recommended that
366 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
370 * When generating safe-prime DH parameters set the recommended private key
371 length equivalent to minimum key lengths as in RFC 7919.
375 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
376 maximum size that is smaller or equal to the digest length to comply with
377 FIPS 186-4 section 5. This is implemented by a new option
378 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
379 `rsa_pss_saltlen` parameter, which is now the default. Signature
380 verification is not affected by this change and continues to work as before.
387 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
388 listed here are only a brief description.
389 The migration guide contains more detailed information related to new features,
390 breaking changes, and mappings for the large list of deprecated functions.
392 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
394 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
396 * Fixed NULL dereference during PKCS7 data verification.
398 A NULL pointer can be dereferenced when signatures are being
399 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
400 algorithm used for the signature is known to the OpenSSL library but
401 the implementation of the hash algorithm is not available the digest
402 initialization will fail. There is a missing check for the return
403 value from the initialization function which later leads to invalid
404 usage of the digest API most likely leading to a crash.
407 PKCS7 data is processed by the SMIME library calls and also by the
408 time stamp (TS) library calls. The TLS implementation in OpenSSL does
409 not call these functions however third party applications would be
410 affected if they call these functions to verify signatures on untrusted
415 * Fixed X.400 address type confusion in X.509 GeneralName.
417 There is a type confusion vulnerability relating to X.400 address processing
418 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
419 but the public structure definition for GENERAL_NAME incorrectly specified
420 the type of the x400Address field as ASN1_TYPE. This field is subsequently
421 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
424 When CRL checking is enabled (i.e. the application sets the
425 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
426 pass arbitrary pointers to a memcmp call, enabling them to read memory
427 contents or enact a denial of service.
432 * Fixed NULL dereference validating DSA public key.
434 An invalid pointer dereference on read can be triggered when an
435 application tries to check a malformed DSA public key by the
436 EVP_PKEY_public_check() function. This will most likely lead
437 to an application crash. This function can be called on public
438 keys supplied from untrusted sources which could allow an attacker
439 to cause a denial of service attack.
441 The TLS implementation in OpenSSL does not call this function
442 but applications might call the function if there are additional
443 security requirements imposed by standards such as FIPS 140-3.
446 *Shane Lontis, Tomáš Mráz*
448 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
450 An invalid pointer dereference on read can be triggered when an
451 application tries to load malformed PKCS7 data with the
452 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
454 The result of the dereference is an application crash which could
455 lead to a denial of service attack. The TLS implementation in OpenSSL
456 does not call this function however third party applications might
457 call these functions on untrusted data.
462 * Fixed Use-after-free following BIO_new_NDEF.
464 The public API function BIO_new_NDEF is a helper function used for
465 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
466 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
467 be called directly by end user applications.
469 The function receives a BIO from the caller, prepends a new BIO_f_asn1
470 filter BIO onto the front of it to form a BIO chain, and then returns
471 the new head of the BIO chain to the caller. Under certain conditions,
472 for example if a CMS recipient public key is invalid, the new filter BIO
473 is freed and the function returns a NULL result indicating a failure.
474 However, in this case, the BIO chain is not properly cleaned up and the
475 BIO passed by the caller still retains internal pointers to the previously
476 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
477 then a use-after-free will occur. This will most likely result in a crash.
480 *Viktor Dukhovni, Matt Caswell*
482 * Fixed Double free after calling PEM_read_bio_ex.
484 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
485 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
486 data. If the function succeeds then the "name_out", "header" and "data"
487 arguments are populated with pointers to buffers containing the relevant
488 decoded data. The caller is responsible for freeing those buffers. It is
489 possible to construct a PEM file that results in 0 bytes of payload data.
490 In this case PEM_read_bio_ex() will return a failure code but will populate
491 the header argument with a pointer to a buffer that has already been freed.
492 If the caller also frees this buffer then a double free will occur. This
493 will most likely lead to a crash.
495 The functions PEM_read_bio() and PEM_read() are simple wrappers around
496 PEM_read_bio_ex() and therefore these functions are also directly affected.
498 These functions are also called indirectly by a number of other OpenSSL
499 functions including PEM_X509_INFO_read_bio_ex() and
500 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
501 internal uses of these functions are not vulnerable because the caller does
502 not free the header argument if PEM_read_bio_ex() returns a failure code.
505 *Kurt Roeckx, Matt Caswell*
507 * Fixed Timing Oracle in RSA Decryption.
509 A timing based side channel exists in the OpenSSL RSA Decryption
510 implementation which could be sufficient to recover a plaintext across
511 a network in a Bleichenbacher style attack. To achieve a successful
512 decryption an attacker would have to be able to send a very large number
513 of trial messages for decryption. The vulnerability affects all RSA padding
514 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
517 *Dmitry Belyavsky, Hubert Kario*
519 * Fixed X.509 Name Constraints Read Buffer Overflow.
521 A read buffer overrun can be triggered in X.509 certificate verification,
522 specifically in name constraint checking. The read buffer overrun might
523 result in a crash which could lead to a denial of service attack.
524 In a TLS client, this can be triggered by connecting to a malicious
525 server. In a TLS server, this can be triggered if the server requests
526 client authentication and a malicious client connects.
531 * Fixed X.509 Policy Constraints Double Locking security issue.
533 If an X.509 certificate contains a malformed policy constraint and
534 policy processing is enabled, then a write lock will be taken twice
535 recursively. On some operating systems (most widely: Windows) this
536 results in a denial of service when the affected process hangs. Policy
537 processing being enabled on a publicly facing server is not considered
538 to be a common setup.
543 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
544 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
545 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
546 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
547 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
548 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
549 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
550 for legacy EC and SM2 keys is also changed similarly to honor the
551 equivalent conversion format flag as specified in the underlying
552 `EC_KEY` object being exported to a provider, when this function is
553 called through `EVP_PKEY_export()`.
557 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
559 * Fixed two buffer overflows in punycode decoding functions.
561 A buffer overrun can be triggered in X.509 certificate verification,
562 specifically in name constraint checking. Note that this occurs after
563 certificate chain signature verification and requires either a CA to
564 have signed the malicious certificate or for the application to continue
565 certificate verification despite failure to construct a path to a trusted
568 In a TLS client, this can be triggered by connecting to a malicious
569 server. In a TLS server, this can be triggered if the server requests
570 client authentication and a malicious client connects.
572 An attacker can craft a malicious email address to overflow
573 an arbitrary number of bytes containing the `.` character (decimal 46)
574 on the stack. This buffer overflow could result in a crash (causing a
578 An attacker can craft a malicious email address to overflow four
579 attacker-controlled bytes on the stack. This buffer overflow could
580 result in a crash (causing a denial of service) or potentially remote code
581 execution depending on stack layout for any given platform/compiler.
586 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
587 parameters in OpenSSL code.
588 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
589 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
590 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
591 Using these invalid names may cause algorithms to use slower methods
592 that ignore the CRT parameters.
596 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
601 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
602 data to be signed before signing the certificate.
606 * Added RIPEMD160 to the default provider.
610 * Ensured that the key share group sent or accepted for the key exchange
611 is allowed for the protocol version.
615 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
617 * OpenSSL supports creating a custom cipher via the legacy
618 EVP_CIPHER_meth_new() function and associated function calls. This function
619 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
620 to use the new provider mechanism in order to implement custom ciphers.
622 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
623 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
624 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
625 and decryption initialisation functions). Instead of using the custom cipher
626 directly it incorrectly tries to fetch an equivalent cipher from the
627 available providers. An equivalent cipher is found based on the NID passed to
628 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
629 given cipher. However it is possible for an application to incorrectly pass
630 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
631 is used in this way the OpenSSL encryption/decryption initialisation function
632 will match the NULL cipher as being equivalent and will fetch this from the
633 available providers. This will succeed if the default provider has been
634 loaded (or if a third party provider has been loaded that offers this
635 cipher). Using the NULL cipher means that the plaintext is emitted as the
638 Applications are only affected by this issue if they call
639 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
640 encryption/decryption initialisation function. Applications that only use
641 SSL/TLS are not impacted by this issue.
646 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
651 * Fixed the linux-mips64 Configure target which was missing the
652 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
657 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
662 * Correctly handle a retransmitted ClientHello in DTLS
666 * Fixed detection of ktls support in cross-compile environment on Linux
670 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
675 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
676 report correct results in some cases
680 * Fix UWP builds by defining VirtualLock
684 * For known safe primes use the minimum key length according to RFC 7919.
685 Longer private key sizes unnecessarily raise the cycles needed to compute the
686 shared secret without any increase of the real security. This fixes a
687 regression from 1.1.1 where these shorter keys were generated for the known
692 * Added the loongarch64 target
696 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
697 only passed to the FIPS provider and not to the default or legacy provider.
701 * Fixed reported performance degradation on aarch64. Restored the
702 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
703 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
704 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
705 The new algorithm is still used for 32 bit targets.
709 * Added a missing header for memcmp that caused compilation failure on some
714 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
716 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
717 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
718 This issue makes the RSA implementation with 2048 bit private keys
719 incorrect on such machines and memory corruption will happen during
720 the computation. As a consequence of the memory corruption an attacker
721 may be able to trigger a remote code execution on the machine performing
724 SSL/TLS servers or other servers using 2048 bit RSA private keys running
725 on machines supporting AVX512IFMA instructions of the X86_64 architecture
726 are affected by this issue.
731 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
732 implementation would not encrypt the entirety of the data under some
733 circumstances. This could reveal sixteen bytes of data that was
734 preexisting in the memory that wasn't written. In the special case of
735 "in place" encryption, sixteen bytes of the plaintext would be revealed.
737 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
738 they are both unaffected.
741 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
743 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
745 * In addition to the c_rehash shell command injection identified in
746 CVE-2022-1292, further bugs where the c_rehash script does not
747 properly sanitise shell metacharacters to prevent command injection have been
750 When the CVE-2022-1292 was fixed it was not discovered that there
751 are other places in the script where the file names of certificates
752 being hashed were possibly passed to a command executed through the shell.
754 This script is distributed by some operating systems in a manner where
755 it is automatically executed. On such operating systems, an attacker
756 could execute arbitrary commands with the privileges of the script.
758 Use of the c_rehash script is considered obsolete and should be replaced
759 by the OpenSSL rehash command line tool.
762 *Daniel Fiala, Tomáš Mráz*
764 * Case insensitive string comparison no longer uses locales. It has instead
765 been directly implemented.
769 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
771 * Case insensitive string comparison is reimplemented via new locale-agnostic
772 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
773 comparison. The previous implementation had problems when the Turkish locale
778 * Fixed a bug in the c_rehash script which was not properly sanitising shell
779 metacharacters to prevent command injection. This script is distributed by
780 some operating systems in a manner where it is automatically executed. On
781 such operating systems, an attacker could execute arbitrary commands with the
782 privileges of the script.
784 Use of the c_rehash script is considered obsolete and should be replaced
785 by the OpenSSL rehash command line tool.
790 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
791 certificate on an OCSP response. The bug caused the function in the case
792 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
793 response (meaning a successful verification) even in the case where the
794 response signing certificate fails to verify.
796 It is anticipated that most users of `OCSP_basic_verify` will not use the
797 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
798 a negative value (indicating a fatal error) in the case of a certificate
799 verification failure. The normal expected return value in this case would be
802 This issue also impacts the command line OpenSSL "ocsp" application. When
803 verifying an ocsp response with the "-no_cert_checks" option the command line
804 application will report that the verification is successful even though it
805 has in fact failed. In this case the incorrect successful response will also
806 be accompanied by error messages showing the failure and contradicting the
807 apparently successful result.
812 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
813 AAD data as the MAC key. This made the MAC key trivially predictable.
815 An attacker could exploit this issue by performing a man-in-the-middle attack
816 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
817 that the modified data would still pass the MAC integrity check.
819 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
820 endpoint will always be rejected by the recipient and the connection will
821 fail at that point. Many application protocols require data to be sent from
822 the client to the server first. Therefore, in such a case, only an OpenSSL
823 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
825 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
826 sent in both directions. In this case both clients and servers could be
827 affected, regardless of the application protocol.
829 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
830 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
831 the handshake when using this ciphersuite.
833 The confidentiality of data is not impacted by this issue, i.e. an attacker
834 cannot decrypt data that has been encrypted using this ciphersuite - they can
837 In order for this attack to work both endpoints must legitimately negotiate
838 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
839 OpenSSL 3.0, and is not available within the default provider or the default
840 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
841 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
842 following must have occurred:
844 1) OpenSSL must have been compiled with the (non-default) compile time option
845 enable-weak-ssl-ciphers
847 2) OpenSSL must have had the legacy provider explicitly loaded (either
848 through application code or via configuration)
850 3) The ciphersuite must have been explicitly added to the ciphersuite list
852 4) The libssl security level must have been set to 0 (default is 1)
854 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
856 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
857 others that both endpoints have in common
862 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
863 occupied by the removed hash table entries.
865 This function is used when decoding certificates or keys. If a long lived
866 process periodically decodes certificates or keys its memory usage will
867 expand without bounds and the process might be terminated by the operating
868 system causing a denial of service. Also traversing the empty hash table
869 entries will take increasingly more time.
871 Typically such long lived processes might be TLS clients or TLS servers
872 configured to accept client certificate authentication.
875 *Hugo Landau, Aliaksei Levin*
877 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
878 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
879 statistics are no longer supported. For compatibility, these statistics are
880 still listed in the output but are now always reported as zero.
884 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
886 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
887 for non-prime moduli.
889 Internally this function is used when parsing certificates that contain
890 elliptic curve public keys in compressed form or explicit elliptic curve
891 parameters with a base point encoded in compressed form.
893 It is possible to trigger the infinite loop by crafting a certificate that
894 has invalid explicit curve parameters.
896 Since certificate parsing happens prior to verification of the certificate
897 signature, any process that parses an externally supplied certificate may thus
898 be subject to a denial of service attack. The infinite loop can also be
899 reached when parsing crafted private keys as they can contain explicit
900 elliptic curve parameters.
902 Thus vulnerable situations include:
904 - TLS clients consuming server certificates
905 - TLS servers consuming client certificates
906 - Hosting providers taking certificates or private keys from customers
907 - Certificate authorities parsing certification requests from subscribers
908 - Anything else which parses ASN.1 elliptic curve parameters
910 Also any other applications that use the BN_mod_sqrt() where the attacker
911 can control the parameter values are vulnerable to this DoS issue.
916 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
917 to the list of ciphersuites providing Perfect Forward Secrecy as
918 required by SECLEVEL >= 3.
920 *Dmitry Belyavskiy, Nicola Tuveri*
922 * Made the AES constant time code for no-asm configurations
923 optional due to the resulting 95% performance degradation.
924 The AES constant time code can be enabled, for no assembly
925 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
929 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
934 * The negative return value handling of the certificate verification callback
935 was reverted. The replacement is to set the verification retry state with
936 the SSL_set_retry_verify() function.
940 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
942 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
943 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
944 verify a certificate supplied by a server. That function may return a
945 negative return value to indicate an internal error (for example out of
946 memory). Such a negative return value is mishandled by OpenSSL and will cause
947 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
948 success and a subsequent call to SSL_get_error() to return the value
949 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
950 returned by OpenSSL if the application has previously called
951 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
952 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
953 totally unexpected and applications may not behave correctly as a result. The
954 exact behaviour will depend on the application but it could result in
955 crashes, infinite loops or other similar incorrect responses.
957 This issue is made more serious in combination with a separate bug in OpenSSL
958 3.0 that will cause X509_verify_cert() to indicate an internal error when
959 processing a certificate chain. This will occur where a certificate does not
960 include the Subject Alternative Name extension but where a Certificate
961 Authority has enforced name constraints. This issue can occur even with valid
967 * Corrected a few file name and file reference bugs in the build,
968 installation and setup scripts, which lead to installation verification
969 failures. Slightly enhanced the installation verification script.
973 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
978 * Fixed PVK encoder to properly query for the passphrase.
982 * Multiple fixes in the OSSL_HTTP API functions.
986 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
987 OSSL_PARAM_INTEGER data type and return error on negative numbers
988 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
989 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
993 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
997 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
1001 * Multiple threading fixes.
1005 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
1009 * Allow fetching an operation from the provider that owns an unexportable key
1010 as a fallback if that is still allowed by the property query.
1014 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
1016 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
1021 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
1022 S390X capability vector to zero. This simplifies testing of different code
1023 paths on S390X architecture.
1027 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
1028 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
1029 SP 800-38D". The communication will fail at this point.
1033 * The EC_GROUP_clear_free() function is deprecated as there is nothing
1034 confidential in EC_GROUP data.
1038 * The byte order mark (BOM) character is ignored if encountered at the
1039 beginning of a PEM-formatted file.
1043 * Added CMS support for the Russian GOST algorithms.
1047 * Due to move of the implementation of cryptographic operations
1048 to the providers, validation of various operation parameters can
1049 be postponed until the actual operation is executed where previously
1050 it happened immediately when an operation parameter was set.
1052 For example when setting an unsupported curve with
1053 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
1054 fail but later keygen operations with the EVP_PKEY_CTX will fail.
1056 *OpenSSL team members and many third party contributors*
1058 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
1059 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
1060 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1061 instead to retrieve these algorithms from a provider.
1065 * On build targets where the multilib postfix is set in the build
1066 configuration the libdir directory was changing based on whether
1067 the lib directory with the multilib postfix exists on the system
1068 or not. This unpredictable behavior was removed and eventual
1069 multilib postfix is now always added to the default libdir. Use
1070 `--libdir=lib` to override the libdir if adding the postfix is
1075 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1076 no longer interoperable with OpenSSL 1.1.1.
1080 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1081 function codes, this function can only cause problems for calling
1086 * Add a configurable flag to output date formats as ISO 8601. Does not
1087 change the default date format.
1091 * Version of MSVC earlier than 1300 could get link warnings, which could
1092 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1093 Support for this flag has been removed.
1097 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1098 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1099 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1100 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1101 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1105 * The signatures of the functions to get and set options on SSL and
1106 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1107 Some source code changes may be required.
1111 * The public definitions of conf_method_st and conf_st have been
1112 deprecated. They will be made opaque in a future release.
1114 *Rich Salz and Tomáš Mráz*
1116 * Client-initiated renegotiation is disabled by default. To allow it, use
1117 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1118 flag, or the "ClientRenegotiation" config parameter as appropriate.
1122 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1123 or modify relative pathname inclusion.
1127 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1128 validated. Please consult the README-FIPS and
1129 README-PROVIDERS files, as well as the migration guide.
1131 *OpenSSL team members and many third party contributors*
1133 * For the key types DH and DHX the allowed settable parameters are now different.
1137 * The openssl commands that read keys, certificates, and CRLs now
1138 automatically detect the PEM or DER format of the input files.
1140 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1142 * Added enhanced PKCS#12 APIs which accept a library context.
1146 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1150 * Added support for Kernel TLS (KTLS).
1152 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1154 * Support for RFC 5746 secure renegotiation is now required by default for
1155 SSL or TLS connections to succeed.
1159 * The signature of the `copy` functional parameter of the
1160 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1161 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1162 the signature of the `pub_decode` functional parameter of the
1163 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1164 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1168 * The error return values from some control calls (ctrl) have changed.
1172 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1176 * Many functions in the EVP_ namespace that are getters of values from
1177 implementations or contexts were renamed to include get or get0 in their
1178 names. Old names are provided as macro aliases for compatibility and
1183 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1184 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1185 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1190 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1193 * The output from the command line applications may have minor
1198 * The output from numerous "printing" may have minor changes.
1202 * Windows thread synchronization uses read/write primitives (SRWLock) when
1203 supported by the OS, otherwise CriticalSection continues to be used.
1207 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1208 work on read only BIO source/sinks that do not support these functions.
1209 This allows piping or redirection of a file BIO using stdin to be buffered
1210 into memory. This is used internally in OSSL_DECODER_from_bio().
1214 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1215 this function would return one of the values OSSL_STORE_INFO_NAME,
1216 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1217 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1218 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1219 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1220 using this function should be amended to handle the changed return value.
1224 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1225 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1226 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1227 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1228 contains more than one certificate identifier: This means that all
1229 certificates referenced there MUST be part of the validation chain.
1233 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1234 RC5, DESX and DES have been moved to the legacy provider.
1238 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1239 RIPEMD-160 have been moved to the legacy provider.
1243 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1248 * The deprecated functions EVP_PKEY_get0_RSA(),
1249 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1250 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1251 well as the similarly named "get1" functions behave differently in
1256 * A number of functions handling low-level keys or engines were deprecated
1257 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1258 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1259 EVP_PKEY_get0_siphash().
1263 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1264 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1265 will need to load the legacy crypto provider. This includes these PBE
1266 algorithms which use this KDF:
1267 - NID_pbeWithMD2AndDES_CBC
1268 - NID_pbeWithMD5AndDES_CBC
1269 - NID_pbeWithSHA1AndRC2_CBC
1270 - NID_pbeWithMD2AndRC2_CBC
1271 - NID_pbeWithMD5AndRC2_CBC
1272 - NID_pbeWithSHA1AndDES_CBC
1276 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1277 BIO_debug_callback() functions.
1281 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1282 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1286 * The RAND_METHOD APIs have been deprecated.
1290 * The SRP APIs have been deprecated.
1294 * Add a compile time option to prevent the caching of provider fetched
1295 algorithms. This is enabled by including the no-cached-fetch option
1296 at configuration time.
1300 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1301 count of PKCS12_DEFAULT_ITER.
1303 *Tomáš Mráz and Sahana Prasad*
1305 * The openssl speed command does not use low-level API calls anymore.
1309 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1312 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1314 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1318 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1319 providers may supply their own group implementations (using either the "key
1320 exchange" or the "key encapsulation" methods) which will automatically be
1321 detected and used by libssl.
1323 *Matt Caswell, Nicola Tuveri*
1325 * The undocumented function X509_certificate_type() has been deprecated;
1329 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1333 * Removed RSA padding mode for SSLv23 (which was only used for
1334 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1335 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1340 * Deprecated the obsolete X9.31 RSA key generation related functions.
1342 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1343 is not allowed to return a value > 1, this is no more taken as failure.
1345 *Viktor Dukhovni and David von Oheimb*
1347 * Deprecated the obsolete X9.31 RSA key generation related functions
1348 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1349 BN_X931_generate_prime_ex().
1353 * The default key generation method for the regular 2-prime RSA keys was
1354 changed to the FIPS 186-4 B.3.6 method.
1358 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1362 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1366 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1367 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1369 *Rich Salz, Richard Levitte, and David von Oheimb*
1371 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1375 * Deprecated `OCSP_parse_url()`.
1379 * Validation of SM2 keys has been separated from the validation of regular EC
1384 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1385 switches: a validation failure triggers an early exit, returning a failure
1386 exit status to the parent process.
1390 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1391 to ignore unknown ciphers.
1395 * The `-cipher-commands` and `-digest-commands` options
1396 of the command line utility `list` have been deprecated.
1397 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1401 * Added convenience functions for generating asymmetric key pairs:
1402 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1403 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1407 * All of the low-level EC_KEY functions have been deprecated.
1409 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1411 * Deprecated all the libcrypto and libssl error string loading
1416 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1417 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1422 * The `-crypt` option to the `passwd` command line tool has been removed.
1426 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1431 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1435 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1436 EVP_PKEY_get1_tls_encodedpoint().
1440 * The security callback, which can be customised by application code, supports
1441 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1442 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1446 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1447 interface. Their functionality remains unchanged.
1451 * Added new option for 'openssl list', '-providers', which will display the
1452 list of loaded providers, their names, version and status. It optionally
1453 displays their gettable parameters.
1457 * Removed EVP_PKEY_set_alias_type().
1461 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1462 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1466 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1467 parameters are still checked for type safety at compile time via helper
1472 * Remove the RAND_DRBG API
1474 *Paul Dale and Matthias St. Pierre*
1476 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1477 as well as actual hostnames.
1481 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1482 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1483 conversely, silently ignore DTLS protocol version bounds when configuring
1484 TLS-based contexts. The commands can be repeated to set bounds of both
1485 types. The same applies with the corresponding "min_protocol" and
1486 "max_protocol" command-line switches, in case some application uses both TLS
1489 SSL_CTX instances that are created for a fixed protocol version (e.g.
1490 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1491 attempts to apply bounds to these protocol versions would result in an
1492 error. Now only the "version-flexible" SSL_CTX instances are subject to
1493 limits in configuration files in command-line options.
1497 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1502 * Reworked the recorded ERR codes to make better space for system errors.
1503 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1504 given code is a system error (true) or an OpenSSL error (false).
1508 * Reworked the test perl framework to better allow parallel testing.
1510 *Nicola Tuveri and David von Oheimb*
1512 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1513 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1517 * 'Configure' has been changed to figure out the configuration target if
1518 none is given on the command line. Consequently, the 'config' script is
1519 now only a mere wrapper. All documentation is changed to only mention
1522 *Rich Salz and Richard Levitte*
1524 * Added a library context `OSSL_LIB_CTX` that applications as well as
1525 other libraries can use to form a separate context within which
1526 libcrypto operations are performed.
1530 * Added various `_ex` functions to the OpenSSL API that support using
1531 a non-default `OSSL_LIB_CTX`.
1535 * Handshake now fails if Extended Master Secret extension is dropped
1540 * Dropped interactive mode from the `openssl` program.
1544 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1546 *David von Oheimb and Shane Lontis*
1548 * Deprecated `EC_METHOD_get_field_type()`.
1552 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1553 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1554 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1558 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1562 * Add CAdES-BES signature verification support, mostly derived
1563 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1565 *Filipe Raimundo da Silva*
1567 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1571 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1572 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1576 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1580 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1581 EC_KEY_precompute_mult().
1585 * Deprecated EC_POINTs_mul().
1589 * Removed FIPS_mode() and FIPS_mode_set().
1593 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1597 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1598 EC_POINT_get_Jprojective_coordinates_GFp().
1602 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1603 arrays to be more easily constructed via a series of utility functions.
1604 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1605 the various push functions and finally convert to a passable OSSL_PARAM
1606 array using OSSL_PARAM_BLD_to_param().
1610 * The security strength of SHA1 and MD5 based signatures in TLS has been
1615 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1616 contain a provider side internal key.
1620 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1624 * Project text documents not yet having a proper file name extension
1625 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1626 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1630 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1631 have been converted to Markdown with the goal to produce documents
1632 which not only look pretty when viewed online in the browser, but
1633 remain well readable inside a plain text editor.
1635 To achieve this goal, a 'minimalistic' Markdown style has been applied
1636 which avoids formatting elements that interfere too much with the
1637 reading flow in the text file. For example, it
1639 * avoids [ATX headings][] and uses [setext headings][] instead
1640 (which works for `<h1>` and `<h2>` headings only).
1641 * avoids [inline links][] and uses [reference links][] instead.
1642 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1644 [ATX headings]: https://github.github.com/gfm/#atx-headings
1645 [setext headings]: https://github.github.com/gfm/#setext-headings
1646 [inline links]: https://github.github.com/gfm/#inline-link
1647 [reference links]: https://github.github.com/gfm/#reference-link
1648 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1649 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1651 *Matthias St. Pierre*
1653 * The test suite is changed to preserve results of each test recipe.
1654 A new directory test-runs/ with subdirectories named like the
1655 test recipes are created in the build tree for this purpose.
1659 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1660 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1661 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1663 *David von Oheimb, Martin Peylo*
1665 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1666 It supports arbitrary request and response content types, GET redirection,
1667 TLS, connections via HTTP(S) proxies, connections and exchange via
1668 user-defined BIOs (allowing implicit connections), persistent connections,
1669 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1670 The legacy OCSP-focused (and only partly documented) API
1671 is retained for backward compatibility, while most of it is deprecated.
1675 * Added `util/check-format.pl`, a tool for checking adherence to the
1676 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1677 The checks performed are incomplete and yield some false positives.
1678 Still the tool should be useful for detecting most typical glitches.
1682 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1683 If domain name resolution yields multiple IP addresses all of them are tried
1684 after `connect()` failures.
1688 * All of the low-level RSA functions have been deprecated.
1692 * X509 certificates signed using SHA1 are no longer allowed at security
1697 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1698 modified to use PKEY APIs. These commands are now in maintenance mode
1699 and no new features will be added to them.
1703 * The command line utility rsautl has been deprecated.
1707 * The command line utilities genrsa and rsa have been modified to use PKEY
1708 APIs. They now write PKCS#8 keys by default. These commands are now in
1709 maintenance mode and no new features will be added to them.
1713 * All of the low-level DH functions have been deprecated.
1715 *Paul Dale and Matt Caswell*
1717 * All of the low-level DSA functions have been deprecated.
1721 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1722 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1726 * Deprecated low-level ECDH and ECDSA functions.
1730 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1734 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1735 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1736 a new formulation to include all the things it can be used for,
1737 as well as words of caution.
1741 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1745 * All of the low-level HMAC functions have been deprecated.
1747 *Paul Dale and David von Oheimb*
1749 * Over two thousand fixes were made to the documentation, including:
1750 - Common options (such as -rand/-writerand, TLS version control, etc)
1751 were refactored and point to newly-enhanced descriptions in openssl.pod.
1752 - Added style conformance for all options (with help from Richard Levitte),
1753 documented all reported missing options, added a CI build to check
1754 that all options are documented and that no unimplemented options
1756 - Documented some internals, such as all use of environment variables.
1757 - Addressed all internal broken L<> references.
1761 * All of the low-level CMAC functions have been deprecated.
1765 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1766 functions have been deprecated.
1768 *Paul Dale and David von Oheimb*
1770 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1771 set of functions. The documentation mentioned negative values for some
1772 errors, but this was never the case, so the mention of negative values
1775 Code that followed the documentation and thereby check with something
1776 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1780 * All of the low-level cipher functions have been deprecated.
1782 *Matt Caswell and Paul Dale*
1784 * Removed include/openssl/opensslconf.h.in and replaced it with
1785 include/openssl/configuration.h.in, which differs in not including
1786 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1787 was added to include both.
1789 This allows internal hacks where one might need to modify the set
1790 of configured macros, for example this if deprecated symbols are
1791 still supposed to be available internally:
1793 #include <openssl/configuration.h>
1795 #undef OPENSSL_NO_DEPRECATED
1796 #define OPENSSL_SUPPRESS_DEPRECATED
1798 #include <openssl/macros.h>
1800 This should not be used by applications that use the exported
1801 symbols, as that will lead to linking errors.
1805 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1806 used in exponentiation with 512-bit moduli. No EC algorithms are
1807 affected. Analysis suggests that attacks against 2-prime RSA1024,
1808 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1809 difficult to perform and are not believed likely. Attacks against DH512
1810 are considered just feasible. However, for an attack the target would
1811 have to re-use the DH512 private key, which is not recommended anyway.
1812 Also applications directly using the low-level API BN_mod_exp may be
1813 affected if they use BN_FLG_CONSTTIME.
1818 * Most memory-debug features have been deprecated, and the functionality
1819 replaced with no-ops.
1823 * Added documentation for the STACK API.
1827 * Introduced a new method type and API, OSSL_ENCODER, to represent
1828 generic encoders. These do the same sort of job that PEM writers
1829 and d2i functions do, but with support for methods supplied by
1830 providers, and the possibility for providers to support other
1835 * Introduced a new method type and API, OSSL_DECODER, to represent
1836 generic decoders. These do the same sort of job that PEM readers
1837 and i2d functions do, but with support for methods supplied by
1838 providers, and the possibility for providers to support other
1843 * Added a .pragma directive to the syntax of configuration files, to
1844 allow varying behavior in a supported and predictable manner.
1845 Currently added pragma:
1849 This allows dollar signs to be a keyword character unless it's
1850 followed by a opening brace or parenthesis. This is useful for
1851 platforms where dollar signs are commonly used in names, such as
1852 volume names and system directory names on VMS.
1856 * Added functionality to create an EVP_PKEY from user data.
1860 * Change the interpretation of the '--api' configuration option to
1861 mean that this is a desired API compatibility level with no
1862 further meaning. The previous interpretation, that this would
1863 also mean to remove all deprecated symbols up to and including
1864 the given version, no requires that 'no-deprecated' is also used
1865 in the configuration.
1867 When building applications, the desired API compatibility level
1868 can be set with the OPENSSL_API_COMPAT macro like before. For
1869 API compatibility version below 3.0, the old style numerical
1870 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1871 For version 3.0 and on, the value is expected to be the decimal
1872 value calculated from the major and minor version like this:
1874 MAJOR * 10000 + MINOR * 100
1878 -DOPENSSL_API_COMPAT=30000 For 3.0
1879 -DOPENSSL_API_COMPAT=30200 For 3.2
1881 To hide declarations that are deprecated up to and including the
1882 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1883 given when building the application as well.
1887 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1888 access to certificate and CRL stores via URIs and OSSL_STORE
1891 This adds the following functions:
1893 - X509_LOOKUP_store()
1894 - X509_STORE_load_file()
1895 - X509_STORE_load_path()
1896 - X509_STORE_load_store()
1897 - SSL_add_store_cert_subjects_to_stack()
1898 - SSL_CTX_set_default_verify_store()
1899 - SSL_CTX_load_verify_file()
1900 - SSL_CTX_load_verify_dir()
1901 - SSL_CTX_load_verify_store()
1905 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1906 The presence of this system service is determined at run-time.
1910 * Added functionality to create an EVP_PKEY context based on data
1911 for methods from providers. This takes an algorithm name and a
1912 property query string and simply stores them, with the intent
1913 that any operation that uses this context will use those strings
1914 to fetch the needed methods implicitly, thereby making the port
1915 of application written for pre-3.0 OpenSSL easier.
1919 * The undocumented function NCONF_WIN32() has been deprecated; for
1920 conversion details see the HISTORY section of doc/man5/config.pod
1924 * Introduced the new functions EVP_DigestSignInit_ex() and
1925 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1926 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1927 pages for further details.
1931 * Over two thousand fixes were made to the documentation, including:
1932 adding missing command flags, better style conformance, documentation
1935 *Rich Salz, Richard Levitte*
1937 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1938 X25519, X448, Ed25519 and Ed448.
1942 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1947 * Deprecated the public definition of `ERR_STATE` as well as the function
1948 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1953 * Added ERR functionality to give callers access to the stored function
1954 names that have replaced the older function code based functions.
1956 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1957 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1958 ERR_peek_error_all() and ERR_peek_last_error_all().
1960 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1961 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1962 ERR_func_error_string().
1966 * Extended testing to be verbose for failing tests only. The make variables
1967 VERBOSE_FAILURE or VF can be used to enable this:
1969 $ make VF=1 test # Unix
1970 $ mms /macro=(VF=1) test ! OpenVMS
1971 $ nmake VF=1 test # Windows
1975 * Added the `-copy_extensions` option to the `x509` command for use with
1976 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1977 all extensions in the request are copied to the certificate or vice versa.
1979 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1981 * Added the `-copy_extensions` option to the `req` command for use with
1982 `-x509`. When given with the `copy` or `copyall` argument,
1983 all extensions in the certification request are copied to the certificate.
1987 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1988 they generate are by default RFC 5280 compliant in the following sense:
1989 There is a subjectKeyIdentifier extension with a hash value of the public key
1990 and for not self-signed certs there is an authorityKeyIdentifier extension
1991 with a keyIdentifier field or issuer information identifying the signing key.
1992 This is done unless some configuration overrides the new default behavior,
1993 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1997 * Added several checks to `X509_verify_cert()` according to requirements in
1998 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1999 (which may be done by using the CLI option `-x509_strict`):
2000 * The basicConstraints of CA certificates must be marked critical.
2001 * CA certificates must explicitly include the keyUsage extension.
2002 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
2003 * The issuer name of any certificate must not be empty.
2004 * The subject name of CA certs, certs with keyUsage crlSign,
2005 and certs without subjectAlternativeName must not be empty.
2006 * If a subjectAlternativeName extension is given it must not be empty.
2007 * The signatureAlgorithm field and the cert signature must be consistent.
2008 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
2009 must not be marked critical.
2010 * The authorityKeyIdentifier must be given for X.509v3 certs
2011 unless they are self-signed.
2012 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
2016 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
2017 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
2021 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2022 used even when parsing explicit parameters, when loading a encoded key
2023 or calling `EC_GROUP_new_from_ecpkparameters()`/
2024 `EC_GROUP_new_from_ecparameters()`.
2025 This prevents bypass of security hardening and performance gains,
2026 especially for curves with specialized EC_METHODs.
2027 By default, if a key encoded with explicit parameters is loaded and later
2028 encoded, the output is still encoded with explicit parameters, even if
2029 internally a "named" EC_GROUP is used for computation.
2033 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2034 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2035 NULL. After this change, only the cofactor parameter can be NULL. It also
2036 does some minimal sanity checks on the passed order.
2041 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2042 An attack is simple, if the first CMS_recipientInfo is valid but the
2043 second CMS_recipientInfo is chosen ciphertext. If the second
2044 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2045 encryption key will be replaced by garbage, and the message cannot be
2046 decoded, but if the RSA decryption fails, the correct encryption key is
2047 used and the recipient will not notice the attack.
2048 As a work around for this potential attack the length of the decrypted
2049 key must be equal to the cipher default key length, in case the
2050 certificate is not given and all recipientInfo are tried out.
2051 The old behaviour can be re-enabled in the CMS code by setting the
2052 CMS_DEBUG_DECRYPT flag.
2056 * Early start up entropy quality from the DEVRANDOM seed source has been
2057 improved for older Linux systems. The RAND subsystem will wait for
2058 /dev/random to be producing output before seeding from /dev/urandom.
2059 The seeded state is stored for future library initialisations using
2060 a system global shared memory segment. The shared memory identifier
2061 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2062 the desired value. The default identifier is 114.
2066 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2067 when primes for RSA keys are computed.
2068 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2069 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2070 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2071 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2072 This avoids possible fingerprinting of newly generated RSA modules.
2076 * Correct the extended master secret constant on EBCDIC systems. Without this
2077 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2078 negotiate EMS will fail. Unfortunately this also means that TLS connections
2079 between EBCDIC systems with this fix, and EBCDIC systems without this
2080 fix will fail if they negotiate EMS.
2084 * Changed the library initialisation so that the config file is now loaded
2085 by default. This was already the case for libssl. It now occurs for both
2086 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2087 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2091 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2092 where the former acts as a replacement for `ERR_put_error()`, and the
2093 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2094 `ERR_raise_data()` adds more flexibility by taking a format string and
2095 an arbitrary number of arguments following it, to be processed with
2100 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2101 to check if a named provider is loaded and available. When called, it
2102 will also activate all fallback providers if such are still present.
2106 * Enforce a minimum DH modulus size of 512 bits.
2110 * Changed DH parameters to generate the order q subgroup instead of 2q.
2111 Previously generated DH parameters are still accepted by DH_check
2112 but DH_generate_key works around that by clearing bit 0 of the
2113 private key for those. This avoids leaking bit 0 of the private key.
2117 * Significantly reduce secure memory usage by the randomness pools.
2121 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2126 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2127 algorithms. An implementation of a key exchange algorithm can be obtained
2128 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2129 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2130 the older EVP_PKEY_derive_init() function. See the man pages for the new
2131 functions for further details.
2135 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2139 * Removed the function names from error messages and deprecated the
2144 * Removed NextStep support and the macro OPENSSL_UNISTD
2148 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2149 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2150 Also removed "export var as function" capability; we do not export
2151 variables, only functions.
2155 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2156 an error and 1 indicating success. In previous versions of OpenSSL this
2157 was a void type. If a key was set longer than the maximum possible this
2162 * Support SM2 signing and verification schemes with X509 certificate.
2166 * Use SHA256 as the default digest for TS query in the `ts` app.
2170 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2174 * Default cipher lists/suites are now available via a function, the
2175 #defines are deprecated.
2179 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2180 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2181 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2185 * Join the directories crypto/x509 and crypto/x509v3
2189 * Added command 'openssl kdf' that uses the EVP_KDF API.
2193 * Added command 'openssl mac' that uses the EVP_MAC API.
2197 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2198 as default directories. Also added the command 'openssl info'
2199 for scripting purposes.
2203 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2208 * Add prediction resistance to the DRBG reseeding process.
2212 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2213 mandated by IEEE Std 1619-2018.
2217 * Added newline escaping functionality to a filename when using openssl dgst.
2218 This output format is to replicate the output format found in the `*sum`
2219 checksum programs. This aims to preserve backward compatibility.
2221 *Matt Eaton, Richard Levitte, and Paul Dale*
2223 * Removed the heartbeat message in DTLS feature, as it has very
2224 little usage and doesn't seem to fulfill a valuable purpose.
2225 The configuration option is now deprecated.
2229 * Changed the output of 'openssl {digestname} < file' to display the
2230 digest name in its output.
2234 * Added a new generic trace API which provides support for enabling
2235 instrumentation through trace output.
2237 *Richard Levitte & Matthias St. Pierre*
2239 * Added build tests for C++. These are generated files that only do one
2240 thing, to include one public OpenSSL head file each. This tests that
2241 the public header files can be usefully included in a C++ application.
2243 This test isn't enabled by default. It can be enabled with the option
2244 'enable-buildtest-c++'.
2248 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2252 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2256 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2260 * Added KMAC to EVP_MAC.
2264 * Added property based algorithm implementation selection framework to
2269 * Added SCA hardening for modular field inversion in EC_GROUP through
2270 a new dedicated field_inv() pointer in EC_METHOD.
2271 This also addresses a leakage affecting conversions from projective
2272 to affine coordinates.
2274 *Billy Bob Brumley, Nicola Tuveri*
2276 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2277 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2278 those algorithms that were already supported through the EVP_PKEY API
2279 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2280 and scrypt are now wrappers that call EVP_KDF.
2284 * Build devcrypto engine as a dynamic engine.
2286 *Eneas U de Queiroz*
2288 * Add keyed BLAKE2 to EVP_MAC.
2292 * Fix a bug in the computation of the endpoint-pair shared secret used
2293 by DTLS over SCTP. This breaks interoperability with older versions
2294 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2295 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2296 interoperability with such broken implementations. However, enabling
2297 this switch breaks interoperability with correct implementations.
2299 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2300 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2304 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2308 * Changed the license to the Apache License v2.0.
2312 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2314 - Major releases (indicated by incrementing the MAJOR release number)
2315 may introduce incompatible API/ABI changes.
2316 - Minor releases (indicated by incrementing the MINOR release number)
2317 may introduce new features but retain API/ABI compatibility.
2318 - Patch releases (indicated by incrementing the PATCH number)
2319 are intended for bug fixes and other improvements of existing
2320 features only (like improving performance or adding documentation)
2321 and retain API/ABI compatibility.
2325 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2329 * Remove the 'dist' target and add a tarball building script. The
2330 'dist' target has fallen out of use, and it shouldn't be
2331 necessary to configure just to create a source distribution.
2335 * Recreate the OS390-Unix config target. It no longer relies on a
2336 special script like it did for OpenSSL pre-1.1.0.
2340 * Instead of having the source directories listed in Configure, add
2341 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2346 * Add GMAC to EVP_MAC.
2350 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2354 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2355 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2356 to facilitate the continued use of MACs through raw private keys in
2357 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2361 * Deprecate ECDH_KDF_X9_62().
2365 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2366 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2367 are retained for backwards compatibility.
2371 * AES-XTS mode now enforces that its two keys are different to mitigate
2372 the attacked described in "Efficient Instantiations of Tweakable
2373 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2374 Details of this attack can be obtained from:
2375 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2379 * Rename the object files, i.e. give them other names than in previous
2380 versions. Their names now include the name of the final product, as
2381 well as its type mnemonic (bin, lib, shlib).
2385 * Added new option for 'openssl list', '-objects', which will display the
2386 list of built in objects, i.e. OIDs with names.
2390 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2391 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2396 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2397 improves application performance by removing data copies and providing
2398 applications with zero-copy system calls such as sendfile and splice.
2402 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2406 * `PKCS12_parse` now maintains the order of the parsed certificates
2407 when outputting them via `*ca` (rather than reversing it).
2411 * Deprecated pthread fork support methods.
2415 * Added support for FFDHE key exchange in TLS 1.3.
2419 * Added a new concept for OpenSSL plugability: providers. This
2420 functionality is designed to replace the ENGINE API and ENGINE
2421 implementations, and to be much more dynamic, allowing provider
2422 authors to introduce new algorithms among other things, as long as
2423 there's an API that supports the algorithm type.
2425 With this concept comes a new core API for interaction between
2426 libcrypto and provider implementations. Public libcrypto functions
2427 that want to use providers do so through this core API.
2429 The main documentation for this core API is found in
2430 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2431 refer to other manuals describing the API specific for supported
2432 algorithm types (also called operations).
2439 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2441 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2443 * Avoid loading of a dynamic engine twice.
2447 * Prioritise DANE TLSA issuer certs over peer certs
2451 * Fixed random API for MacOS prior to 10.12
2453 These MacOS versions don't support the CommonCrypto APIs
2457 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2459 * Fixed an SM2 Decryption Buffer Overflow.
2461 In order to decrypt SM2 encrypted data an application is expected to
2462 call the API function EVP_PKEY_decrypt(). Typically an application will
2463 call this function twice. The first time, on entry, the "out" parameter
2464 can be NULL and, on exit, the "outlen" parameter is populated with the
2465 buffer size required to hold the decrypted plaintext. The application
2466 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2467 again, but this time passing a non-NULL value for the "out" parameter.
2469 A bug in the implementation of the SM2 decryption code means that the
2470 calculation of the buffer size required to hold the plaintext returned
2471 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2472 size required by the second call. This can lead to a buffer overflow
2473 when EVP_PKEY_decrypt() is called by the application a second time with
2474 a buffer that is too small.
2476 A malicious attacker who is able present SM2 content for decryption to
2477 an application could cause attacker chosen data to overflow the buffer
2478 by up to a maximum of 62 bytes altering the contents of other data held
2479 after the buffer, possibly changing application behaviour or causing
2480 the application to crash. The location of the buffer is application
2481 dependent but is typically heap allocated.
2486 * Fixed various read buffer overruns processing ASN.1 strings
2488 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2489 structure which contains a buffer holding the string data and a field
2490 holding the buffer length. This contrasts with normal C strings which
2491 are represented as a buffer for the string data which is terminated
2492 with a NUL (0) byte.
2494 Although not a strict requirement, ASN.1 strings that are parsed using
2495 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2496 well as any string whose value has been set with the ASN1_STRING_set()
2497 function will additionally NUL terminate the byte array in the
2498 ASN1_STRING structure.
2500 However, it is possible for applications to directly construct valid
2501 ASN1_STRING structures which do not NUL terminate the byte array by
2502 directly setting the "data" and "length" fields in the ASN1_STRING
2503 array. This can also happen by using the ASN1_STRING_set0() function.
2505 Numerous OpenSSL functions that print ASN.1 data have been found to
2506 assume that the ASN1_STRING byte array will be NUL terminated, even
2507 though this is not guaranteed for strings that have been directly
2508 constructed. Where an application requests an ASN.1 structure to be
2509 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2510 been directly constructed by the application without NUL terminating
2511 the "data" field, then a read buffer overrun can occur.
2513 The same thing can also occur during name constraints processing
2514 of certificates (for example if a certificate has been directly
2515 constructed by the application instead of loading it via the OpenSSL
2516 parsing functions, and the certificate contains non NUL terminated
2517 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2518 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2520 If a malicious actor can cause an application to directly construct an
2521 ASN1_STRING and then process it through one of the affected OpenSSL
2522 functions then this issue could be hit. This might result in a crash
2523 (causing a Denial of Service attack). It could also result in the
2524 disclosure of private memory contents (such as private keys, or
2525 sensitive plaintext).
2530 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2532 * Fixed a problem with verifying a certificate chain when using the
2533 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2534 the certificates present in a certificate chain. It is not set by default.
2536 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2537 the chain that have explicitly encoded elliptic curve parameters was added
2538 as an additional strict check.
2540 An error in the implementation of this check meant that the result of a
2541 previous check to confirm that certificates in the chain are valid CA
2542 certificates was overwritten. This effectively bypasses the check
2543 that non-CA certificates must not be able to issue other certificates.
2545 If a "purpose" has been configured then there is a subsequent opportunity
2546 for checks that the certificate is a valid CA. All of the named "purpose"
2547 values implemented in libcrypto perform this check. Therefore, where
2548 a purpose is set the certificate chain will still be rejected even when the
2549 strict flag has been used. A purpose is set by default in libssl client and
2550 server certificate verification routines, but it can be overridden or
2551 removed by an application.
2553 In order to be affected, an application must explicitly set the
2554 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2555 for the certificate verification or, in the case of TLS client or server
2556 applications, override the default purpose.
2561 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2562 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2563 renegotiation ClientHello omits the signature_algorithms extension (where it
2564 was present in the initial ClientHello), but includes a
2565 signature_algorithms_cert extension then a NULL pointer dereference will
2566 result, leading to a crash and a denial of service attack.
2568 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2569 (which is the default configuration). OpenSSL TLS clients are not impacted by
2573 *Peter Kästle and Samuel Sapalski*
2575 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2577 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2578 create a unique hash value based on the issuer and serial number data
2579 contained within an X509 certificate. However, it was failing to correctly
2580 handle any errors that may occur while parsing the issuer field (which might
2581 occur if the issuer field is maliciously constructed). This may subsequently
2582 result in a NULL pointer deref and a crash leading to a potential denial of
2588 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2589 padding mode to correctly check for rollback attacks. This is considered a
2590 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2595 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2596 functions. Previously they could overflow the output length argument in some
2597 cases where the input length is close to the maximum permissible length for
2598 an integer on the platform. In such cases the return value from the function
2599 call would be 1 (indicating success), but the output length value would be
2600 negative. This could cause applications to behave incorrectly or crash.
2605 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2606 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2607 could be exploited in a side channel attack to recover the password. Since
2608 the attack is local host only this is outside of the current OpenSSL
2609 threat model and therefore no CVE is assigned.
2611 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2616 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2618 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2619 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2620 If an attacker can control both items being compared then this could lead
2621 to a possible denial of service attack. OpenSSL itself uses the
2622 GENERAL_NAME_cmp function for two purposes:
2623 1) Comparing CRL distribution point names between an available CRL and a
2624 CRL distribution point embedded in an X509 certificate
2625 2) When verifying that a timestamp response token signer matches the
2626 timestamp authority name (exposed via the API functions
2627 TS_RESP_verify_response and TS_RESP_verify_token)
2632 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2634 * Certificates with explicit curve parameters are now disallowed in
2635 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2639 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2640 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2641 conversely, silently ignore DTLS protocol version bounds when configuring
2642 TLS-based contexts. The commands can be repeated to set bounds of both
2643 types. The same applies with the corresponding "min_protocol" and
2644 "max_protocol" command-line switches, in case some application uses both TLS
2647 SSL_CTX instances that are created for a fixed protocol version (e.g.
2648 TLSv1_server_method()) also silently ignore version bounds. Previously
2649 attempts to apply bounds to these protocol versions would result in an
2650 error. Now only the "version-flexible" SSL_CTX instances are subject to
2651 limits in configuration files in command-line options.
2655 * Handshake now fails if Extended Master Secret extension is dropped
2660 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2662 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2664 * Fixed segmentation fault in SSL_check_chain()
2665 Server or client applications that call the SSL_check_chain() function
2666 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2667 dereference as a result of incorrect handling of the
2668 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2669 or unrecognised signature algorithm is received from the peer. This could
2670 be exploited by a malicious peer in a Denial of Service attack.
2675 * Added AES consttime code for no-asm configurations
2676 an optional constant time support for AES was added
2677 when building openssl for no-asm.
2678 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2679 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2680 At this time this feature is by default disabled.
2681 It will be enabled by default in 3.0.
2685 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2687 * Revert the change of EOF detection while reading in libssl to avoid
2688 regressions in applications depending on the current way of reporting
2689 the EOF. As the existing method is not fully accurate the change to
2690 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2691 branch and will be present in the 3.0 release.
2695 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2696 when primes for RSA keys are computed.
2697 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2698 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2699 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2700 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2701 This avoids possible fingerprinting of newly generated RSA modules.
2705 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2707 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2708 while reading in libssl then we would report an error back to the
2709 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2710 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2711 therefore give a hint as to what went wrong.
2715 * Check that ed25519 and ed448 are allowed by the security level. Previously
2716 signature algorithms not using an MD were not being checked that they were
2717 allowed by the security level.
2721 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2722 was not quite right. The behaviour was not consistent between resumption
2723 and normal handshakes, and also not quite consistent with historical
2724 behaviour. The behaviour in various scenarios has been clarified and
2725 it has been updated to make it match historical behaviour as closely as
2730 * *[VMS only]* The header files that the VMS compilers include automatically,
2731 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2732 that the C++ compiler doesn't understand. This is a shortcoming in the
2733 compiler, but can be worked around with `__cplusplus` guards.
2735 C++ applications that use OpenSSL libraries must be compiled using the
2736 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2737 functions. Otherwise, only functions with symbols of less than 31
2738 characters can be used, as the linker will not be able to successfully
2739 resolve symbols with longer names.
2743 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2744 The presence of this system service is determined at run-time.
2748 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2753 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2755 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2756 number generator (RNG). This was intended to include protection in the
2757 event of a fork() system call in order to ensure that the parent and child
2758 processes did not share the same RNG state. However, this protection was not
2759 being used in the default case.
2761 A partial mitigation for this issue is that the output from a high
2762 precision timer is mixed into the RNG state so the likelihood of a parent
2763 and child process sharing state is significantly reduced.
2765 If an application already calls OPENSSL_init_crypto() explicitly using
2766 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2769 *Matthias St. Pierre*
2771 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2772 used even when parsing explicit parameters, when loading a encoded key
2773 or calling `EC_GROUP_new_from_ecpkparameters()`/
2774 `EC_GROUP_new_from_ecparameters()`.
2775 This prevents bypass of security hardening and performance gains,
2776 especially for curves with specialized EC_METHODs.
2777 By default, if a key encoded with explicit parameters is loaded and later
2778 encoded, the output is still encoded with explicit parameters, even if
2779 internally a "named" EC_GROUP is used for computation.
2783 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2784 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2785 NULL. After this change, only the cofactor parameter can be NULL. It also
2786 does some minimal sanity checks on the passed order.
2791 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2792 An attack is simple, if the first CMS_recipientInfo is valid but the
2793 second CMS_recipientInfo is chosen ciphertext. If the second
2794 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2795 encryption key will be replaced by garbage, and the message cannot be
2796 decoded, but if the RSA decryption fails, the correct encryption key is
2797 used and the recipient will not notice the attack.
2798 As a work around for this potential attack the length of the decrypted
2799 key must be equal to the cipher default key length, in case the
2800 certificate is not given and all recipientInfo are tried out.
2801 The old behaviour can be re-enabled in the CMS code by setting the
2802 CMS_DEBUG_DECRYPT flag.
2807 * Early start up entropy quality from the DEVRANDOM seed source has been
2808 improved for older Linux systems. The RAND subsystem will wait for
2809 /dev/random to be producing output before seeding from /dev/urandom.
2810 The seeded state is stored for future library initialisations using
2811 a system global shared memory segment. The shared memory identifier
2812 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2813 the desired value. The default identifier is 114.
2817 * Correct the extended master secret constant on EBCDIC systems. Without this
2818 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2819 negotiate EMS will fail. Unfortunately this also means that TLS connections
2820 between EBCDIC systems with this fix, and EBCDIC systems without this
2821 fix will fail if they negotiate EMS.
2825 * Use Windows installation paths in the mingw builds
2827 Mingw isn't a POSIX environment per se, which means that Windows
2828 paths should be used for installation.
2833 * Changed DH_check to accept parameters with order q and 2q subgroups.
2834 With order 2q subgroups the bit 0 of the private key is not secret
2835 but DH_generate_key works around that by clearing bit 0 of the
2836 private key for those. This avoids leaking bit 0 of the private key.
2840 * Significantly reduce secure memory usage by the randomness pools.
2844 * Revert the DEVRANDOM_WAIT feature for Linux systems
2846 The DEVRANDOM_WAIT feature added a select() call to wait for the
2847 /dev/random device to become readable before reading from the
2848 /dev/urandom device.
2850 It turned out that this change had negative side effects on
2851 performance which were not acceptable. After some discussion it
2852 was decided to revert this feature and leave it up to the OS
2853 resp. the platform maintainer to ensure a proper initialization
2854 during early boot time.
2856 *Matthias St. Pierre*
2858 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2860 * Add build tests for C++. These are generated files that only do one
2861 thing, to include one public OpenSSL head file each. This tests that
2862 the public header files can be usefully included in a C++ application.
2864 This test isn't enabled by default. It can be enabled with the option
2865 'enable-buildtest-c++'.
2869 * Enable SHA3 pre-hashing for ECDSA and DSA.
2873 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2874 This changes the size when using the `genpkey` command when no size is given.
2875 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2876 generation commands to use 2048 bits by default.
2880 * Reorganize the manual pages to consistently have RETURN VALUES,
2881 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2882 util/fix-doc-nits accordingly.
2884 *Paul Yang, Joshua Lock*
2886 * Add the missing accessor EVP_PKEY_get0_engine()
2890 * Have commands like `s_client` and `s_server` output the signature scheme
2891 along with other cipher suite parameters when debugging.
2895 * Make OPENSSL_config() error agnostic again.
2899 * Do the error handling in RSA decryption constant time.
2903 * Prevent over long nonces in ChaCha20-Poly1305.
2905 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2906 for every encryption operation. RFC 7539 specifies that the nonce value
2907 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2908 and front pads the nonce with 0 bytes if it is less than 12
2909 bytes. However it also incorrectly allows a nonce to be set of up to 16
2910 bytes. In this case only the last 12 bytes are significant and any
2911 additional leading bytes are ignored.
2913 It is a requirement of using this cipher that nonce values are
2914 unique. Messages encrypted using a reused nonce value are susceptible to
2915 serious confidentiality and integrity attacks. If an application changes
2916 the default nonce length to be longer than 12 bytes and then makes a
2917 change to the leading bytes of the nonce expecting the new value to be a
2918 new unique nonce then such an application could inadvertently encrypt
2919 messages with a reused nonce.
2921 Additionally the ignored bytes in a long nonce are not covered by the
2922 integrity guarantee of this cipher. Any application that relies on the
2923 integrity of these ignored leading bytes of a long nonce may be further
2924 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2925 is safe because no such use sets such a long nonce value. However user
2926 applications that use this cipher directly and set a non-default nonce
2927 length to be longer than 12 bytes may be vulnerable.
2929 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2935 * Add DEVRANDOM_WAIT feature for Linux systems
2937 On older Linux systems where the getrandom() system call is not available,
2938 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2939 Contrary to getrandom(), the /dev/urandom device will not block during
2940 early boot when the kernel CSPRNG has not been seeded yet.
2942 To mitigate this known weakness, use select() to wait for /dev/random to
2943 become readable before reading from /dev/urandom.
2945 * Ensure that SM2 only uses SM3 as digest algorithm
2949 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2951 * Change the info callback signals for the start and end of a post-handshake
2952 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2953 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2954 confused by this and assume that a TLSv1.2 renegotiation has started. This
2955 can break KeyUpdate handling. Instead we no longer signal the start and end
2956 of a post handshake message exchange (although the messages themselves are
2957 still signalled). This could break some applications that were expecting
2958 the old signals. However without this KeyUpdate is not usable for many
2963 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2965 * Timing vulnerability in DSA signature generation
2967 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2968 timing side channel attack. An attacker could use variations in the signing
2969 algorithm to recover the private key.
2971 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2976 * Timing vulnerability in ECDSA signature generation
2978 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2979 timing side channel attack. An attacker could use variations in the signing
2980 algorithm to recover the private key.
2982 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2987 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2988 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2989 of two gigabytes and the error handling improved.
2991 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2992 categorized as a normal bug, not a security issue, because the DRBG reseeds
2993 automatically and is fully functional even without additional randomness
2994 provided by the application.
2996 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2998 * Add a new ClientHello callback. Provides a callback interface that gives
2999 the application the ability to adjust the nascent SSL object at the
3000 earliest stage of ClientHello processing, immediately after extensions have
3001 been collected but before they have been processed. In particular, this
3002 callback can adjust the supported TLS versions in response to the contents
3007 * Add SM2 base algorithm support.
3011 * s390x assembly pack: add (improved) hardware-support for the following
3012 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
3013 aes-cfb/cfb8, aes-ecb.
3017 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3018 parameter is no longer accepted, as it leads to a corrupt table. NULL
3019 pem_str is reserved for alias entries only.
3023 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3024 step for prime curves. The new implementation is based on formulae from
3025 differential addition-and-doubling in homogeneous projective coordinates
3026 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
3027 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
3028 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
3029 to work in projective coordinates.
3031 *Billy Bob Brumley, Nicola Tuveri*
3033 * Change generating and checking of primes so that the error rate of not
3034 being prime depends on the intended use based on the size of the input.
3035 For larger primes this will result in more rounds of Miller-Rabin.
3036 The maximal error rate for primes with more than 1080 bits is lowered
3039 *Kurt Roeckx, Annie Yousar*
3041 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3045 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
3046 moving between systems, and to avoid confusion when a Windows build is
3047 done with mingw vs with MSVC. For POSIX installs, there's still a
3048 symlink or copy named 'tsget' to avoid that confusion as well.
3052 * Revert blinding in ECDSA sign and instead make problematic addition
3053 length-invariant. Switch even to fixed-length Montgomery multiplication.
3057 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3058 step for binary curves. The new implementation is based on formulae from
3059 differential addition-and-doubling in mixed Lopez-Dahab projective
3060 coordinates, modified to independently blind the operands.
3062 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3064 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3065 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3066 EC_METHODs to implement their own specialized "ladder step", to take
3067 advantage of more favorable coordinate systems or more efficient
3068 differential addition-and-doubling algorithms.
3070 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3072 * Modified the random device based seed sources to keep the relevant
3073 file descriptors open rather than reopening them on each access.
3074 This allows such sources to operate in a chroot() jail without
3075 the associated device nodes being available. This behaviour can be
3076 controlled using RAND_keep_random_devices_open().
3080 * Numerous side-channel attack mitigations have been applied. This may have
3081 performance impacts for some algorithms for the benefit of improved
3082 security. Specific changes are noted in this change log by their respective
3087 * AIX shared library support overhaul. Switch to AIX "natural" way of
3088 handling shared libraries, which means collecting shared objects of
3089 different versions and bitnesses in one common archive. This allows to
3090 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3091 doesn't affect the way 3rd party applications are linked, only how
3092 multi-version installation is managed.
3096 * Make ec_group_do_inverse_ord() more robust and available to other
3097 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3098 mitigations are applied to the fallback BN_mod_inverse().
3099 When using this function rather than BN_mod_inverse() directly, new
3100 EC cryptosystem implementations are then safer-by-default.
3104 * Add coordinate blinding for EC_POINT and implement projective
3105 coordinate blinding for generic prime curves as a countermeasure to
3106 chosen point SCA attacks.
3108 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3110 * Add blinding to ECDSA and DSA signatures to protect against side channel
3111 attacks discovered by Keegan Ryan (NCC Group).
3115 * Enforce checking in the `pkeyutl` command to ensure that the input
3116 length does not exceed the maximum supported digest length when performing
3117 a sign, verify or verifyrecover operation.
3121 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3122 I/O in combination with something like select() or poll() will hang. This
3123 can be turned off again using SSL_CTX_clear_mode().
3124 Many applications do not properly handle non-application data records, and
3125 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3126 around the problems in those applications, but can also break some.
3127 It's recommended to read the manpages about SSL_read(), SSL_write(),
3128 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3129 SSL_CTX_set_read_ahead() again.
3133 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3134 now allow empty (zero character) pass phrases.
3138 * Apply blinding to binary field modular inversion and remove patent
3139 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3143 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3144 binary and prime elliptic curves.
3148 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3149 constant time fixed point multiplication.
3153 * Revise elliptic curve scalar multiplication with timing attack
3154 defenses: ec_wNAF_mul redirects to a constant time implementation
3155 when computing fixed point and variable point multiplication (which
3156 in OpenSSL are mostly used with secret scalars in keygen, sign,
3157 ECDH derive operations).
3158 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3161 * Updated CONTRIBUTING
3165 * Updated DRBG / RAND to request nonce and additional low entropy
3166 randomness from the system.
3168 *Matthias St. Pierre*
3170 * Updated 'openssl rehash' to use OpenSSL consistent default.
3174 * Moved the load of the ssl_conf module to libcrypto, which helps
3175 loading engines that libssl uses before libssl is initialised.
3179 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3183 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3185 *Ingo Schwarze, Rich Salz*
3187 * Added output of accepting IP address and port for 'openssl s_server'
3191 * Added a new API for TLSv1.3 ciphersuites:
3192 SSL_CTX_set_ciphersuites()
3193 SSL_set_ciphersuites()
3197 * Memory allocation failures consistently add an error to the error
3202 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3203 in libcrypto when run as setuid/setgid.
3207 * Load any config file by default when libssl is used.
3211 * Added new public header file <openssl/rand_drbg.h> and documentation
3212 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3214 *Matthias St. Pierre*
3216 * QNX support removed (cannot find contributors to get their approval
3217 for the license change).
3221 * TLSv1.3 replay protection for early data has been implemented. See the
3222 SSL_read_early_data() man page for further details.
3226 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3227 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3228 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3229 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3230 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3231 configuration has been separated out. See the ciphers man page or the
3232 SSL_CTX_set_ciphersuites() man page for more information.
3236 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3237 in responder mode now supports the new "-multi" option, which
3238 spawns the specified number of child processes to handle OCSP
3239 requests. The "-timeout" option now also limits the OCSP
3240 responder's patience to wait to receive the full client request
3241 on a newly accepted connection. Child processes are respawned
3242 as needed, and the CA index file is automatically reloaded
3243 when changed. This makes it possible to run the "ocsp" responder
3244 as a long-running service, making the OpenSSL CA somewhat more
3245 feature-complete. In this mode, most diagnostic messages logged
3246 after entering the event loop are logged via syslog(3) rather than
3251 * Added support for X448 and Ed448. Heavily based on original work by
3256 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3257 objects loaded. This adds the functions OSSL_STORE_expect() and
3258 OSSL_STORE_find() as well as needed tools to construct searches and
3259 get the search data out of them.
3263 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3264 version of OpenSSL should review their configuration settings to ensure
3265 that they are still appropriate for TLSv1.3. For further information see:
3266 <https://wiki.openssl.org/index.php/TLS1.3>
3270 * Grand redesign of the OpenSSL random generator
3272 The default RAND method now utilizes an AES-CTR DRBG according to
3273 NIST standard SP 800-90Ar1. The new random generator is essentially
3274 a port of the default random generator from the OpenSSL FIPS 2.0
3275 object module. It is a hybrid deterministic random bit generator
3276 using an AES-CTR bit stream and which seeds and reseeds itself
3277 automatically using trusted system entropy sources.
3279 Some of its new features are:
3280 - Support for multiple DRBG instances with seed chaining.
3281 - The default RAND method makes use of a DRBG.
3282 - There is a public and private DRBG instance.
3283 - The DRBG instances are fork-safe.
3284 - Keep all global DRBG instances on the secure heap if it is enabled.
3285 - The public and private DRBG instance are per thread for lock free
3288 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3290 * Changed Configure so it only says what it does and doesn't dump
3291 so much data. Instead, ./configdata.pm should be used as a script
3292 to display all sorts of configuration data.
3296 * Added processing of "make variables" to Configure.
3300 * Added SHA512/224 and SHA512/256 algorithm support.
3304 * The last traces of Netware support, first removed in 1.1.0, have
3309 * Get rid of Makefile.shared, and in the process, make the processing
3310 of certain files (rc.obj, or the .def/.map/.opt files produced from
3311 the ordinal files) more visible and hopefully easier to trace and
3312 debug (or make silent).
3316 * Make it possible to have environment variable assignments as
3317 arguments to config / Configure.
3321 * Add multi-prime RSA (RFC 8017) support.
3325 * Add SM3 implemented according to GB/T 32905-2016
3326 *Jack Lloyd <jack.lloyd@ribose.com>,*
3327 *Ronald Tse <ronald.tse@ribose.com>,*
3328 *Erick Borsboom <erick.borsboom@ribose.com>*
3330 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3331 as documented in RFC6066.
3332 Based on a patch from Tomasz Moń
3334 *Filipe Raimundo da Silva*
3336 * Add SM4 implemented according to GB/T 32907-2016.
3337 *Jack Lloyd <jack.lloyd@ribose.com>,*
3338 *Ronald Tse <ronald.tse@ribose.com>,*
3339 *Erick Borsboom <erick.borsboom@ribose.com>*
3341 * Reimplement -newreq-nodes and ERR_error_string_n; the
3342 original author does not agree with the license change.
3346 * Add ARIA AEAD TLS support.
3350 * Some macro definitions to support VS6 have been removed. Visual
3351 Studio 6 has not worked since 1.1.0
3355 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3356 without clearing the errors.
3360 * Add "atfork" functions. If building on a system that without
3361 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3362 requirements. The RAND facility now uses/requires this.
3370 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3371 not possible to disable entirely. However, it's still possible to
3372 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3375 To disable, configure with 'no-ui-console'. 'no-ui' is still
3376 possible to use as an alias. Check at compile time with the
3377 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3378 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3382 * Add a STORE module, which implements a uniform and URI based reader of
3383 stores that can contain keys, certificates, CRLs and numerous other
3384 objects. The main API is loosely based on a few stdio functions,
3385 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3386 OSSL_STORE_error and OSSL_STORE_close.
3387 The implementation uses backends called "loaders" to implement arbitrary
3388 URI schemes. There is one built in "loader" for the 'file' scheme.
3392 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3393 then adjusted to work on FreeBSD 8.4 as well.
3394 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3395 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3399 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3400 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3401 error code calls like this:
3403 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3405 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3406 that can be encoded in C. For the foreseeable future, this will only
3409 *Richard Levitte and Tim Hudson*
3411 * Removed BSD cryptodev engine.
3415 * Add a build target 'build_all_generated', to build all generated files
3416 and only that. This can be used to prepare everything that requires
3417 things like perl for a system that lacks perl and then move everything
3418 to that system and do the rest of the build there.
3422 * In the UI interface, make it possible to duplicate the user data. This
3423 can be used by engines that need to retain the data for a longer time
3424 than just the call where this user data is passed.
3428 * Ignore the '-named_curve auto' value for compatibility of applications
3431 *Tomáš Mráz <tmraz@fedoraproject.org>*
3433 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3434 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3435 alerts across multiple records (some of which could be empty). In practice
3436 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3437 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3438 support this at all. Supporting it adds significant complexity to the
3439 record layer, and its removal is unlikely to cause interoperability
3444 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3445 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3446 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3451 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3452 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3454 *Richard Levitte, Andy Polyakov*
3456 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3461 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3462 platform rather than 'mingw'.
3466 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3467 success if they are asked to add an object which already exists
3468 in the store. This change cascades to other functions which load
3469 certificates and CRLs.
3473 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3474 facilitate stack unwinding even from assembly subroutines.
3478 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3479 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3483 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3484 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3485 which is the minimum version we support.
3489 * Certificate time validation (X509_cmp_time) enforces stricter
3490 compliance with RFC 5280. Fractional seconds and timezone offsets
3491 are no longer allowed.
3495 * Add support for ARIA
3499 * s_client will now send the Server Name Indication (SNI) extension by
3500 default unless the new "-noservername" option is used. The server name is
3501 based on the host provided to the "-connect" option unless overridden by
3502 using "-servername".
3506 * Add support for SipHash
3510 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3511 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3512 prevent issues where no progress is being made and the peer continually
3513 sends unrecognised record types, using up resources processing them.
3517 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3518 using the algorithm defined in
3519 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3523 * Heartbeat support has been removed; the ABI is changed for now.
3525 *Richard Levitte, Rich Salz*
3527 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3531 * The RSA "null" method, which was partially supported to avoid patent
3532 issues, has been replaced to always returns NULL.
3539 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3541 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3542 used even when parsing explicit parameters, when loading a encoded key
3543 or calling `EC_GROUP_new_from_ecpkparameters()`/
3544 `EC_GROUP_new_from_ecparameters()`.
3545 This prevents bypass of security hardening and performance gains,
3546 especially for curves with specialized EC_METHODs.
3547 By default, if a key encoded with explicit parameters is loaded and later
3548 encoded, the output is still encoded with explicit parameters, even if
3549 internally a "named" EC_GROUP is used for computation.
3553 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3554 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3555 NULL. After this change, only the cofactor parameter can be NULL. It also
3556 does some minimal sanity checks on the passed order.
3561 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3562 An attack is simple, if the first CMS_recipientInfo is valid but the
3563 second CMS_recipientInfo is chosen ciphertext. If the second
3564 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3565 encryption key will be replaced by garbage, and the message cannot be
3566 decoded, but if the RSA decryption fails, the correct encryption key is
3567 used and the recipient will not notice the attack.
3568 As a work around for this potential attack the length of the decrypted
3569 key must be equal to the cipher default key length, in case the
3570 certificate is not given and all recipientInfo are tried out.
3571 The old behaviour can be re-enabled in the CMS code by setting the
3572 CMS_DEBUG_DECRYPT flag.
3577 * Use Windows installation paths in the mingw builds
3579 Mingw isn't a POSIX environment per se, which means that Windows
3580 paths should be used for installation.
3585 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3587 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3588 This changes the size when using the `genpkey` command when no size is given.
3589 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3590 generation commands to use 2048 bits by default.
3594 * Prevent over long nonces in ChaCha20-Poly1305.
3596 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3597 for every encryption operation. RFC 7539 specifies that the nonce value
3598 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3599 and front pads the nonce with 0 bytes if it is less than 12
3600 bytes. However it also incorrectly allows a nonce to be set of up to 16
3601 bytes. In this case only the last 12 bytes are significant and any
3602 additional leading bytes are ignored.
3604 It is a requirement of using this cipher that nonce values are
3605 unique. Messages encrypted using a reused nonce value are susceptible to
3606 serious confidentiality and integrity attacks. If an application changes
3607 the default nonce length to be longer than 12 bytes and then makes a
3608 change to the leading bytes of the nonce expecting the new value to be a
3609 new unique nonce then such an application could inadvertently encrypt
3610 messages with a reused nonce.
3612 Additionally the ignored bytes in a long nonce are not covered by the
3613 integrity guarantee of this cipher. Any application that relies on the
3614 integrity of these ignored leading bytes of a long nonce may be further
3615 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3616 is safe because no such use sets such a long nonce value. However user
3617 applications that use this cipher directly and set a non-default nonce
3618 length to be longer than 12 bytes may be vulnerable.
3620 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3626 * Added SCA hardening for modular field inversion in EC_GROUP through
3627 a new dedicated field_inv() pointer in EC_METHOD.
3628 This also addresses a leakage affecting conversions from projective
3629 to affine coordinates.
3631 *Billy Bob Brumley, Nicola Tuveri*
3633 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3634 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3638 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3642 * Remove the 'dist' target and add a tarball building script. The
3643 'dist' target has fallen out of use, and it shouldn't be
3644 necessary to configure just to create a source distribution.
3648 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3650 * Timing vulnerability in DSA signature generation
3652 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3653 timing side channel attack. An attacker could use variations in the signing
3654 algorithm to recover the private key.
3656 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3661 * Timing vulnerability in ECDSA signature generation
3663 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3664 timing side channel attack. An attacker could use variations in the signing
3665 algorithm to recover the private key.
3667 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3672 * Add coordinate blinding for EC_POINT and implement projective
3673 coordinate blinding for generic prime curves as a countermeasure to
3674 chosen point SCA attacks.
3676 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3678 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3680 * Client DoS due to large DH parameter
3682 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3683 malicious server can send a very large prime value to the client. This will
3684 cause the client to spend an unreasonably long period of time generating a
3685 key for this prime resulting in a hang until the client has finished. This
3686 could be exploited in a Denial Of Service attack.
3688 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3693 * Cache timing vulnerability in RSA Key Generation
3695 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3696 a cache timing side channel attack. An attacker with sufficient access to
3697 mount cache timing attacks during the RSA key generation process could
3698 recover the private key.
3700 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3701 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3706 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3707 parameter is no longer accepted, as it leads to a corrupt table. NULL
3708 pem_str is reserved for alias entries only.
3712 * Revert blinding in ECDSA sign and instead make problematic addition
3713 length-invariant. Switch even to fixed-length Montgomery multiplication.
3717 * Change generating and checking of primes so that the error rate of not
3718 being prime depends on the intended use based on the size of the input.
3719 For larger primes this will result in more rounds of Miller-Rabin.
3720 The maximal error rate for primes with more than 1080 bits is lowered
3723 *Kurt Roeckx, Annie Yousar*
3725 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3729 * Add blinding to ECDSA and DSA signatures to protect against side channel
3730 attacks discovered by Keegan Ryan (NCC Group).
3734 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3735 now allow empty (zero character) pass phrases.
3739 * Certificate time validation (X509_cmp_time) enforces stricter
3740 compliance with RFC 5280. Fractional seconds and timezone offsets
3741 are no longer allowed.
3745 * Fixed a text canonicalisation bug in CMS
3747 Where a CMS detached signature is used with text content the text goes
3748 through a canonicalisation process first prior to signing or verifying a
3749 signature. This process strips trailing space at the end of lines, converts
3750 line terminators to CRLF and removes additional trailing line terminators
3751 at the end of a file. A bug in the canonicalisation process meant that
3752 some characters, such as form-feed, were incorrectly treated as whitespace
3753 and removed. This is contrary to the specification (RFC5485). This fix
3754 could mean that detached text data signed with an earlier version of
3755 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3756 signed with a fixed OpenSSL may fail to verify with an earlier version of
3757 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3758 and use the "-binary" flag (for the "cms" command line application) or set
3759 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3763 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3765 * Constructed ASN.1 types with a recursive definition could exceed the stack
3767 Constructed ASN.1 types with a recursive definition (such as can be found
3768 in PKCS7) could eventually exceed the stack given malicious input with
3769 excessive recursion. This could result in a Denial Of Service attack. There
3770 are no such structures used within SSL/TLS that come from untrusted sources
3771 so this is considered safe.
3773 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3779 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3781 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3782 effectively reduced to only comparing the least significant bit of each
3783 byte. This allows an attacker to forge messages that would be considered as
3784 authenticated in an amount of tries lower than that guaranteed by the
3785 security claims of the scheme. The module can only be compiled by the
3786 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3788 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3794 * Add a build target 'build_all_generated', to build all generated files
3795 and only that. This can be used to prepare everything that requires
3796 things like perl for a system that lacks perl and then move everything
3797 to that system and do the rest of the build there.
3801 * Backport SSL_OP_NO_RENGOTIATION
3803 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3804 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3805 changes this is no longer possible in 1.1.0. Therefore, the new
3806 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3807 1.1.0 to provide equivalent functionality.
3809 Note that if an application built against 1.1.0h headers (or above) is run
3810 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3811 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3815 * Removed the OS390-Unix config target. It relied on a script that doesn't
3820 * rsaz_1024_mul_avx2 overflow bug on x86_64
3822 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3823 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3824 Analysis suggests that attacks against RSA and DSA as a result of this
3825 defect would be very difficult to perform and are not believed likely.
3826 Attacks against DH1024 are considered just feasible, because most of the
3827 work necessary to deduce information about a private key may be performed
3828 offline. The amount of resources required for such an attack would be
3829 significant. However, for an attack on TLS to be meaningful, the server
3830 would have to share the DH1024 private key among multiple clients, which is
3831 no longer an option since CVE-2016-0701.
3833 This only affects processors that support the AVX2 but not ADX extensions
3834 like Intel Haswell (4th generation).
3836 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3837 was originally found via the OSS-Fuzz project.
3842 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3844 * bn_sqrx8x_internal carry bug on x86_64
3846 There is a carry propagating bug in the x86_64 Montgomery squaring
3847 procedure. No EC algorithms are affected. Analysis suggests that attacks
3848 against RSA and DSA as a result of this defect would be very difficult to
3849 perform and are not believed likely. Attacks against DH are considered just
3850 feasible (although very difficult) because most of the work necessary to
3851 deduce information about a private key may be performed offline. The amount
3852 of resources required for such an attack would be very significant and
3853 likely only accessible to a limited number of attackers. An attacker would
3854 additionally need online access to an unpatched system using the target
3855 private key in a scenario with persistent DH parameters and a private
3856 key that is shared between multiple clients.
3858 This only affects processors that support the BMI1, BMI2 and ADX extensions
3859 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3861 This issue was reported to OpenSSL by the OSS-Fuzz project.
3866 * Malformed X.509 IPAddressFamily could cause OOB read
3868 If an X.509 certificate has a malformed IPAddressFamily extension,
3869 OpenSSL could do a one-byte buffer overread. The most likely result
3870 would be an erroneous display of the certificate in text format.
3872 This issue was reported to OpenSSL by the OSS-Fuzz project.
3877 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3879 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3880 platform rather than 'mingw'.
3884 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3885 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3886 which is the minimum version we support.
3890 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3892 * Encrypt-Then-Mac renegotiation crash
3894 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3895 negotiated where it was not in the original handshake (or vice-versa) then
3896 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3897 and servers are affected.
3899 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3904 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3906 * Truncated packet could crash via OOB read
3908 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3909 cipher is being used, then a truncated packet can cause that host to
3910 perform an out-of-bounds read, usually resulting in a crash.
3912 This issue was reported to OpenSSL by Robert Święcki of Google.
3917 * Bad (EC)DHE parameters cause a client crash
3919 If a malicious server supplies bad parameters for a DHE or ECDHE key
3920 exchange then this can result in the client attempting to dereference a
3921 NULL pointer leading to a client crash. This could be exploited in a Denial
3924 This issue was reported to OpenSSL by Guido Vranken.
3929 * BN_mod_exp may produce incorrect results on x86_64
3931 There is a carry propagating bug in the x86_64 Montgomery squaring
3932 procedure. No EC algorithms are affected. Analysis suggests that attacks
3933 against RSA and DSA as a result of this defect would be very difficult to
3934 perform and are not believed likely. Attacks against DH are considered just
3935 feasible (although very difficult) because most of the work necessary to
3936 deduce information about a private key may be performed offline. The amount
3937 of resources required for such an attack would be very significant and
3938 likely only accessible to a limited number of attackers. An attacker would
3939 additionally need online access to an unpatched system using the target
3940 private key in a scenario with persistent DH parameters and a private
3941 key that is shared between multiple clients. For example this can occur by
3942 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3943 similar to CVE-2015-3193 but must be treated as a separate problem.
3945 This issue was reported to OpenSSL by the OSS-Fuzz project.
3950 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3952 * ChaCha20/Poly1305 heap-buffer-overflow
3954 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3955 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3956 crash. This issue is not considered to be exploitable beyond a DoS.
3958 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3963 * CMS Null dereference
3965 Applications parsing invalid CMS structures can crash with a NULL pointer
3966 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3967 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3968 structure callback if an attempt is made to free certain invalid encodings.
3969 Only CHOICE structures using a callback which do not handle NULL value are
3972 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3977 * Montgomery multiplication may produce incorrect results
3979 There is a carry propagating bug in the Broadwell-specific Montgomery
3980 multiplication procedure that handles input lengths divisible by, but
3981 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3982 and DH private keys are impossible. This is because the subroutine in
3983 question is not used in operations with the private key itself and an input
3984 of the attacker's direct choice. Otherwise the bug can manifest itself as
3985 transient authentication and key negotiation failures or reproducible
3986 erroneous outcome of public-key operations with specially crafted input.
3987 Among EC algorithms only Brainpool P-512 curves are affected and one
3988 presumably can attack ECDH key negotiation. Impact was not analyzed in
3989 detail, because pre-requisites for attack are considered unlikely. Namely
3990 multiple clients have to choose the curve in question and the server has to
3991 share the private key among them, neither of which is default behaviour.
3992 Even then only clients that chose the curve will be affected.
3994 This issue was publicly reported as transient failures and was not
3995 initially recognized as a security issue. Thanks to Richard Morgan for
3996 providing reproducible case.
4001 * Removed automatic addition of RPATH in shared libraries and executables,
4002 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
4006 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
4008 * Fix Use After Free for large message sizes
4010 The patch applied to address CVE-2016-6307 resulted in an issue where if a
4011 message larger than approx 16k is received then the underlying buffer to
4012 store the incoming message is reallocated and moved. Unfortunately a
4013 dangling pointer to the old location is left which results in an attempt to
4014 write to the previously freed location. This is likely to result in a
4015 crash, however it could potentially lead to execution of arbitrary code.
4017 This issue only affects OpenSSL 1.1.0a.
4019 This issue was reported to OpenSSL by Robert Święcki.
4024 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
4026 * OCSP Status Request extension unbounded memory growth
4028 A malicious client can send an excessively large OCSP Status Request
4029 extension. If that client continually requests renegotiation, sending a
4030 large OCSP Status Request extension each time, then there will be unbounded
4031 memory growth on the server. This will eventually lead to a Denial Of
4032 Service attack through memory exhaustion. Servers with a default
4033 configuration are vulnerable even if they do not support OCSP. Builds using
4034 the "no-ocsp" build time option are not affected.
4036 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4041 * SSL_peek() hang on empty record
4043 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
4044 sends an empty record. This could be exploited by a malicious peer in a
4045 Denial Of Service attack.
4047 This issue was reported to OpenSSL by Alex Gaynor.
4052 * Excessive allocation of memory in tls_get_message_header() and
4053 dtls1_preprocess_fragment()
4055 A (D)TLS message includes 3 bytes for its length in the header for the
4056 message. This would allow for messages up to 16Mb in length. Messages of
4057 this length are excessive and OpenSSL includes a check to ensure that a
4058 peer is sending reasonably sized messages in order to avoid too much memory
4059 being consumed to service a connection. A flaw in the logic of version
4060 1.1.0 means that memory for the message is allocated too early, prior to
4061 the excessive message length check. Due to way memory is allocated in
4062 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4063 to service a connection. This could lead to a Denial of Service through
4064 memory exhaustion. However, the excessive message length check still takes
4065 place, and this would cause the connection to immediately fail. Assuming
4066 that the application calls SSL_free() on the failed connection in a timely
4067 manner then the 21Mb of allocated memory will then be immediately freed
4068 again. Therefore, the excessive memory allocation will be transitory in
4069 nature. This then means that there is only a security impact if:
4071 1) The application does not call SSL_free() in a timely manner in the event
4072 that the connection fails
4074 2) The application is working in a constrained environment where there is
4075 very little free memory
4077 3) The attacker initiates multiple connection attempts such that there are
4078 multiple connections in a state where memory has been allocated for the
4079 connection; SSL_free() has not yet been called; and there is insufficient
4080 memory to service the multiple requests.
4082 Except in the instance of (1) above any Denial Of Service is likely to be
4083 transitory because as soon as the connection fails the memory is
4084 subsequently freed again in the SSL_free() call. However there is an
4085 increased risk during this period of application crashes due to the lack of
4086 memory - which would then mean a more serious Denial of Service.
4088 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4089 (CVE-2016-6307 and CVE-2016-6308)
4093 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4094 had to be removed. Primary reason is that vendor assembler can't
4095 assemble our modules with -KPIC flag. As result it, assembly
4096 support, was not even available as option. But its lack means
4097 lack of side-channel resistant code, which is incompatible with
4098 security by todays standards. Fortunately gcc is readily available
4099 prepackaged option, which we firmly point at...
4103 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4105 * Windows command-line tool supports UTF-8 opt-in option for arguments
4106 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4107 (to any value) allows Windows user to access PKCS#12 file generated
4108 with Windows CryptoAPI and protected with non-ASCII password, as well
4109 as files generated under UTF-8 locale on Linux also protected with
4114 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4115 have been disabled by default and removed from DEFAULT, just like RC4.
4116 See the RC4 item below to re-enable both.
4120 * The method for finding the storage location for the Windows RAND seed file
4121 has changed. First we check %RANDFILE%. If that is not set then we check
4122 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4123 all else fails we fall back to C:\.
4127 * The EVP_EncryptUpdate() function has had its return type changed from void
4128 to int. A return of 0 indicates and error while a return of 1 indicates
4133 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4134 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4135 off the constant time implementation for RSA, DSA and DH have been made
4136 no-ops and deprecated.
4140 * Windows RAND implementation was simplified to only get entropy by
4141 calling CryptGenRandom(). Various other RAND-related tickets
4144 *Joseph Wylie Yandle, Rich Salz*
4146 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4147 and `OPENSSL_LH_`, respectively. The old names are available
4148 with API compatibility. They new names are now completely documented.
4152 * Unify TYPE_up_ref(obj) methods signature.
4153 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4154 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4155 int (instead of void) like all others TYPE_up_ref() methods.
4156 So now these methods also check the return value of CRYPTO_atomic_add(),
4157 and the validity of object reference counter.
4159 *fdasilvayy@gmail.com*
4161 * With Windows Visual Studio builds, the .pdb files are installed
4162 alongside the installed libraries and executables. For a static
4163 library installation, ossl_static.pdb is the associate compiler
4164 generated .pdb file to be used when linking programs.
4168 * Remove openssl.spec. Packaging files belong with the packagers.
4172 * Automatic Darwin/OSX configuration has had a refresh, it will now
4173 recognise x86_64 architectures automatically. You can still decide
4174 to build for a different bitness with the environment variable
4175 KERNEL_BITS (can be 32 or 64), for example:
4177 KERNEL_BITS=32 ./config
4181 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4182 256 bit AES and HMAC with SHA256.
4186 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4190 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4194 * To enable users to have their own config files and build file templates,
4195 Configure looks in the directory indicated by the environment variable
4196 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4197 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4198 name and is used as is.
4202 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4203 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4204 X509_CERT_FILE_CTX was removed.
4208 * "shared" builds are now the default. To create only static libraries use
4209 the "no-shared" Configure option.
4213 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4214 All of these option have not worked for some while and are fundamental
4219 * Make various cleanup routines no-ops and mark them as deprecated. Most
4220 global cleanup functions are no longer required because they are handled
4221 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4222 Explicitly de-initing can cause problems (e.g. where a library that uses
4223 OpenSSL de-inits, but an application is still using it). The affected
4224 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4225 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4226 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4227 COMP_zlib_cleanup().
4231 * --strict-warnings no longer enables runtime debugging options
4232 such as REF_DEBUG. Instead, debug options are automatically
4233 enabled with '--debug' builds.
4235 *Andy Polyakov, Emilia Käsper*
4237 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4238 have been moved out of the public header files. New functions for managing
4239 these have been added.
4243 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4244 objects have been moved out of the public header files. New
4245 functions for managing these have been added.
4249 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4250 have been moved out of the public header files. New functions for managing
4251 these have been added.
4255 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4256 moved out of the public header files. New functions for managing these
4261 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4265 * Removed the mk1mf build scripts.
4269 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4270 it is always safe to #include a header now.
4274 * Removed the aged BC-32 config and all its supporting scripts
4278 * Removed support for Ultrix, Netware, and OS/2.
4282 * Add support for HKDF.
4284 *Alessandro Ghedini*
4286 * Add support for blake2b and blake2s
4290 * Added support for "pipelining". Ciphers that have the
4291 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4292 encryptions/decryptions simultaneously. There are currently no built-in
4293 ciphers with this property but the expectation is that engines will be able
4294 to offer it to significantly improve throughput. Support has been extended
4295 into libssl so that multiple records for a single connection can be
4296 processed in one go (for >=TLS 1.1).
4300 * Added the AFALG engine. This is an async capable engine which is able to
4301 offload work to the Linux kernel. In this initial version it only supports
4302 AES128-CBC. The kernel must be version 4.1.0 or greater.
4306 * OpenSSL now uses a new threading API. It is no longer necessary to
4307 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4308 are two supported threading models: pthreads and windows threads. It is
4309 also possible to configure OpenSSL at compile time for "no-threads". The
4310 old threading API should no longer be used. The functions have been
4311 replaced with "no-op" compatibility macros.
4313 *Alessandro Ghedini, Matt Caswell*
4315 * Modify behavior of ALPN to invoke callback after SNI/servername
4316 callback, such that updates to the SSL_CTX affect ALPN.
4320 * Add SSL_CIPHER queries for authentication and key-exchange.
4324 * Changes to the DEFAULT cipherlist:
4325 - Prefer (EC)DHE handshakes over plain RSA.
4326 - Prefer AEAD ciphers over legacy ciphers.
4327 - Prefer ECDSA over RSA when both certificates are available.
4328 - Prefer TLSv1.2 ciphers/PRF.
4329 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4334 * Change the ECC default curve list to be this, in order: x25519,
4335 secp256r1, secp521r1, secp384r1.
4339 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4340 disabled by default. They can be re-enabled using the
4341 enable-weak-ssl-ciphers option to Configure.
4345 * If the server has ALPN configured, but supports no protocols that the
4346 client advertises, send a fatal "no_application_protocol" alert.
4347 This behaviour is SHALL in RFC 7301, though it isn't universally
4348 implemented by other servers.
4352 * Add X25519 support.
4353 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4354 for public and private key encoding using the format documented in
4355 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4356 key generation and key derivation.
4358 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4363 * Deprecate SRP_VBASE_get_by_user.
4364 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4365 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4366 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4367 seed, even if the seed is configured.
4369 Users should use SRP_VBASE_get1_by_user instead. Note that in
4370 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4371 also that even though configuring the SRP seed attempts to hide
4372 invalid usernames by continuing the handshake with fake
4373 credentials, this behaviour is not constant time and no strong
4374 guarantees are made that the handshake is indistinguishable from
4375 that of a valid user.
4379 * Configuration change; it's now possible to build dynamic engines
4380 without having to build shared libraries and vice versa. This
4381 only applies to the engines in `engines/`, those in `crypto/engine/`
4382 will always be built into libcrypto (i.e. "static").
4384 Building dynamic engines is enabled by default; to disable, use
4385 the configuration option "disable-dynamic-engine".
4387 The only requirements for building dynamic engines are the
4388 presence of the DSO module and building with position independent
4389 code, so they will also automatically be disabled if configuring
4390 with "disable-dso" or "disable-pic".
4392 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4393 are also taken away from openssl/opensslconf.h, as they are
4398 * Configuration change; if there is a known flag to compile
4399 position independent code, it will always be applied on the
4400 libcrypto and libssl object files, and never on the application
4401 object files. This means other libraries that use routines from
4402 libcrypto / libssl can be made into shared libraries regardless
4403 of how OpenSSL was configured.
4405 If this isn't desirable, the configuration options "disable-pic"
4406 or "no-pic" can be used to disable the use of PIC. This will
4407 also disable building shared libraries and dynamic engines.
4411 * Removed JPAKE code. It was experimental and has no wide use.
4415 * The INSTALL_PREFIX Makefile variable has been renamed to
4416 DESTDIR. That makes for less confusion on what this variable
4417 is for. Also, the configuration option --install_prefix is
4422 * Heartbeat for TLS has been removed and is disabled by default
4423 for DTLS; configure with enable-heartbeats. Code that uses the
4424 old #define's might need to be updated.
4426 *Emilia Käsper, Rich Salz*
4428 * Rename REF_CHECK to REF_DEBUG.
4432 * New "unified" build system
4434 The "unified" build system is aimed to be a common system for all
4435 platforms we support. With it comes new support for VMS.
4437 This system builds supports building in a different directory tree
4438 than the source tree. It produces one Makefile (for unix family
4439 or lookalikes), or one descrip.mms (for VMS).
4441 The source of information to make the Makefile / descrip.mms is
4442 small files called 'build.info', holding the necessary
4443 information for each directory with source to compile, and a
4444 template in Configurations, like unix-Makefile.tmpl or
4447 With this change, the library names were also renamed on Windows
4448 and on VMS. They now have names that are closer to the standard
4449 on Unix, and include the major version number, and in certain
4450 cases, the architecture they are built for. See "Notes on shared
4451 libraries" in INSTALL.
4453 We rely heavily on the perl module Text::Template.
4457 * Added support for auto-initialisation and de-initialisation of the library.
4458 OpenSSL no longer requires explicit init or deinit routines to be called,
4459 except in certain circumstances. See the OPENSSL_init_crypto() and
4460 OPENSSL_init_ssl() man pages for further information.
4464 * The arguments to the DTLSv1_listen function have changed. Specifically the
4465 "peer" argument is now expected to be a BIO_ADDR object.
4467 * Rewrite of BIO networking library. The BIO library lacked consistent
4468 support of IPv6, and adding it required some more extensive
4469 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4470 which hold all types of addresses and chains of address information.
4471 It also introduces a new API, with functions like BIO_socket,
4472 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4473 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4474 have been adapted accordingly.
4478 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4483 * CRIME protection: disable compression by default, even if OpenSSL is
4484 compiled with zlib enabled. Applications can still enable compression
4485 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4486 using the SSL_CONF library to configure compression.
4490 * The signature of the session callback configured with
4491 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4492 was explicitly marked as `const unsigned char*` instead of
4497 * Always DPURIFY. Remove the use of uninitialized memory in the
4498 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4502 * Removed many obsolete configuration items, including
4503 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4504 MD2_CHAR, MD2_INT, MD2_LONG
4506 IDEA_SHORT, IDEA_LONG
4507 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4509 *Rich Salz, with advice from Andy Polyakov*
4511 * Many BN internals have been moved to an internal header file.
4513 *Rich Salz with help from Andy Polyakov*
4515 * Configuration and writing out the results from it has changed.
4516 Files such as Makefile include/openssl/opensslconf.h and are now
4517 produced through general templates, such as Makefile.in and
4518 crypto/opensslconf.h.in and some help from the perl module
4521 Also, the center of configuration information is no longer
4522 Makefile. Instead, Configure produces a perl module in
4523 configdata.pm which holds most of the config data (in the hash
4524 table %config), the target data that comes from the target
4525 configuration in one of the `Configurations/*.conf` files (in
4530 * To clarify their intended purposes, the Configure options
4531 --prefix and --openssldir change their semantics, and become more
4532 straightforward and less interdependent.
4534 --prefix shall be used exclusively to give the location INSTALLTOP
4535 where programs, scripts, libraries, include files and manuals are
4536 going to be installed. The default is now /usr/local.
4538 --openssldir shall be used exclusively to give the default
4539 location OPENSSLDIR where certificates, private keys, CRLs are
4540 managed. This is also where the default openssl.cnf gets
4542 If the directory given with this option is a relative path, the
4543 values of both the --prefix value and the --openssldir value will
4544 be combined to become OPENSSLDIR.
4545 The default for --openssldir is INSTALLTOP/ssl.
4547 Anyone who uses --openssldir to specify where OpenSSL is to be
4548 installed MUST change to use --prefix instead.
4552 * The GOST engine was out of date and therefore it has been removed. An up
4553 to date GOST engine is now being maintained in an external repository.
4554 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4555 support for GOST ciphersuites (these are only activated if a GOST engine
4560 * EGD is no longer supported by default; use enable-egd when
4563 *Ben Kaduk and Rich Salz*
4565 * The distribution now has Makefile.in files, which are used to
4566 create Makefile's when Configure is run. *Configure must be run
4567 before trying to build now.*
4571 * The return value for SSL_CIPHER_description() for error conditions
4576 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4578 Obtaining and performing DNSSEC validation of TLSA records is
4579 the application's responsibility. The application provides
4580 the TLSA records of its choice to OpenSSL, and these are then
4581 used to authenticate the peer.
4583 The TLSA records need not even come from DNS. They can, for
4584 example, be used to implement local end-entity certificate or
4585 trust-anchor "pinning", where the "pin" data takes the form
4586 of TLSA records, which can augment or replace verification
4587 based on the usual WebPKI public certification authorities.
4591 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4592 continues to support deprecated interfaces in default builds.
4593 However, applications are strongly advised to compile their
4594 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4595 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4596 or the 1.1.0 releases.
4598 In environments in which all applications have been ported to
4599 not use any deprecated interfaces OpenSSL's Configure script
4600 should be used with the --api=1.1.0 option to entirely remove
4601 support for the deprecated features from the library and
4602 unconditionally disable them in the installed headers.
4603 Essentially the same effect can be achieved with the "no-deprecated"
4604 argument to Configure, except that this will always restrict
4605 the build to just the latest API, rather than a fixed API
4608 As applications are ported to future revisions of the API,
4609 they should update their compile-time OPENSSL_API_COMPAT define
4610 accordingly, but in most cases should be able to continue to
4611 compile with later releases.
4613 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4614 0x10000000L and 0x00908000L, respectively. However those
4615 versions did not support the OPENSSL_API_COMPAT feature, and
4616 so applications are not typically tested for explicit support
4617 of just the undeprecated features of either release.
4621 * Add support for setting the minimum and maximum supported protocol.
4622 It can bet set via the SSL_set_min_proto_version() and
4623 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4624 MaxProtocol. It's recommended to use the new APIs to disable
4625 protocols instead of disabling individual protocols using
4626 SSL_set_options() or SSL_CONF's Protocol. This change also
4627 removes support for disabling TLS 1.2 in the OpenSSL TLS
4628 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4632 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4636 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4637 and integrates ECDSA and ECDH functionality into EC. Implementations can
4638 now redirect key generation and no longer need to convert to or from
4641 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4642 include the ec.h header file instead.
4646 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4647 ciphers who are no longer supported and drops support the ephemeral RSA key
4648 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4652 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4653 opaque. For HMAC_CTX, the following constructors and destructors
4656 HMAC_CTX *HMAC_CTX_new(void);
4657 void HMAC_CTX_free(HMAC_CTX *ctx);
4659 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4660 destroy such methods has been added. See EVP_MD_meth_new(3) and
4661 EVP_CIPHER_meth_new(3) for documentation.
4664 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4665 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4666 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4667 an already created structure.
4668 2) For consistency with the majority of our object creators and
4669 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4670 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4671 for deprecated builds.
4675 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4676 cryptographic operations to be performed asynchronously as long as an
4677 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4678 further details. Libssl has also had this capability integrated with the
4679 introduction of the new mode SSL_MODE_ASYNC and associated error
4680 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4681 pages. This work was developed in partnership with Intel Corp.
4685 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4686 always enabled now. If you want to disable the support you should
4687 exclude it using the list of supported ciphers. This also means that the
4688 "-no_ecdhe" option has been removed from s_server.
4692 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4693 SSL_{CTX_}set1_curves() which can set a list.
4697 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4698 curve you want to support using SSL_{CTX_}set1_curves().
4702 * State machine rewrite. The state machine code has been significantly
4703 refactored in order to remove much duplication of code and solve issues
4704 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4705 further details). This change does have some associated API changes.
4706 Notably the SSL_state() function has been removed and replaced by
4707 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4708 SSL_set_state() has been removed altogether. The previous handshake states
4709 defined in ssl.h and ssl3.h have also been removed.
4713 * All instances of the string "ssleay" in the public API were replaced
4714 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4715 Some error codes related to internal RSA_eay API's were renamed.
4719 * The demo files in crypto/threads were moved to demo/threads.
4723 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4726 *Matt Caswell, Rich Salz*
4728 * New ASN.1 embed macro.
4730 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4731 structure is not allocated: it is part of the parent. That is instead of
4739 This reduces memory fragmentation and make it impossible to accidentally
4740 set a mandatory field to NULL.
4742 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4743 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4744 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4749 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4753 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4754 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4755 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4756 DES and RC4 ciphersuites.
4760 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4761 This changes the decoding behaviour for some invalid messages,
4762 though the change is mostly in the more lenient direction, and
4763 legacy behaviour is preserved as much as possible.
4767 * Fix no-stdio build.
4768 *David Woodhouse <David.Woodhouse@intel.com> and also*
4769 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4771 * New testing framework
4772 The testing framework has been largely rewritten and is now using
4773 perl and the perl modules Test::Harness and an extended variant of
4774 Test::More called OpenSSL::Test to do its work. All test scripts in
4775 test/ have been rewritten into test recipes, and all direct calls to
4776 executables in test/Makefile have become individual recipes using the
4777 simplified testing OpenSSL::Test::Simple.
4779 For documentation on our testing modules, do:
4781 perldoc test/testlib/OpenSSL/Test/Simple.pm
4782 perldoc test/testlib/OpenSSL/Test.pm
4786 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4787 are used; the latter aborts on memory leaks (usually checked on exit).
4788 Some undocumented "set malloc, etc., hooks" functions were removed
4789 and others were changed. All are now documented.
4793 * In DSA_generate_parameters_ex, if the provided seed is too short,
4796 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4798 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4799 from RFC4279, RFC4785, RFC5487, RFC5489.
4801 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4802 original RSA_PSK patch.
4806 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4807 era flag was never set throughout the codebase (only read). Also removed
4808 SSL3_FLAGS_POP_BUFFER which was only used if
4809 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4813 * Changed the default name options in the "ca", "crl", "req" and "x509"
4814 to be "oneline" instead of "compat".
4818 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4819 not aware of clients that still exhibit this bug, and the workaround
4820 hasn't been working properly for a while.
4824 * The return type of BIO_number_read() and BIO_number_written() as well as
4825 the corresponding num_read and num_write members in the BIO structure has
4826 changed from unsigned long to uint64_t. On platforms where an unsigned
4827 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4832 * Given the pervasive nature of TLS extensions it is inadvisable to run
4833 OpenSSL without support for them. It also means that maintaining
4834 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4835 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4839 * Removed support for the two export grade static DH ciphersuites
4840 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4841 were newly added (along with a number of other static DH ciphersuites) to
4842 1.0.2. However the two export ones have *never* worked since they were
4843 introduced. It seems strange in any case to be adding new export
4844 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4848 * Version negotiation has been rewritten. In particular SSLv23_method(),
4849 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4850 and turned into macros which simply call the new preferred function names
4851 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4852 should use the new names instead. Also as part of this change the ssl23.h
4853 header file has been removed.
4857 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4858 code and the associated standard is no longer considered fit-for-purpose.
4862 * RT2547 was closed. When generating a private key, try to make the
4863 output file readable only by the owner. This behavior change might
4864 be noticeable when interacting with other software.
4866 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4871 * Added HTTP GET support to the ocsp command.
4875 * Changed default digest for the dgst and enc commands from MD5 to
4880 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4884 * Added support for TLS extended master secret from
4885 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4886 initial patch which was a great help during development.
4890 * All libssl internal structures have been removed from the public header
4891 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4892 now redundant). Users should not attempt to access internal structures
4893 directly. Instead they should use the provided API functions.
4897 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4898 Access to deprecated functions can be re-enabled by running config with
4899 "enable-deprecated". In addition applications wishing to use deprecated
4900 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4901 will, by default, disable some transitive includes that previously existed
4902 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4906 * Added support for OCB mode. OpenSSL has been granted a patent license
4907 compatible with the OpenSSL license for use of OCB. Details are available
4908 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4909 for OCB can be removed by calling config with no-ocb.
4913 * SSLv2 support has been removed. It still supports receiving an SSLv2
4914 compatible client hello.
4918 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4919 done while fixing the error code for the key-too-small case.
4921 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4923 * CA.sh has been removed; use CA.pl instead.
4927 * Removed old DES API.
4931 * Remove various unsupported platforms:
4937 Sinix/ReliantUNIX RM400
4942 16-bit platforms such as WIN16
4946 * Clean up OPENSSL_NO_xxx #define's
4947 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4948 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4949 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4950 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4951 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4952 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4953 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4954 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4955 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4956 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4960 * Cleaned up dead code
4961 Remove all but one '#ifdef undef' which is to be looked at.
4965 * Clean up calling of xxx_free routines.
4966 Just like free(), fix most of the xxx_free routines to accept
4967 NULL. Remove the non-null checks from callers. Save much code.
4971 * Add secure heap for storage of private keys (when possible).
4972 Add BIO_s_secmem(), CBIGNUM, etc.
4973 Contributed by Akamai Technologies under our Corporate CLA.
4977 * Experimental support for a new, fast, unbiased prime candidate generator,
4978 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4980 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4982 * New output format NSS in the sess_id command line tool. This allows
4983 exporting the session id and the master key in NSS keylog format.
4985 *Martin Kaiser <martin@kaiser.cx>*
4987 * Harmonize version and its documentation. -f flag is used to display
4990 *mancha <mancha1@zoho.com>*
4992 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4993 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4995 *mancha <mancha1@zoho.com>*
4997 * Fix some double frees. These are not thought to be exploitable.
4999 *mancha <mancha1@zoho.com>*
5001 * A missing bounds check in the handling of the TLS heartbeat extension
5002 can be used to reveal up to 64k of memory to a connected client or
5005 Thanks for Neel Mehta of Google Security for discovering this bug and to
5006 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
5007 preparing the fix ([CVE-2014-0160])
5009 *Adam Langley, Bodo Moeller*
5011 * Fix for the attack described in the paper "Recovering OpenSSL
5012 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
5013 by Yuval Yarom and Naomi Benger. Details can be obtained from:
5014 <http://eprint.iacr.org/2014/140>
5016 Thanks to Yuval Yarom and Naomi Benger for discovering this
5017 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5019 *Yuval Yarom and Naomi Benger*
5021 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5022 this fixes a limitation in previous versions of OpenSSL.
5026 * Experimental encrypt-then-mac support.
5028 Experimental support for encrypt then mac from
5029 draft-gutmann-tls-encrypt-then-mac-02.txt
5031 To enable it set the appropriate extension number (0x42 for the test
5032 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
5034 For non-compliant peers (i.e. just about everything) this should have no
5037 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
5041 * Add EVP support for key wrapping algorithms, to avoid problems with
5042 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5043 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5044 algorithms and include tests cases.
5048 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
5053 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5054 MGF1 digest and OAEP label.
5058 * Make openssl verify return errors.
5060 *Chris Palmer <palmer@google.com> and Ben Laurie*
5062 * New function ASN1_TIME_diff to calculate the difference between two
5063 ASN1_TIME structures or one structure and the current time.
5067 * Update fips_test_suite to support multiple command line options. New
5068 test to induce all self test errors in sequence and check expected
5073 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5074 sign or verify all in one operation.
5078 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5079 test programs and fips_test_suite. Includes functionality to parse
5080 the minimal script output of fipsalgest.pl directly.
5084 * Add authorisation parameter to FIPS_module_mode_set().
5088 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5092 * Use separate DRBG fields for internal and external flags. New function
5093 FIPS_drbg_health_check() to perform on demand health checking. Add
5094 generation tests to fips_test_suite with reduced health check interval to
5095 demonstrate periodic health checking. Add "nodh" option to
5096 fips_test_suite to skip very slow DH test.
5100 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5105 * More extensive health check for DRBG checking many more failure modes.
5106 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5107 combination: call this in fips_test_suite.
5111 * Add support for canonical generation of DSA parameter 'g'. See
5114 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5115 POST to handle HMAC cases.
5119 * Add functions FIPS_module_version() and FIPS_module_version_text()
5120 to return numerical and string versions of the FIPS module number.
5124 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5125 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5126 outside the validated module in the FIPS capable OpenSSL.
5130 * Minor change to DRBG entropy callback semantics. In some cases
5131 there is no multiple of the block length between min_len and
5132 max_len. Allow the callback to return more than max_len bytes
5133 of entropy but discard any extra: it is the callback's responsibility
5134 to ensure that the extra data discarded does not impact the
5135 requested amount of entropy.
5139 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5140 information in FIPS186-3, SP800-57 and SP800-131A.
5144 * CCM support via EVP. Interface is very similar to GCM case except we
5145 must supply all data in one chunk (i.e. no update, final) and the
5146 message length must be supplied if AAD is used. Add algorithm test
5151 * Initial version of POST overhaul. Add POST callback to allow the status
5152 of POST to be monitored and/or failures induced. Modify fips_test_suite
5153 to use callback. Always run all selftests even if one fails.
5157 * XTS support including algorithm test driver in the fips_gcmtest program.
5158 Note: this does increase the maximum key length from 32 to 64 bytes but
5159 there should be no binary compatibility issues as existing applications
5160 will never use XTS mode.
5164 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5165 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5166 performs algorithm blocking for unapproved PRNG types. Also do not
5167 set PRNG type in FIPS_mode_set(): leave this to the application.
5168 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5169 the standard OpenSSL PRNG: set additional data to a date time vector.
5173 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5174 This shouldn't present any incompatibility problems because applications
5175 shouldn't be using these directly and any that are will need to rethink
5176 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5180 * Extensive self tests and health checking required by SP800-90 DRBG.
5181 Remove strength parameter from FIPS_drbg_instantiate and always
5182 instantiate at maximum supported strength.
5186 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5190 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5194 * New function DH_compute_key_padded() to compute a DH key and pad with
5195 leading zeroes if needed: this complies with SP800-56A et al.
5199 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5200 anything, incomplete, subject to change and largely untested at present.
5204 * Modify fipscanisteronly build option to only build the necessary object
5205 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5209 * Add experimental option FIPSSYMS to give all symbols in
5210 fipscanister.o and FIPS or fips prefix. This will avoid
5211 conflicts with future versions of OpenSSL. Add perl script
5212 util/fipsas.pl to preprocess assembly language source files
5213 and rename any affected symbols.
5217 * Add selftest checks and algorithm block of non-fips algorithms in
5218 FIPS mode. Remove DES2 from selftests.
5222 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5223 return internal method without any ENGINE dependencies. Add new
5224 tiny fips sign and verify functions.
5228 * New build option no-ec2m to disable characteristic 2 code.
5232 * New build option "fipscanisteronly". This only builds fipscanister.o
5233 and (currently) associated fips utilities. Uses the file Makefile.fips
5234 instead of Makefile.org as the prototype.
5238 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5239 Update fips_gcmtest to use IV generator.
5243 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5244 setting output buffer to NULL. The `*Final` function must be
5245 called although it will not retrieve any additional data. The tag
5246 can be set or retrieved with a ctrl. The IV length is by default 12
5247 bytes (96 bits) but can be set to an alternative value. If the IV
5248 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5253 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5254 underlying do_cipher function handles all cipher semantics itself
5255 including padding and finalisation. This is useful if (for example)
5256 an ENGINE cipher handles block padding itself. The behaviour of
5257 do_cipher is subtly changed if this flag is set: the return value
5258 is the number of characters written to the output buffer (zero is
5259 no longer an error code) or a negative error code. Also if the
5260 input buffer is NULL and length 0 finalisation should be performed.
5264 * If a candidate issuer certificate is already part of the constructed
5265 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5269 * Improve forward-security support: add functions
5271 void SSL_CTX_set_not_resumable_session_callback(
5272 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5273 void SSL_set_not_resumable_session_callback(
5274 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5276 for use by SSL/TLS servers; the callback function will be called whenever a
5277 new session is created, and gets to decide whether the session may be
5278 cached to make it resumable (return 0) or not (return 1). (As by the
5279 SSL/TLS protocol specifications, the session_id sent by the server will be
5280 empty to indicate that the session is not resumable; also, the server will
5281 not generate RFC 4507 (RFC 5077) session tickets.)
5283 A simple reasonable callback implementation is to return is_forward_secure.
5284 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5285 by the SSL/TLS server library, indicating whether it can provide forward
5288 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5290 * New -verify_name option in command line utilities to set verification
5295 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5296 Add CMAC pkey methods.
5300 * Experimental renegotiation in s_server -www mode. If the client
5301 browses /reneg connection is renegotiated. If /renegcert it is
5302 renegotiated requesting a certificate.
5306 * Add an "external" session cache for debugging purposes to s_server. This
5307 should help trace issues which normally are only apparent in deployed
5308 multi-process servers.
5312 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5313 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5314 BIO_set_cipher() and some obscure PEM functions were changed so they
5315 can now return an error. The RAND changes required a change to the
5316 RAND_METHOD structure.
5320 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5321 a gcc attribute to warn if the result of a function is ignored. This
5322 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5323 whose return value is often ignored.
5327 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5328 These allow SCTs (signed certificate timestamps) to be requested and
5329 validated when establishing a connection.
5331 *Rob Percival <robpercival@google.com>*
5336 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5338 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5339 used even when parsing explicit parameters, when loading a encoded key
5340 or calling `EC_GROUP_new_from_ecpkparameters()`/
5341 `EC_GROUP_new_from_ecparameters()`.
5342 This prevents bypass of security hardening and performance gains,
5343 especially for curves with specialized EC_METHODs.
5344 By default, if a key encoded with explicit parameters is loaded and later
5345 encoded, the output is still encoded with explicit parameters, even if
5346 internally a "named" EC_GROUP is used for computation.
5350 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5351 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5352 NULL. After this change, only the cofactor parameter can be NULL. It also
5353 does some minimal sanity checks on the passed order.
5358 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5359 An attack is simple, if the first CMS_recipientInfo is valid but the
5360 second CMS_recipientInfo is chosen ciphertext. If the second
5361 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5362 encryption key will be replaced by garbage, and the message cannot be
5363 decoded, but if the RSA decryption fails, the correct encryption key is
5364 used and the recipient will not notice the attack.
5365 As a work around for this potential attack the length of the decrypted
5366 key must be equal to the cipher default key length, in case the
5367 certificate is not given and all recipientInfo are tried out.
5368 The old behaviour can be re-enabled in the CMS code by setting the
5369 CMS_DEBUG_DECRYPT flag.
5374 * Document issue with installation paths in diverse Windows builds
5376 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5377 binaries and run-time config file.
5382 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5384 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5385 This changes the size when using the `genpkey` command when no size is given.
5386 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5387 generation commands to use 2048 bits by default.
5391 * Add FIPS support for Android Arm 64-bit
5393 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5394 Module in Version 2.0.10. For some reason, the corresponding target
5395 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5396 built with FIPS support on Android Arm 64-bit. This omission has been
5399 *Matthias St. Pierre*
5401 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5403 * 0-byte record padding oracle
5405 If an application encounters a fatal protocol error and then calls
5406 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5407 then OpenSSL can respond differently to the calling application if a 0 byte
5408 record is received with invalid padding compared to if a 0 byte record is
5409 received with an invalid MAC. If the application then behaves differently
5410 based on that in a way that is detectable to the remote peer, then this
5411 amounts to a padding oracle that could be used to decrypt data.
5413 In order for this to be exploitable "non-stitched" ciphersuites must be in
5414 use. Stitched ciphersuites are optimised implementations of certain
5415 commonly used ciphersuites. Also the application must call SSL_shutdown()
5416 twice even if a protocol error has occurred (applications should not do
5417 this but some do anyway).
5419 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5420 Aviram, with additional investigation by Steven Collison and Andrew
5421 Hourselt. It was reported to OpenSSL on 10th December 2018.
5426 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5430 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5432 * Microarchitecture timing vulnerability in ECC scalar multiplication
5434 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5435 shown to be vulnerable to a microarchitecture timing side channel attack.
5436 An attacker with sufficient access to mount local timing attacks during
5437 ECDSA signature generation could recover the private key.
5439 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5440 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5446 * Timing vulnerability in DSA signature generation
5448 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5449 timing side channel attack. An attacker could use variations in the signing
5450 algorithm to recover the private key.
5452 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5457 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5458 Module, accidentally introduced while backporting security fixes from the
5459 development branch and hindering the use of ECC in FIPS mode.
5463 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5465 * Client DoS due to large DH parameter
5467 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5468 malicious server can send a very large prime value to the client. This will
5469 cause the client to spend an unreasonably long period of time generating a
5470 key for this prime resulting in a hang until the client has finished. This
5471 could be exploited in a Denial Of Service attack.
5473 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5478 * Cache timing vulnerability in RSA Key Generation
5480 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5481 a cache timing side channel attack. An attacker with sufficient access to
5482 mount cache timing attacks during the RSA key generation process could
5483 recover the private key.
5485 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5486 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5491 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5492 parameter is no longer accepted, as it leads to a corrupt table. NULL
5493 pem_str is reserved for alias entries only.
5497 * Revert blinding in ECDSA sign and instead make problematic addition
5498 length-invariant. Switch even to fixed-length Montgomery multiplication.
5502 * Change generating and checking of primes so that the error rate of not
5503 being prime depends on the intended use based on the size of the input.
5504 For larger primes this will result in more rounds of Miller-Rabin.
5505 The maximal error rate for primes with more than 1080 bits is lowered
5508 *Kurt Roeckx, Annie Yousar*
5510 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5514 * Add blinding to ECDSA and DSA signatures to protect against side channel
5515 attacks discovered by Keegan Ryan (NCC Group).
5519 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5520 now allow empty (zero character) pass phrases.
5524 * Certificate time validation (X509_cmp_time) enforces stricter
5525 compliance with RFC 5280. Fractional seconds and timezone offsets
5526 are no longer allowed.
5530 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5532 * Constructed ASN.1 types with a recursive definition could exceed the stack
5534 Constructed ASN.1 types with a recursive definition (such as can be found
5535 in PKCS7) could eventually exceed the stack given malicious input with
5536 excessive recursion. This could result in a Denial Of Service attack. There
5537 are no such structures used within SSL/TLS that come from untrusted sources
5538 so this is considered safe.
5540 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5546 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5548 * Read/write after SSL object in error state
5550 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5551 mechanism. The intent was that if a fatal error occurred during a handshake
5552 then OpenSSL would move into the error state and would immediately fail if
5553 you attempted to continue the handshake. This works as designed for the
5554 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5555 SSL_connect()), however due to a bug it does not work correctly if
5556 SSL_read() or SSL_write() is called directly. In that scenario, if the
5557 handshake fails then a fatal error will be returned in the initial function
5558 call. If SSL_read()/SSL_write() is subsequently called by the application
5559 for the same SSL object then it will succeed and the data is passed without
5560 being decrypted/encrypted directly from the SSL/TLS record layer.
5562 In order to exploit this issue an application bug would have to be present
5563 that resulted in a call to SSL_read()/SSL_write() being issued after having
5564 already received a fatal error.
5566 This issue was reported to OpenSSL by David Benjamin (Google).
5571 * rsaz_1024_mul_avx2 overflow bug on x86_64
5573 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5574 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5575 Analysis suggests that attacks against RSA and DSA as a result of this
5576 defect would be very difficult to perform and are not believed likely.
5577 Attacks against DH1024 are considered just feasible, because most of the
5578 work necessary to deduce information about a private key may be performed
5579 offline. The amount of resources required for such an attack would be
5580 significant. However, for an attack on TLS to be meaningful, the server
5581 would have to share the DH1024 private key among multiple clients, which is
5582 no longer an option since CVE-2016-0701.
5584 This only affects processors that support the AVX2 but not ADX extensions
5585 like Intel Haswell (4th generation).
5587 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5588 was originally found via the OSS-Fuzz project.
5593 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5595 * bn_sqrx8x_internal carry bug on x86_64
5597 There is a carry propagating bug in the x86_64 Montgomery squaring
5598 procedure. No EC algorithms are affected. Analysis suggests that attacks
5599 against RSA and DSA as a result of this defect would be very difficult to
5600 perform and are not believed likely. Attacks against DH are considered just
5601 feasible (although very difficult) because most of the work necessary to
5602 deduce information about a private key may be performed offline. The amount
5603 of resources required for such an attack would be very significant and
5604 likely only accessible to a limited number of attackers. An attacker would
5605 additionally need online access to an unpatched system using the target
5606 private key in a scenario with persistent DH parameters and a private
5607 key that is shared between multiple clients.
5609 This only affects processors that support the BMI1, BMI2 and ADX extensions
5610 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5612 This issue was reported to OpenSSL by the OSS-Fuzz project.
5617 * Malformed X.509 IPAddressFamily could cause OOB read
5619 If an X.509 certificate has a malformed IPAddressFamily extension,
5620 OpenSSL could do a one-byte buffer overread. The most likely result
5621 would be an erroneous display of the certificate in text format.
5623 This issue was reported to OpenSSL by the OSS-Fuzz project.
5627 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5629 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5630 platform rather than 'mingw'.
5634 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5636 * Truncated packet could crash via OOB read
5638 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5639 cipher is being used, then a truncated packet can cause that host to
5640 perform an out-of-bounds read, usually resulting in a crash.
5642 This issue was reported to OpenSSL by Robert Święcki of Google.
5647 * BN_mod_exp may produce incorrect results on x86_64
5649 There is a carry propagating bug in the x86_64 Montgomery squaring
5650 procedure. No EC algorithms are affected. Analysis suggests that attacks
5651 against RSA and DSA as a result of this defect would be very difficult to
5652 perform and are not believed likely. Attacks against DH are considered just
5653 feasible (although very difficult) because most of the work necessary to
5654 deduce information about a private key may be performed offline. The amount
5655 of resources required for such an attack would be very significant and
5656 likely only accessible to a limited number of attackers. An attacker would
5657 additionally need online access to an unpatched system using the target
5658 private key in a scenario with persistent DH parameters and a private
5659 key that is shared between multiple clients. For example this can occur by
5660 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5661 similar to CVE-2015-3193 but must be treated as a separate problem.
5663 This issue was reported to OpenSSL by the OSS-Fuzz project.
5668 * Montgomery multiplication may produce incorrect results
5670 There is a carry propagating bug in the Broadwell-specific Montgomery
5671 multiplication procedure that handles input lengths divisible by, but
5672 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5673 and DH private keys are impossible. This is because the subroutine in
5674 question is not used in operations with the private key itself and an input
5675 of the attacker's direct choice. Otherwise the bug can manifest itself as
5676 transient authentication and key negotiation failures or reproducible
5677 erroneous outcome of public-key operations with specially crafted input.
5678 Among EC algorithms only Brainpool P-512 curves are affected and one
5679 presumably can attack ECDH key negotiation. Impact was not analyzed in
5680 detail, because pre-requisites for attack are considered unlikely. Namely
5681 multiple clients have to choose the curve in question and the server has to
5682 share the private key among them, neither of which is default behaviour.
5683 Even then only clients that chose the curve will be affected.
5685 This issue was publicly reported as transient failures and was not
5686 initially recognized as a security issue. Thanks to Richard Morgan for
5687 providing reproducible case.
5692 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5693 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5694 prevent issues where no progress is being made and the peer continually
5695 sends unrecognised record types, using up resources processing them.
5699 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5701 * Missing CRL sanity check
5703 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5704 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5705 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5707 This issue only affects the OpenSSL 1.0.2i
5712 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5714 * OCSP Status Request extension unbounded memory growth
5716 A malicious client can send an excessively large OCSP Status Request
5717 extension. If that client continually requests renegotiation, sending a
5718 large OCSP Status Request extension each time, then there will be unbounded
5719 memory growth on the server. This will eventually lead to a Denial Of
5720 Service attack through memory exhaustion. Servers with a default
5721 configuration are vulnerable even if they do not support OCSP. Builds using
5722 the "no-ocsp" build time option are not affected.
5724 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5729 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5732 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5738 * OOB write in MDC2_Update()
5740 An overflow can occur in MDC2_Update() either if called directly or
5741 through the EVP_DigestUpdate() function using MDC2. If an attacker
5742 is able to supply very large amounts of input data after a previous
5743 call to EVP_EncryptUpdate() with a partial block then a length check
5744 can overflow resulting in a heap corruption.
5746 The amount of data needed is comparable to SIZE_MAX which is impractical
5749 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5754 * Malformed SHA512 ticket DoS
5756 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5757 DoS attack where a malformed ticket will result in an OOB read which will
5760 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5761 a custom server callback and ticket lookup mechanism.
5763 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5768 * OOB write in BN_bn2dec()
5770 The function BN_bn2dec() does not check the return value of BN_div_word().
5771 This can cause an OOB write if an application uses this function with an
5772 overly large BIGNUM. This could be a problem if an overly large certificate
5773 or CRL is printed out from an untrusted source. TLS is not affected because
5774 record limits will reject an oversized certificate before it is parsed.
5776 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5781 * OOB read in TS_OBJ_print_bio()
5783 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5784 the total length the OID text representation would use and not the amount
5785 of data written. This will result in OOB reads when large OIDs are
5788 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5793 * Pointer arithmetic undefined behaviour
5795 Avoid some undefined pointer arithmetic
5797 A common idiom in the codebase is to check limits in the following manner:
5800 Where "p" points to some malloc'd data of SIZE bytes and
5803 "len" here could be from some externally supplied data (e.g. from a TLS
5806 The rules of C pointer arithmetic are such that "p + len" is only well
5807 defined where len <= SIZE. Therefore the above idiom is actually
5808 undefined behaviour.
5810 For example this could cause problems if some malloc implementation
5811 provides an address for "p" such that "p + len" actually overflows for
5812 values of len that are too big and therefore p + len < limit.
5814 This issue was reported to OpenSSL by Guido Vranken
5819 * Constant time flag not preserved in DSA signing
5821 Operations in the DSA signing algorithm should run in constant time in
5822 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5823 implementation means that a non-constant time codepath is followed for
5824 certain operations. This has been demonstrated through a cache-timing
5825 attack to be sufficient for an attacker to recover the private DSA key.
5827 This issue was reported by César Pereida (Aalto University), Billy Brumley
5828 (Tampere University of Technology), and Yuval Yarom (The University of
5829 Adelaide and NICTA).
5834 * DTLS buffered message DoS
5836 In a DTLS connection where handshake messages are delivered out-of-order
5837 those messages that OpenSSL is not yet ready to process will be buffered
5838 for later use. Under certain circumstances, a flaw in the logic means that
5839 those messages do not get removed from the buffer even though the handshake
5840 has been completed. An attacker could force up to approx. 15 messages to
5841 remain in the buffer when they are no longer required. These messages will
5842 be cleared when the DTLS connection is closed. The default maximum size for
5843 a message is 100k. Therefore, the attacker could force an additional 1500k
5844 to be consumed per connection. By opening many simultaneous connections an
5845 attacker could cause a DoS attack through memory exhaustion.
5847 This issue was reported to OpenSSL by Quan Luo.
5852 * DTLS replay protection DoS
5854 A flaw in the DTLS replay attack protection mechanism means that records
5855 that arrive for future epochs update the replay protection "window" before
5856 the MAC for the record has been validated. This could be exploited by an
5857 attacker by sending a record for the next epoch (which does not have to
5858 decrypt or have a valid MAC), with a very large sequence number. This means
5859 that all subsequent legitimate packets are dropped causing a denial of
5860 service for a specific DTLS connection.
5862 This issue was reported to OpenSSL by the OCAP audit team.
5867 * Certificate message OOB reads
5869 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5870 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5871 theoretical DoS risk but this has not been observed in practice on common
5874 The messages affected are client certificate, client certificate request
5875 and server certificate. As a result the attack can only be performed
5876 against a client or a server which enables client authentication.
5878 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5883 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5885 * Prevent padding oracle in AES-NI CBC MAC check
5887 A MITM attacker can use a padding oracle attack to decrypt traffic
5888 when the connection uses an AES CBC cipher and the server support
5891 This issue was introduced as part of the fix for Lucky 13 padding
5892 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5893 constant time by making sure that always the same bytes are read and
5894 compared against either the MAC or padding bytes. But it no longer
5895 checked that there was enough data to have both the MAC and padding
5898 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5902 * Fix EVP_EncodeUpdate overflow
5904 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5905 Base64 encoding of binary data. If an attacker is able to supply very large
5906 amounts of input data then a length check can overflow resulting in a heap
5909 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5910 the `PEM_write_bio*` family of functions. These are mainly used within the
5911 OpenSSL command line applications, so any application which processes data
5912 from an untrusted source and outputs it as a PEM file should be considered
5913 vulnerable to this issue. User applications that call these APIs directly
5914 with large amounts of untrusted data may also be vulnerable.
5916 This issue was reported by Guido Vranken.
5921 * Fix EVP_EncryptUpdate overflow
5923 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5924 is able to supply very large amounts of input data after a previous call to
5925 EVP_EncryptUpdate() with a partial block then a length check can overflow
5926 resulting in a heap corruption. Following an analysis of all OpenSSL
5927 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5928 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5929 the first called function after an EVP_EncryptInit(), and therefore that
5930 specific call must be safe. The second form is where the length passed to
5931 EVP_EncryptUpdate() can be seen from the code to be some small value and
5932 therefore there is no possibility of an overflow. Since all instances are
5933 one of these two forms, it is believed that there can be no overflows in
5934 internal code due to this problem. It should be noted that
5935 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5936 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5937 of these calls have also been analysed too and it is believed there are no
5938 instances in internal usage where an overflow could occur.
5940 This issue was reported by Guido Vranken.
5945 * Prevent ASN.1 BIO excessive memory allocation
5947 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5948 a short invalid encoding can cause allocation of large amounts of memory
5949 potentially consuming excessive resources or exhausting memory.
5951 Any application parsing untrusted data through d2i BIO functions is
5952 affected. The memory based functions such as d2i_X509() are *not* affected.
5953 Since the memory based functions are used by the TLS library, TLS
5954 applications are not affected.
5956 This issue was reported by Brian Carpenter.
5963 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5964 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5965 in arbitrary stack data being returned in the buffer.
5967 This issue was reported by Guido Vranken.
5972 * Modify behavior of ALPN to invoke callback after SNI/servername
5973 callback, such that updates to the SSL_CTX affect ALPN.
5977 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5982 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5983 methods are enabled and ssl2 is disabled the methods return NULL.
5987 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5989 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5990 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5991 provide any "EXPORT" or "LOW" strength ciphers.
5995 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5996 is by default disabled at build-time. Builds that are not configured with
5997 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5998 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5999 will need to explicitly call either of:
6001 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6003 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6005 as appropriate. Even if either of those is used, or the application
6006 explicitly uses the version-specific SSLv2_method() or its client and
6007 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6008 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6009 ciphers, and SSLv2 56-bit DES are no longer available.
6014 * Fix a double-free in DSA code
6016 A double free bug was discovered when OpenSSL parses malformed DSA private
6017 keys and could lead to a DoS attack or memory corruption for applications
6018 that receive DSA private keys from untrusted sources. This scenario is
6021 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6027 * Disable SRP fake user seed to address a server memory leak.
6029 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6031 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6032 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6033 was changed to ignore the "fake user" SRP seed, even if the seed
6036 Users should use SRP_VBASE_get1_by_user instead. Note that in
6037 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6038 also that even though configuring the SRP seed attempts to hide
6039 invalid usernames by continuing the handshake with fake
6040 credentials, this behaviour is not constant time and no strong
6041 guarantees are made that the handshake is indistinguishable from
6042 that of a valid user.
6047 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6049 In the BN_hex2bn function the number of hex digits is calculated using an
6050 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6051 large values of `i` this can result in `bn_expand` not allocating any
6052 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6053 field as NULL leading to a subsequent NULL ptr deref. For very large values
6054 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6055 In this case memory is allocated to the internal BIGNUM data field, but it
6056 is insufficiently sized leading to heap corruption. A similar issue exists
6057 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6058 is ever called by user applications with very large untrusted hex/dec data.
6059 This is anticipated to be a rare occurrence.
6061 All OpenSSL internal usage of these functions use data that is not expected
6062 to be untrusted, e.g. config file data or application command line
6063 arguments. If user developed applications generate config file data based
6064 on untrusted data then it is possible that this could also lead to security
6065 consequences. This is also anticipated to be rare.
6067 This issue was reported to OpenSSL by Guido Vranken.
6072 * Fix memory issues in `BIO_*printf` functions
6074 The internal `fmtstr` function used in processing a "%s" format string in
6075 the `BIO_*printf` functions could overflow while calculating the length of a
6076 string and cause an OOB read when printing very long strings.
6078 Additionally the internal `doapr_outch` function can attempt to write to an
6079 OOB memory location (at an offset from the NULL pointer) in the event of a
6080 memory allocation failure. In 1.0.2 and below this could be caused where
6081 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6082 could be in processing a very long "%s" format string. Memory leaks can
6085 The first issue may mask the second issue dependent on compiler behaviour.
6086 These problems could enable attacks where large amounts of untrusted data
6087 is passed to the `BIO_*printf` functions. If applications use these functions
6088 in this way then they could be vulnerable. OpenSSL itself uses these
6089 functions when printing out human-readable dumps of ASN.1 data. Therefore
6090 applications that print this data could be vulnerable if the data is from
6091 untrusted sources. OpenSSL command line applications could also be
6092 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6093 as command line arguments.
6095 Libssl is not considered directly vulnerable. Additionally certificates etc
6096 received via remote connections via libssl are also unlikely to be able to
6097 trigger these issues because of message size limits enforced within libssl.
6099 This issue was reported to OpenSSL Guido Vranken.
6104 * Side channel attack on modular exponentiation
6106 A side-channel attack was found which makes use of cache-bank conflicts on
6107 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6108 of RSA keys. The ability to exploit this issue is limited as it relies on
6109 an attacker who has control of code in a thread running on the same
6110 hyper-threaded core as the victim thread which is performing decryptions.
6112 This issue was reported to OpenSSL by Yuval Yarom, The University of
6113 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6114 Nadia Heninger, University of Pennsylvania with more information at
6115 <http://cachebleed.info>.
6120 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6121 if no keysize is specified with default_bits. This fixes an
6122 omission in an earlier change that changed all RSA/DSA key generation
6123 commands to use 2048 bits by default.
6127 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6129 * DH small subgroups
6131 Historically OpenSSL only ever generated DH parameters based on "safe"
6132 primes. More recently (in version 1.0.2) support was provided for
6133 generating X9.42 style parameter files such as those required for RFC 5114
6134 support. The primes used in such files may not be "safe". Where an
6135 application is using DH configured with parameters based on primes that are
6136 not "safe" then an attacker could use this fact to find a peer's private
6137 DH exponent. This attack requires that the attacker complete multiple
6138 handshakes in which the peer uses the same private DH exponent. For example
6139 this could be used to discover a TLS server's private DH exponent if it's
6140 reusing the private DH exponent or it's using a static DH ciphersuite.
6142 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6143 TLS. It is not on by default. If the option is not set then the server
6144 reuses the same private DH exponent for the life of the server process and
6145 would be vulnerable to this attack. It is believed that many popular
6146 applications do set this option and would therefore not be at risk.
6148 The fix for this issue adds an additional check where a "q" parameter is
6149 available (as is the case in X9.42 based parameters). This detects the
6150 only known attack, and is the only possible defense for static DH
6151 ciphersuites. This could have some performance impact.
6153 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6154 default and cannot be disabled. This could have some performance impact.
6156 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6161 * SSLv2 doesn't block disabled ciphers
6163 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6164 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6165 been disabled, provided that the SSLv2 protocol was not also disabled via
6168 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6169 and Sebastian Schinzel.
6174 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6176 * BN_mod_exp may produce incorrect results on x86_64
6178 There is a carry propagating bug in the x86_64 Montgomery squaring
6179 procedure. No EC algorithms are affected. Analysis suggests that attacks
6180 against RSA and DSA as a result of this defect would be very difficult to
6181 perform and are not believed likely. Attacks against DH are considered just
6182 feasible (although very difficult) because most of the work necessary to
6183 deduce information about a private key may be performed offline. The amount
6184 of resources required for such an attack would be very significant and
6185 likely only accessible to a limited number of attackers. An attacker would
6186 additionally need online access to an unpatched system using the target
6187 private key in a scenario with persistent DH parameters and a private
6188 key that is shared between multiple clients. For example this can occur by
6189 default in OpenSSL DHE based SSL/TLS ciphersuites.
6191 This issue was reported to OpenSSL by Hanno Böck.
6196 * Certificate verify crash with missing PSS parameter
6198 The signature verification routines will crash with a NULL pointer
6199 dereference if presented with an ASN.1 signature using the RSA PSS
6200 algorithm and absent mask generation function parameter. Since these
6201 routines are used to verify certificate signature algorithms this can be
6202 used to crash any certificate verification operation and exploited in a
6203 DoS attack. Any application which performs certificate verification is
6204 vulnerable including OpenSSL clients and servers which enable client
6207 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6212 * X509_ATTRIBUTE memory leak
6214 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6215 memory. This structure is used by the PKCS#7 and CMS routines so any
6216 application which reads PKCS#7 or CMS data from untrusted sources is
6217 affected. SSL/TLS is not affected.
6219 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6225 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6226 This changes the decoding behaviour for some invalid messages,
6227 though the change is mostly in the more lenient direction, and
6228 legacy behaviour is preserved as much as possible.
6232 * In DSA_generate_parameters_ex, if the provided seed is too short,
6235 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6237 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6239 * Alternate chains certificate forgery
6241 During certificate verification, OpenSSL will attempt to find an
6242 alternative certificate chain if the first attempt to build such a chain
6243 fails. An error in the implementation of this logic can mean that an
6244 attacker could cause certain checks on untrusted certificates to be
6245 bypassed, such as the CA flag, enabling them to use a valid leaf
6246 certificate to act as a CA and "issue" an invalid certificate.
6248 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6253 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6255 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6256 incompatibility in the handling of HMAC. The previous ABI has now been
6261 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6263 * Malformed ECParameters causes infinite loop
6265 When processing an ECParameters structure OpenSSL enters an infinite loop
6266 if the curve specified is over a specially malformed binary polynomial
6269 This can be used to perform denial of service against any
6270 system which processes public keys, certificate requests or
6271 certificates. This includes TLS clients and TLS servers with
6272 client authentication enabled.
6274 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6279 * Exploitable out-of-bounds read in X509_cmp_time
6281 X509_cmp_time does not properly check the length of the ASN1_TIME
6282 string and can read a few bytes out of bounds. In addition,
6283 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6286 An attacker can use this to craft malformed certificates and CRLs of
6287 various sizes and potentially cause a segmentation fault, resulting in
6288 a DoS on applications that verify certificates or CRLs. TLS clients
6289 that verify CRLs are affected. TLS clients and servers with client
6290 authentication enabled may be affected if they use custom verification
6293 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6294 independently by Hanno Böck.
6299 * PKCS7 crash with missing EnvelopedContent
6301 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6302 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6303 with missing content and trigger a NULL pointer dereference on parsing.
6305 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6306 structures from untrusted sources are affected. OpenSSL clients and
6307 servers are not affected.
6309 This issue was reported to OpenSSL by Michal Zalewski (Google).
6314 * CMS verify infinite loop with unknown hash function
6316 When verifying a signedData message the CMS code can enter an infinite loop
6317 if presented with an unknown hash function OID. This can be used to perform
6318 denial of service against any system which verifies signedData messages using
6320 This issue was reported to OpenSSL by Johannes Bauer.
6325 * Race condition handling NewSessionTicket
6327 If a NewSessionTicket is received by a multi-threaded client when attempting to
6328 reuse a previous ticket then a race condition can occur potentially leading to
6329 a double free of the ticket data.
6334 * Only support 256-bit or stronger elliptic curves with the
6335 'ecdh_auto' setting (server) or by default (client). Of supported
6336 curves, prefer P-256 (both).
6340 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6342 * ClientHello sigalgs DoS fix
6344 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6345 invalid signature algorithms extension a NULL pointer dereference will
6346 occur. This can be exploited in a DoS attack against the server.
6348 This issue was was reported to OpenSSL by David Ramos of Stanford
6352 *Stephen Henson and Matt Caswell*
6354 * Multiblock corrupted pointer fix
6356 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6357 feature only applies on 64 bit x86 architecture platforms that support AES
6358 NI instructions. A defect in the implementation of "multiblock" can cause
6359 OpenSSL's internal write buffer to become incorrectly set to NULL when
6360 using non-blocking IO. Typically, when the user application is using a
6361 socket BIO for writing, this will only result in a failed connection.
6362 However if some other BIO is used then it is likely that a segmentation
6363 fault will be triggered, thus enabling a potential DoS attack.
6365 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6370 * Segmentation fault in DTLSv1_listen fix
6372 The DTLSv1_listen function is intended to be stateless and processes the
6373 initial ClientHello from many peers. It is common for user code to loop
6374 over the call to DTLSv1_listen until a valid ClientHello is received with
6375 an associated cookie. A defect in the implementation of DTLSv1_listen means
6376 that state is preserved in the SSL object from one invocation to the next
6377 that can lead to a segmentation fault. Errors processing the initial
6378 ClientHello can trigger this scenario. An example of such an error could be
6379 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6382 This issue was reported to OpenSSL by Per Allansson.
6387 * Segmentation fault in ASN1_TYPE_cmp fix
6389 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6390 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6391 certificate signature algorithm consistency this can be used to crash any
6392 certificate verification operation and exploited in a DoS attack. Any
6393 application which performs certificate verification is vulnerable including
6394 OpenSSL clients and servers which enable client authentication.
6399 * Segmentation fault for invalid PSS parameters fix
6401 The signature verification routines will crash with a NULL pointer
6402 dereference if presented with an ASN.1 signature using the RSA PSS
6403 algorithm and invalid parameters. Since these routines are used to verify
6404 certificate signature algorithms this can be used to crash any
6405 certificate verification operation and exploited in a DoS attack. Any
6406 application which performs certificate verification is vulnerable including
6407 OpenSSL clients and servers which enable client authentication.
6409 This issue was was reported to OpenSSL by Brian Carpenter.
6414 * ASN.1 structure reuse memory corruption fix
6416 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6417 memory corruption via an invalid write. Such reuse is and has been
6418 strongly discouraged and is believed to be rare.
6420 Applications that parse structures containing CHOICE or ANY DEFINED BY
6421 components may be affected. Certificate parsing (d2i_X509 and related
6422 functions) are however not affected. OpenSSL clients and servers are
6428 * PKCS7 NULL pointer dereferences fix
6430 The PKCS#7 parsing code does not handle missing outer ContentInfo
6431 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6432 missing content and trigger a NULL pointer dereference on parsing.
6434 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6435 otherwise parse PKCS#7 structures from untrusted sources are
6436 affected. OpenSSL clients and servers are not affected.
6438 This issue was reported to OpenSSL by Michal Zalewski (Google).
6443 * DoS via reachable assert in SSLv2 servers fix
6445 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6446 servers that both support SSLv2 and enable export cipher suites by sending
6447 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6449 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6450 (OpenSSL development team).
6455 * Empty CKE with client auth and DHE fix
6457 If client auth is used then a server can seg fault in the event of a DHE
6458 ciphersuite being selected and a zero length ClientKeyExchange message
6459 being sent by the client. This could be exploited in a DoS attack.
6464 * Handshake with unseeded PRNG fix
6466 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6467 with an unseeded PRNG. The conditions are:
6468 - The client is on a platform where the PRNG has not been seeded
6469 automatically, and the user has not seeded manually
6470 - A protocol specific client method version has been used (i.e. not
6471 SSL_client_methodv23)
6472 - A ciphersuite is used that does not require additional random data from
6473 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6475 If the handshake succeeds then the client random that has been used will
6476 have been generated from a PRNG with insufficient entropy and therefore the
6477 output may be predictable.
6479 For example using the following command with an unseeded openssl will
6480 succeed on an unpatched platform:
6482 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6487 * Use After Free following d2i_ECPrivatekey error fix
6489 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6490 could cause a use after free condition. This, in turn, could cause a double
6491 free in several private key parsing functions (such as d2i_PrivateKey
6492 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6493 for applications that receive EC private keys from untrusted
6494 sources. This scenario is considered rare.
6496 This issue was discovered by the BoringSSL project and fixed in their
6502 * X509_to_X509_REQ NULL pointer deref fix
6504 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6505 the certificate key is invalid. This function is rarely used in practice.
6507 This issue was discovered by Brian Carpenter.
6512 * Removed the export ciphers from the DEFAULT ciphers
6516 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6518 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6519 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6520 So far those who have to target multiple platforms would compromise
6521 and argue that binary targeting say ARMv5 would still execute on
6522 ARMv8. "Universal" build resolves this compromise by providing
6523 near-optimal performance even on newer platforms.
6527 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6528 (other platforms pending).
6530 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6532 * Add support for the SignedCertificateTimestampList certificate and
6533 OCSP response extensions from RFC6962.
6537 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6538 for corner cases. (Certain input points at infinity could lead to
6539 bogus results, with non-infinity inputs mapped to infinity too.)
6543 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6544 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6545 common cases are optimized and there still is room for further
6546 improvements. Vector Permutation AES for Altivec is also added.
6550 * Add support for little-endian ppc64 Linux target.
6552 *Marcelo Cerri (IBM)*
6554 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6555 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6556 are optimized and there still is room for further improvements.
6557 Both 32- and 64-bit modes are supported.
6559 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6561 * Improved ARMv7 NEON support.
6565 * Support for SPARC Architecture 2011 crypto extensions, first
6566 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6567 SHA256/512, MD5, GHASH and modular exponentiation.
6569 *Andy Polyakov, David Miller*
6571 * Accelerated modular exponentiation for Intel processors, a.k.a.
6574 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6576 * Support for new and upcoming Intel processors, including AVX2,
6577 BMI and SHA ISA extensions. This includes additional "stitched"
6578 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6581 This work was sponsored by Intel Corp.
6585 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6586 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6587 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6591 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6592 this fixes a limitation in previous versions of OpenSSL.
6596 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6597 MGF1 digest and OAEP label.
6601 * Add EVP support for key wrapping algorithms, to avoid problems with
6602 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6603 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6604 algorithms and include tests cases.
6608 * Add functions to allocate and set the fields of an ECDSA_METHOD
6611 *Douglas E. Engert, Steve Henson*
6613 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6614 difference in days and seconds between two tm or ASN1_TIME structures.
6618 * Add -rev test option to s_server to just reverse order of characters
6619 received by client and send back to server. Also prints an abbreviated
6620 summary of the connection parameters.
6624 * New option -brief for s_client and s_server to print out a brief summary
6625 of connection parameters.
6629 * Add callbacks for arbitrary TLS extensions.
6631 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6633 * New option -crl_download in several openssl utilities to download CRLs
6634 from CRLDP extension in certificates.
6638 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6642 * New function X509_CRL_diff to generate a delta CRL from the difference
6643 of two full CRLs. Add support to "crl" utility.
6647 * New functions to set lookup_crls function and to retrieve
6648 X509_STORE from X509_STORE_CTX.
6652 * Print out deprecated issuer and subject unique ID fields in
6657 * Extend OCSP I/O functions so they can be used for simple general purpose
6658 HTTP as well as OCSP. New wrapper function which can be used to download
6659 CRLs using the OCSP API.
6663 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6667 * `SSL_CONF*` functions. These provide a common framework for application
6668 configuration using configuration files or command lines.
6672 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6673 message callback and prints the results. Needs compile time option
6674 "enable-ssl-trace". New options to s_client and s_server to enable
6679 * New ctrl and macro to retrieve supported points extensions.
6680 Print out extension in s_server and s_client.
6684 * New functions to retrieve certificate signature and signature
6689 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6694 * New Suite B modes for TLS code. These use and enforce the requirements
6695 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6696 only use Suite B curves. The Suite B modes can be set by using the
6697 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6701 * New chain verification flags for Suite B levels of security. Check
6702 algorithms are acceptable when flags are set in X509_verify_cert.
6706 * Make tls1_check_chain return a set of flags indicating checks passed
6707 by a certificate chain. Add additional tests to handle client
6708 certificates: checks for matching certificate type and issuer name
6713 * If an attempt is made to use a signature algorithm not in the peer
6714 preference list abort the handshake. If client has no suitable
6715 signature algorithms in response to a certificate request do not
6716 use the certificate.
6720 * If server EC tmp key is not in client preference list abort handshake.
6724 * Add support for certificate stores in CERT structure. This makes it
6725 possible to have different stores per SSL structure or one store in
6726 the parent SSL_CTX. Include distinct stores for certificate chain
6727 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6728 to build and store a certificate chain in CERT structure: returning
6729 an error if the chain cannot be built: this will allow applications
6730 to test if a chain is correctly configured.
6732 Note: if the CERT based stores are not set then the parent SSL_CTX
6733 store is used to retain compatibility with existing behaviour.
6737 * New function ssl_set_client_disabled to set a ciphersuite disabled
6738 mask based on the current session, check mask when sending client
6739 hello and checking the requested ciphersuite.
6743 * New ctrls to retrieve and set certificate types in a certificate
6744 request message. Print out received values in s_client. If certificate
6745 types is not set with custom values set sensible values based on
6746 supported signature algorithms.
6750 * Support for distinct client and server supported signature algorithms.
6754 * Add certificate callback. If set this is called whenever a certificate
6755 is required by client or server. An application can decide which
6756 certificate chain to present based on arbitrary criteria: for example
6757 supported signature algorithms. Add very simple example to s_server.
6758 This fixes many of the problems and restrictions of the existing client
6759 certificate callback: for example you can now clear an existing
6760 certificate and specify the whole chain.
6764 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6765 the certificate can be used for (if anything). Set valid_flags field
6766 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6767 to have similar checks in it.
6769 Add new "cert_flags" field to CERT structure and include a "strict mode".
6770 This enforces some TLS certificate requirements (such as only permitting
6771 certificate signature algorithms contained in the supported algorithms
6772 extension) which some implementations ignore: this option should be used
6773 with caution as it could cause interoperability issues.
6777 * Update and tidy signature algorithm extension processing. Work out
6778 shared signature algorithms based on preferences and peer algorithms
6779 and print them out in s_client and s_server. Abort handshake if no
6780 shared signature algorithms.
6784 * Add new functions to allow customised supported signature algorithms
6785 for SSL and SSL_CTX structures. Add options to s_client and s_server
6790 * New function SSL_certs_clear() to delete all references to certificates
6791 from an SSL structure. Before this once a certificate had been added
6792 it couldn't be removed.
6796 * Integrate hostname, email address and IP address checking with certificate
6797 verification. New verify options supporting checking in openssl utility.
6801 * Fixes and wildcard matching support to hostname and email checking
6802 functions. Add manual page.
6804 *Florian Weimer (Red Hat Product Security Team)*
6806 * New functions to check a hostname email or IP address against a
6807 certificate. Add options x509 utility to print results of checks against
6812 * Fix OCSP checking.
6814 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6816 * Initial experimental support for explicitly trusted non-root CAs.
6817 OpenSSL still tries to build a complete chain to a root but if an
6818 intermediate CA has a trust setting included that is used. The first
6819 setting is used: whether to trust (e.g., -addtrust option to the x509
6824 * Add -trusted_first option which attempts to find certificates in the
6825 trusted store even if an untrusted chain is also supplied.
6829 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6830 platform support for Linux and Android.
6834 * Support for linux-x32, ILP32 environment in x86_64 framework.
6838 * Experimental multi-implementation support for FIPS capable OpenSSL.
6839 When in FIPS mode the approved implementations are used as normal,
6840 when not in FIPS mode the internal unapproved versions are used instead.
6841 This means that the FIPS capable OpenSSL isn't forced to use the
6842 (often lower performance) FIPS implementations outside FIPS mode.
6846 * Transparently support X9.42 DH parameters when calling
6847 PEM_read_bio_DHparameters. This means existing applications can handle
6848 the new parameter format automatically.
6852 * Initial experimental support for X9.42 DH parameter format: mainly
6853 to support use of 'q' parameter for RFC5114 parameters.
6857 * Add DH parameters from RFC5114 including test data to dhtest.
6861 * Support for automatic EC temporary key parameter selection. If enabled
6862 the most preferred EC parameters are automatically used instead of
6863 hardcoded fixed parameters. Now a server just has to call:
6864 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6865 support ECDH and use the most appropriate parameters.
6869 * Enhance and tidy EC curve and point format TLS extension code. Use
6870 static structures instead of allocation if default values are used.
6871 New ctrls to set curves we wish to support and to retrieve shared curves.
6872 Print out shared curves in s_server. New options to s_server and s_client
6873 to set list of supported curves.
6877 * New ctrls to retrieve supported signature algorithms and
6878 supported curve values as an array of NIDs. Extend openssl utility
6879 to print out received values.
6883 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6884 between NIDs and the more common NIST names such as "P-256". Enhance
6885 ecparam utility and ECC method to recognise the NIST names for curves.
6889 * Enhance SSL/TLS certificate chain handling to support different
6890 chains for each certificate instead of one chain in the parent SSL_CTX.
6894 * Support for fixed DH ciphersuite client authentication: where both
6895 server and client use DH certificates with common parameters.
6899 * Support for fixed DH ciphersuites: those requiring DH server
6904 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6906 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6907 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6908 X509_CINF_get_signature were reverted post internal team review.
6913 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6915 * OCSP Status Request extension unbounded memory growth
6917 A malicious client can send an excessively large OCSP Status Request
6918 extension. If that client continually requests renegotiation, sending a
6919 large OCSP Status Request extension each time, then there will be unbounded
6920 memory growth on the server. This will eventually lead to a Denial Of
6921 Service attack through memory exhaustion. Servers with a default
6922 configuration are vulnerable even if they do not support OCSP. Builds using
6923 the "no-ocsp" build time option are not affected.
6925 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6930 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6933 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6939 * OOB write in MDC2_Update()
6941 An overflow can occur in MDC2_Update() either if called directly or
6942 through the EVP_DigestUpdate() function using MDC2. If an attacker
6943 is able to supply very large amounts of input data after a previous
6944 call to EVP_EncryptUpdate() with a partial block then a length check
6945 can overflow resulting in a heap corruption.
6947 The amount of data needed is comparable to SIZE_MAX which is impractical
6950 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6955 * Malformed SHA512 ticket DoS
6957 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6958 DoS attack where a malformed ticket will result in an OOB read which will
6961 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6962 a custom server callback and ticket lookup mechanism.
6964 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6969 * OOB write in BN_bn2dec()
6971 The function BN_bn2dec() does not check the return value of BN_div_word().
6972 This can cause an OOB write if an application uses this function with an
6973 overly large BIGNUM. This could be a problem if an overly large certificate
6974 or CRL is printed out from an untrusted source. TLS is not affected because
6975 record limits will reject an oversized certificate before it is parsed.
6977 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6982 * OOB read in TS_OBJ_print_bio()
6984 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6985 the total length the OID text representation would use and not the amount
6986 of data written. This will result in OOB reads when large OIDs are
6989 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6994 * Pointer arithmetic undefined behaviour
6996 Avoid some undefined pointer arithmetic
6998 A common idiom in the codebase is to check limits in the following manner:
7001 Where "p" points to some malloc'd data of SIZE bytes and
7004 "len" here could be from some externally supplied data (e.g. from a TLS
7007 The rules of C pointer arithmetic are such that "p + len" is only well
7008 defined where len <= SIZE. Therefore, the above idiom is actually
7009 undefined behaviour.
7011 For example this could cause problems if some malloc implementation
7012 provides an address for "p" such that "p + len" actually overflows for
7013 values of len that are too big and therefore p + len < limit.
7015 This issue was reported to OpenSSL by Guido Vranken
7020 * Constant time flag not preserved in DSA signing
7022 Operations in the DSA signing algorithm should run in constant time in
7023 order to avoid side channel attacks. A flaw in the OpenSSL DSA
7024 implementation means that a non-constant time codepath is followed for
7025 certain operations. This has been demonstrated through a cache-timing
7026 attack to be sufficient for an attacker to recover the private DSA key.
7028 This issue was reported by César Pereida (Aalto University), Billy Brumley
7029 (Tampere University of Technology), and Yuval Yarom (The University of
7030 Adelaide and NICTA).
7035 * DTLS buffered message DoS
7037 In a DTLS connection where handshake messages are delivered out-of-order
7038 those messages that OpenSSL is not yet ready to process will be buffered
7039 for later use. Under certain circumstances, a flaw in the logic means that
7040 those messages do not get removed from the buffer even though the handshake
7041 has been completed. An attacker could force up to approx. 15 messages to
7042 remain in the buffer when they are no longer required. These messages will
7043 be cleared when the DTLS connection is closed. The default maximum size for
7044 a message is 100k. Therefore, the attacker could force an additional 1500k
7045 to be consumed per connection. By opening many simultaneous connections an
7046 attacker could cause a DoS attack through memory exhaustion.
7048 This issue was reported to OpenSSL by Quan Luo.
7053 * DTLS replay protection DoS
7055 A flaw in the DTLS replay attack protection mechanism means that records
7056 that arrive for future epochs update the replay protection "window" before
7057 the MAC for the record has been validated. This could be exploited by an
7058 attacker by sending a record for the next epoch (which does not have to
7059 decrypt or have a valid MAC), with a very large sequence number. This means
7060 that all subsequent legitimate packets are dropped causing a denial of
7061 service for a specific DTLS connection.
7063 This issue was reported to OpenSSL by the OCAP audit team.
7068 * Certificate message OOB reads
7070 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7071 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7072 theoretical DoS risk but this has not been observed in practice on common
7075 The messages affected are client certificate, client certificate request
7076 and server certificate. As a result the attack can only be performed
7077 against a client or a server which enables client authentication.
7079 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7084 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7086 * Prevent padding oracle in AES-NI CBC MAC check
7088 A MITM attacker can use a padding oracle attack to decrypt traffic
7089 when the connection uses an AES CBC cipher and the server support
7092 This issue was introduced as part of the fix for Lucky 13 padding
7093 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7094 constant time by making sure that always the same bytes are read and
7095 compared against either the MAC or padding bytes. But it no longer
7096 checked that there was enough data to have both the MAC and padding
7099 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7104 * Fix EVP_EncodeUpdate overflow
7106 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7107 Base64 encoding of binary data. If an attacker is able to supply very large
7108 amounts of input data then a length check can overflow resulting in a heap
7111 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7112 the `PEM_write_bio*` family of functions. These are mainly used within the
7113 OpenSSL command line applications, so any application which processes data
7114 from an untrusted source and outputs it as a PEM file should be considered
7115 vulnerable to this issue. User applications that call these APIs directly
7116 with large amounts of untrusted data may also be vulnerable.
7118 This issue was reported by Guido Vranken.
7123 * Fix EVP_EncryptUpdate overflow
7125 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7126 is able to supply very large amounts of input data after a previous call to
7127 EVP_EncryptUpdate() with a partial block then a length check can overflow
7128 resulting in a heap corruption. Following an analysis of all OpenSSL
7129 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7130 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7131 the first called function after an EVP_EncryptInit(), and therefore that
7132 specific call must be safe. The second form is where the length passed to
7133 EVP_EncryptUpdate() can be seen from the code to be some small value and
7134 therefore there is no possibility of an overflow. Since all instances are
7135 one of these two forms, it is believed that there can be no overflows in
7136 internal code due to this problem. It should be noted that
7137 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7138 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7139 of these calls have also been analysed too and it is believed there are no
7140 instances in internal usage where an overflow could occur.
7142 This issue was reported by Guido Vranken.
7147 * Prevent ASN.1 BIO excessive memory allocation
7149 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7150 a short invalid encoding can casuse allocation of large amounts of memory
7151 potentially consuming excessive resources or exhausting memory.
7153 Any application parsing untrusted data through d2i BIO functions is
7154 affected. The memory based functions such as d2i_X509() are *not* affected.
7155 Since the memory based functions are used by the TLS library, TLS
7156 applications are not affected.
7158 This issue was reported by Brian Carpenter.
7165 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7166 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7167 in arbitrary stack data being returned in the buffer.
7169 This issue was reported by Guido Vranken.
7174 * Modify behavior of ALPN to invoke callback after SNI/servername
7175 callback, such that updates to the SSL_CTX affect ALPN.
7179 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7184 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7185 methods are enabled and ssl2 is disabled the methods return NULL.
7189 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7191 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7192 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7193 provide any "EXPORT" or "LOW" strength ciphers.
7197 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7198 is by default disabled at build-time. Builds that are not configured with
7199 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7200 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7201 will need to explicitly call either of:
7203 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7205 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7207 as appropriate. Even if either of those is used, or the application
7208 explicitly uses the version-specific SSLv2_method() or its client and
7209 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7210 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7211 ciphers, and SSLv2 56-bit DES are no longer available.
7216 * Fix a double-free in DSA code
7218 A double free bug was discovered when OpenSSL parses malformed DSA private
7219 keys and could lead to a DoS attack or memory corruption for applications
7220 that receive DSA private keys from untrusted sources. This scenario is
7223 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7229 * Disable SRP fake user seed to address a server memory leak.
7231 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7233 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7234 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7235 was changed to ignore the "fake user" SRP seed, even if the seed
7238 Users should use SRP_VBASE_get1_by_user instead. Note that in
7239 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7240 also that even though configuring the SRP seed attempts to hide
7241 invalid usernames by continuing the handshake with fake
7242 credentials, this behaviour is not constant time and no strong
7243 guarantees are made that the handshake is indistinguishable from
7244 that of a valid user.
7249 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7251 In the BN_hex2bn function the number of hex digits is calculated using an
7252 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7253 large values of `i` this can result in `bn_expand` not allocating any
7254 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7255 field as NULL leading to a subsequent NULL ptr deref. For very large values
7256 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7257 In this case memory is allocated to the internal BIGNUM data field, but it
7258 is insufficiently sized leading to heap corruption. A similar issue exists
7259 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7260 is ever called by user applications with very large untrusted hex/dec data.
7261 This is anticipated to be a rare occurrence.
7263 All OpenSSL internal usage of these functions use data that is not expected
7264 to be untrusted, e.g. config file data or application command line
7265 arguments. If user developed applications generate config file data based
7266 on untrusted data then it is possible that this could also lead to security
7267 consequences. This is also anticipated to be rare.
7269 This issue was reported to OpenSSL by Guido Vranken.
7274 * Fix memory issues in `BIO_*printf` functions
7276 The internal `fmtstr` function used in processing a "%s" format string in
7277 the `BIO_*printf` functions could overflow while calculating the length of a
7278 string and cause an OOB read when printing very long strings.
7280 Additionally the internal `doapr_outch` function can attempt to write to an
7281 OOB memory location (at an offset from the NULL pointer) in the event of a
7282 memory allocation failure. In 1.0.2 and below this could be caused where
7283 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7284 could be in processing a very long "%s" format string. Memory leaks can
7287 The first issue may mask the second issue dependent on compiler behaviour.
7288 These problems could enable attacks where large amounts of untrusted data
7289 is passed to the `BIO_*printf` functions. If applications use these functions
7290 in this way then they could be vulnerable. OpenSSL itself uses these
7291 functions when printing out human-readable dumps of ASN.1 data. Therefore
7292 applications that print this data could be vulnerable if the data is from
7293 untrusted sources. OpenSSL command line applications could also be
7294 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7295 as command line arguments.
7297 Libssl is not considered directly vulnerable. Additionally certificates etc
7298 received via remote connections via libssl are also unlikely to be able to
7299 trigger these issues because of message size limits enforced within libssl.
7301 This issue was reported to OpenSSL Guido Vranken.
7306 * Side channel attack on modular exponentiation
7308 A side-channel attack was found which makes use of cache-bank conflicts on
7309 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7310 of RSA keys. The ability to exploit this issue is limited as it relies on
7311 an attacker who has control of code in a thread running on the same
7312 hyper-threaded core as the victim thread which is performing decryptions.
7314 This issue was reported to OpenSSL by Yuval Yarom, The University of
7315 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7316 Nadia Heninger, University of Pennsylvania with more information at
7317 <http://cachebleed.info>.
7322 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7323 if no keysize is specified with default_bits. This fixes an
7324 omission in an earlier change that changed all RSA/DSA key generation
7325 commands to use 2048 bits by default.
7329 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7331 * Protection for DH small subgroup attacks
7333 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7334 switched on by default and cannot be disabled. This could have some
7339 * SSLv2 doesn't block disabled ciphers
7341 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7342 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7343 been disabled, provided that the SSLv2 protocol was not also disabled via
7346 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7347 and Sebastian Schinzel.
7352 * Reject DH handshakes with parameters shorter than 1024 bits.
7356 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7358 * Certificate verify crash with missing PSS parameter
7360 The signature verification routines will crash with a NULL pointer
7361 dereference if presented with an ASN.1 signature using the RSA PSS
7362 algorithm and absent mask generation function parameter. Since these
7363 routines are used to verify certificate signature algorithms this can be
7364 used to crash any certificate verification operation and exploited in a
7365 DoS attack. Any application which performs certificate verification is
7366 vulnerable including OpenSSL clients and servers which enable client
7369 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7374 * X509_ATTRIBUTE memory leak
7376 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7377 memory. This structure is used by the PKCS#7 and CMS routines so any
7378 application which reads PKCS#7 or CMS data from untrusted sources is
7379 affected. SSL/TLS is not affected.
7381 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7387 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7388 This changes the decoding behaviour for some invalid messages,
7389 though the change is mostly in the more lenient direction, and
7390 legacy behaviour is preserved as much as possible.
7394 * In DSA_generate_parameters_ex, if the provided seed is too short,
7395 use a random seed, as already documented.
7397 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7399 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7401 * Alternate chains certificate forgery
7403 During certificate verfification, OpenSSL will attempt to find an
7404 alternative certificate chain if the first attempt to build such a chain
7405 fails. An error in the implementation of this logic can mean that an
7406 attacker could cause certain checks on untrusted certificates to be
7407 bypassed, such as the CA flag, enabling them to use a valid leaf
7408 certificate to act as a CA and "issue" an invalid certificate.
7410 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7416 * Race condition handling PSK identify hint
7418 If PSK identity hints are received by a multi-threaded client then
7419 the values are wrongly updated in the parent SSL_CTX structure. This can
7420 result in a race condition potentially leading to a double free of the
7426 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7428 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7429 incompatibility in the handling of HMAC. The previous ABI has now been
7432 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7434 * Malformed ECParameters causes infinite loop
7436 When processing an ECParameters structure OpenSSL enters an infinite loop
7437 if the curve specified is over a specially malformed binary polynomial
7440 This can be used to perform denial of service against any
7441 system which processes public keys, certificate requests or
7442 certificates. This includes TLS clients and TLS servers with
7443 client authentication enabled.
7445 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7450 * Exploitable out-of-bounds read in X509_cmp_time
7452 X509_cmp_time does not properly check the length of the ASN1_TIME
7453 string and can read a few bytes out of bounds. In addition,
7454 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7457 An attacker can use this to craft malformed certificates and CRLs of
7458 various sizes and potentially cause a segmentation fault, resulting in
7459 a DoS on applications that verify certificates or CRLs. TLS clients
7460 that verify CRLs are affected. TLS clients and servers with client
7461 authentication enabled may be affected if they use custom verification
7464 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7465 independently by Hanno Böck.
7470 * PKCS7 crash with missing EnvelopedContent
7472 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7473 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7474 with missing content and trigger a NULL pointer dereference on parsing.
7476 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7477 structures from untrusted sources are affected. OpenSSL clients and
7478 servers are not affected.
7480 This issue was reported to OpenSSL by Michal Zalewski (Google).
7485 * CMS verify infinite loop with unknown hash function
7487 When verifying a signedData message the CMS code can enter an infinite loop
7488 if presented with an unknown hash function OID. This can be used to perform
7489 denial of service against any system which verifies signedData messages using
7491 This issue was reported to OpenSSL by Johannes Bauer.
7496 * Race condition handling NewSessionTicket
7498 If a NewSessionTicket is received by a multi-threaded client when attempting to
7499 reuse a previous ticket then a race condition can occur potentially leading to
7500 a double free of the ticket data.
7505 * Reject DH handshakes with parameters shorter than 768 bits.
7507 *Kurt Roeckx and Emilia Kasper*
7509 * dhparam: generate 2048-bit parameters by default.
7511 *Kurt Roeckx and Emilia Kasper*
7513 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7515 * Segmentation fault in ASN1_TYPE_cmp fix
7517 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7518 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7519 certificate signature algorithm consistency this can be used to crash any
7520 certificate verification operation and exploited in a DoS attack. Any
7521 application which performs certificate verification is vulnerable including
7522 OpenSSL clients and servers which enable client authentication.
7527 * ASN.1 structure reuse memory corruption fix
7529 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7530 memory corruption via an invalid write. Such reuse is and has been
7531 strongly discouraged and is believed to be rare.
7533 Applications that parse structures containing CHOICE or ANY DEFINED BY
7534 components may be affected. Certificate parsing (d2i_X509 and related
7535 functions) are however not affected. OpenSSL clients and servers are
7541 * PKCS7 NULL pointer dereferences fix
7543 The PKCS#7 parsing code does not handle missing outer ContentInfo
7544 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7545 missing content and trigger a NULL pointer dereference on parsing.
7547 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7548 otherwise parse PKCS#7 structures from untrusted sources are
7549 affected. OpenSSL clients and servers are not affected.
7551 This issue was reported to OpenSSL by Michal Zalewski (Google).
7556 * DoS via reachable assert in SSLv2 servers fix
7558 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7559 servers that both support SSLv2 and enable export cipher suites by sending
7560 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7562 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7563 (OpenSSL development team).
7568 * Use After Free following d2i_ECPrivatekey error fix
7570 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7571 could cause a use after free condition. This, in turn, could cause a double
7572 free in several private key parsing functions (such as d2i_PrivateKey
7573 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7574 for applications that receive EC private keys from untrusted
7575 sources. This scenario is considered rare.
7577 This issue was discovered by the BoringSSL project and fixed in their
7583 * X509_to_X509_REQ NULL pointer deref fix
7585 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7586 the certificate key is invalid. This function is rarely used in practice.
7588 This issue was discovered by Brian Carpenter.
7593 * Removed the export ciphers from the DEFAULT ciphers
7597 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7599 * Build fixes for the Windows and OpenVMS platforms
7601 *Matt Caswell and Richard Levitte*
7603 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7605 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7606 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7607 dereference. This could lead to a Denial Of Service attack. Thanks to
7608 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7613 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7614 dtls1_buffer_record function under certain conditions. In particular this
7615 could occur if an attacker sent repeated DTLS records with the same
7616 sequence number but for the next epoch. The memory leak could be exploited
7617 by an attacker in a Denial of Service attack through memory exhaustion.
7618 Thanks to Chris Mueller for reporting this issue.
7623 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7624 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7625 method would be set to NULL which could later result in a NULL pointer
7626 dereference. Thanks to Frank Schmirler for reporting this issue.
7631 * Abort handshake if server key exchange message is omitted for ephemeral
7634 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7635 reporting this issue.
7640 * Remove non-export ephemeral RSA code on client and server. This code
7641 violated the TLS standard by allowing the use of temporary RSA keys in
7642 non-export ciphersuites and could be used by a server to effectively
7643 downgrade the RSA key length used to a value smaller than the server
7644 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7645 INRIA or reporting this issue.
7650 * Fixed issue where DH client certificates are accepted without verification.
7651 An OpenSSL server will accept a DH certificate for client authentication
7652 without the certificate verify message. This effectively allows a client to
7653 authenticate without the use of a private key. This only affects servers
7654 which trust a client certificate authority which issues certificates
7655 containing DH keys: these are extremely rare and hardly ever encountered.
7656 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7662 * Ensure that the session ID context of an SSL is updated when its
7663 SSL_CTX is updated via SSL_set_SSL_CTX.
7665 The session ID context is typically set from the parent SSL_CTX,
7666 and can vary with the CTX.
7670 * Fix various certificate fingerprint issues.
7672 By using non-DER or invalid encodings outside the signed portion of a
7673 certificate the fingerprint can be changed without breaking the signature.
7674 Although no details of the signed portion of the certificate can be changed
7675 this can cause problems with some applications: e.g. those using the
7676 certificate fingerprint for blacklists.
7678 1. Reject signatures with non zero unused bits.
7680 If the BIT STRING containing the signature has non zero unused bits reject
7681 the signature. All current signature algorithms require zero unused bits.
7683 2. Check certificate algorithm consistency.
7685 Check the AlgorithmIdentifier inside TBS matches the one in the
7686 certificate signature. NB: this will result in signature failure
7687 errors for some broken certificates.
7689 Thanks to Konrad Kraszewski from Google for reporting this issue.
7691 3. Check DSA/ECDSA signatures use DER.
7693 Re-encode DSA/ECDSA signatures and compare with the original received
7694 signature. Return an error if there is a mismatch.
7696 This will reject various cases including garbage after signature
7697 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7698 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7699 (negative or with leading zeroes).
7701 Further analysis was conducted and fixes were developed by Stephen Henson
7702 of the OpenSSL core team.
7708 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7709 results on some platforms, including x86_64. This bug occurs at random
7710 with a very low probability, and is not known to be exploitable in any
7711 way, though its exact impact is difficult to determine. Thanks to Pieter
7712 Wuille (Blockstream) who reported this issue and also suggested an initial
7713 fix. Further analysis was conducted by the OpenSSL development team and
7714 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7715 the OpenSSL core team.
7720 * Do not resume sessions on the server if the negotiated protocol
7721 version does not match the session's version. Resuming with a different
7722 version, while not strictly forbidden by the RFC, is of questionable
7723 sanity and breaks all known clients.
7725 *David Benjamin, Emilia Käsper*
7727 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7728 early CCS messages during renegotiation. (Note that because
7729 renegotiation is encrypted, this early CCS was not exploitable.)
7733 * Tighten client-side session ticket handling during renegotiation:
7734 ensure that the client only accepts a session ticket if the server sends
7735 the extension anew in the ServerHello. Previously, a TLS client would
7736 reuse the old extension state and thus accept a session ticket if one was
7737 announced in the initial ServerHello.
7739 Similarly, ensure that the client requires a session ticket if one
7740 was advertised in the ServerHello. Previously, a TLS client would
7741 ignore a missing NewSessionTicket message.
7745 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7749 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7750 sends a carefully crafted handshake message, to cause OpenSSL to fail
7751 to free up to 64k of memory causing a memory leak. This could be
7752 exploited in a Denial Of Service attack. This issue affects OpenSSL
7753 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7754 whether SRTP is used or configured. Implementations of OpenSSL that
7755 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7757 The fix was developed by the OpenSSL team.
7762 * Session Ticket Memory Leak.
7764 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7765 integrity of that ticket is first verified. In the event of a session
7766 ticket integrity check failing, OpenSSL will fail to free memory
7767 causing a memory leak. By sending a large number of invalid session
7768 tickets an attacker could exploit this issue in a Denial Of Service
7774 * Build option no-ssl3 is incomplete.
7776 When OpenSSL is configured with "no-ssl3" as a build option, servers
7777 could accept and complete an SSL 3.0 handshake, and clients could be
7778 configured to send them.
7781 *Akamai and the OpenSSL team*
7783 * Add support for TLS_FALLBACK_SCSV.
7784 Client applications doing fallback retries should call
7785 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7788 *Adam Langley, Bodo Moeller*
7790 * Add additional DigestInfo checks.
7792 Re-encode DigestInto in DER and check against the original when
7793 verifying RSA signature: this will reject any improperly encoded
7794 DigestInfo structures.
7796 Note: this is a precautionary measure and no attacks are currently known.
7800 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7802 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7803 SRP code can be overrun an internal buffer. Add sanity check that
7804 g, A, B < N to SRP code.
7806 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7807 Group for discovering this issue.
7812 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7813 TLS 1.0 instead of higher protocol versions when the ClientHello message
7814 is badly fragmented. This allows a man-in-the-middle attacker to force a
7815 downgrade to TLS 1.0 even if both the server and the client support a
7816 higher protocol version, by modifying the client's TLS records.
7818 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7819 researching this issue.
7824 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7825 to a denial of service attack. A malicious server can crash the client
7826 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7827 ciphersuite and sending carefully crafted handshake messages.
7829 Thanks to Felix Gröbert (Google) for discovering and researching this
7835 * By sending carefully crafted DTLS packets an attacker could cause openssl
7836 to leak memory. This can be exploited through a Denial of Service attack.
7837 Thanks to Adam Langley for discovering and researching this issue.
7842 * An attacker can force openssl to consume large amounts of memory whilst
7843 processing DTLS handshake messages. This can be exploited through a
7844 Denial of Service attack.
7845 Thanks to Adam Langley for discovering and researching this issue.
7850 * An attacker can force an error condition which causes openssl to crash
7851 whilst processing DTLS packets due to memory being freed twice. This
7852 can be exploited through a Denial of Service attack.
7853 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7859 * If a multithreaded client connects to a malicious server using a resumed
7860 session and the server sends an ec point format extension it could write
7861 up to 255 bytes to freed memory.
7863 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7869 * A malicious server can crash an OpenSSL client with a null pointer
7870 dereference (read) by specifying an SRP ciphersuite even though it was not
7871 properly negotiated with the client. This can be exploited through a
7872 Denial of Service attack.
7874 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7875 discovering and researching this issue.
7880 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7881 X509_name_oneline, X509_name_print_ex et al. to leak some information
7882 from the stack. Applications may be affected if they echo pretty printing
7883 output to the attacker.
7885 Thanks to Ivan Fratric (Google) for discovering this issue.
7888 *Emilia Käsper, and Steve Henson*
7890 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7891 for corner cases. (Certain input points at infinity could lead to
7892 bogus results, with non-infinity inputs mapped to infinity too.)
7896 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7898 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7899 handshake can force the use of weak keying material in OpenSSL
7900 SSL/TLS clients and servers.
7902 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7903 researching this issue. ([CVE-2014-0224])
7905 *KIKUCHI Masashi, Steve Henson*
7907 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7908 OpenSSL DTLS client the code can be made to recurse eventually crashing
7911 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7914 *Imre Rad, Steve Henson*
7916 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7917 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7918 client or server. This is potentially exploitable to run arbitrary
7919 code on a vulnerable client or server.
7921 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7923 *Jüri Aedla, Steve Henson*
7925 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7926 are subject to a denial of service attack.
7928 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7929 this issue. ([CVE-2014-3470])
7931 *Felix Gröbert, Ivan Fratric, Steve Henson*
7933 * Harmonize version and its documentation. -f flag is used to display
7936 *mancha <mancha1@zoho.com>*
7938 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7939 in i2d_ECPrivateKey.
7941 *mancha <mancha1@zoho.com>*
7943 * Fix some double frees. These are not thought to be exploitable.
7945 *mancha <mancha1@zoho.com>*
7947 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7949 * A missing bounds check in the handling of the TLS heartbeat extension
7950 can be used to reveal up to 64k of memory to a connected client or
7953 Thanks for Neel Mehta of Google Security for discovering this bug and to
7954 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7955 preparing the fix ([CVE-2014-0160])
7957 *Adam Langley, Bodo Moeller*
7959 * Fix for the attack described in the paper "Recovering OpenSSL
7960 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7961 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7962 <http://eprint.iacr.org/2014/140>
7964 Thanks to Yuval Yarom and Naomi Benger for discovering this
7965 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7967 *Yuval Yarom and Naomi Benger*
7969 * TLS pad extension: draft-agl-tls-padding-03
7971 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7972 TLS client Hello record length value would otherwise be > 255 and
7973 less that 512 pad with a dummy extension containing zeroes so it
7974 is at least 512 bytes long.
7976 *Adam Langley, Steve Henson*
7978 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7980 * Fix for TLS record tampering bug. A carefully crafted invalid
7981 handshake could crash OpenSSL with a NULL pointer exception.
7982 Thanks to Anton Johansson for reporting this issues.
7985 * Keep original DTLS digest and encryption contexts in retransmission
7986 structures so we can use the previous session parameters if they need
7987 to be resent. ([CVE-2013-6450])
7991 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7992 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7993 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7994 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7995 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7996 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7998 *Rob Stradling, Adam Langley*
8000 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
8002 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
8003 supporting platforms or when small records were transferred.
8005 *Andy Polyakov, Steve Henson*
8007 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
8009 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8011 This addresses the flaw in CBC record processing discovered by
8012 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8013 at: <http://www.isg.rhul.ac.uk/tls/>
8015 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8016 Security Group at Royal Holloway, University of London
8017 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8018 Emilia Käsper for the initial patch.
8021 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8023 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
8024 ciphersuites which can be exploited in a denial of service attack.
8025 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
8026 and detecting this bug and to Wolfgang Ettlinger
8027 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
8032 * Return an error when checking OCSP signatures when key is NULL.
8033 This fixes a DoS attack. ([CVE-2013-0166])
8037 * Make openssl verify return errors.
8039 *Chris Palmer <palmer@google.com> and Ben Laurie*
8041 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8042 the right response is stapled. Also change SSL_get_certificate()
8043 so it returns the certificate actually sent.
8044 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8046 *Rob Stradling <rob.stradling@comodo.com>*
8048 * Fix possible deadlock when decoding public keys.
8052 * Don't use TLS 1.0 record version number in initial client hello
8057 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
8059 * Sanity check record length before skipping explicit IV in TLS
8060 1.2, 1.1 and DTLS to fix DoS attack.
8062 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8063 fuzzing as a service testing platform.
8068 * Initialise tkeylen properly when encrypting CMS messages.
8069 Thanks to Solar Designer of Openwall for reporting this issue.
8073 * In FIPS mode don't try to use composite ciphers as they are not
8078 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8080 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8081 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8082 mean any application compiled against OpenSSL 1.0.0 headers setting
8083 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8084 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8085 0x10000000L Any application which was previously compiled against
8086 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8087 will need to be recompiled as a result. Letting be results in
8088 inability to disable specifically TLS 1.1 and in client context,
8089 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8093 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8094 disable just protocol X, but all protocols above X *if* there are
8095 protocols *below* X still enabled. In more practical terms it means
8096 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8097 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8098 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8103 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8105 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8106 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8107 in CRYPTO_realloc_clean.
8109 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8110 issue and to Adam Langley <agl@chromium.org> for fixing it.
8113 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8115 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8119 * Workarounds for some broken servers that "hang" if a client hello
8120 record length exceeds 255 bytes.
8122 1. Do not use record version number > TLS 1.0 in initial client
8123 hello: some (but not all) hanging servers will now work.
8124 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8125 the number of ciphers sent in the client hello. This should be
8126 set to an even number, such as 50, for example by passing:
8127 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8128 Most broken servers should now work.
8129 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8130 TLS 1.2 client support entirely.
8134 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8138 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8140 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8141 STRING form instead of a DigestInfo.
8145 * The format used for MDC2 RSA signatures is inconsistent between EVP
8146 and the RSA_sign/RSA_verify functions. This was made more apparent when
8147 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8148 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8149 the correct format in RSA_verify so both forms transparently work.
8153 * Some servers which support TLS 1.0 can choke if we initially indicate
8154 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8155 encrypted premaster secret. As a workaround use the maximum permitted
8156 client version in client hello, this should keep such servers happy
8157 and still work with previous versions of OpenSSL.
8161 * Add support for TLS/DTLS heartbeats.
8163 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8165 * Add support for SCTP.
8167 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8169 * Improved PRNG seeding for VOS.
8171 *Paul Green <Paul.Green@stratus.com>*
8173 * Extensive assembler packs updates, most notably:
8175 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8176 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8177 - x86_64: bit-sliced AES implementation;
8178 - ARM: NEON support, contemporary platforms optimizations;
8179 - s390x: z196 support;
8180 - `*`: GHASH and GF(2^m) multiplication implementations;
8184 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8185 (removal of unnecessary code)
8187 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8189 * Add TLS key material exporter from RFC 5705.
8193 * Add DTLS-SRTP negotiation from RFC 5764.
8197 * Add Next Protocol Negotiation,
8198 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8199 disabled with a no-npn flag to config or Configure. Code donated
8202 *Adam Langley <agl@google.com> and Ben Laurie*
8204 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8205 NIST-P256, NIST-P521, with constant-time single point multiplication on
8206 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8207 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8208 Code made available under Apache License version 2.0.
8210 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8211 line to include this in your build of OpenSSL, and run "make depend" (or
8212 "make update"). This enables the following EC_METHODs:
8214 EC_GFp_nistp224_method()
8215 EC_GFp_nistp256_method()
8216 EC_GFp_nistp521_method()
8218 EC_GROUP_new_by_curve_name() will automatically use these (while
8219 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8222 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8224 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8225 all platforms. Move ssize_t definition from e_os.h to the public
8226 header file e_os2.h as it now appears in public header file cms.h
8230 * New -sigopt option to the ca, req and x509 utilities. Additional
8231 signature parameters can be passed using this option and in
8236 * Add RSA PSS signing function. This will generate and set the
8237 appropriate AlgorithmIdentifiers for PSS based on those in the
8238 corresponding EVP_MD_CTX structure. No application support yet.
8242 * Support for companion algorithm specific ASN1 signing routines.
8243 New function ASN1_item_sign_ctx() signs a pre-initialised
8244 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8245 the appropriate parameters.
8249 * Add new algorithm specific ASN1 verification initialisation function
8250 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8251 handling will be the same no matter what EVP_PKEY_METHOD is used.
8252 Add a PSS handler to support verification of PSS signatures: checked
8253 against a number of sample certificates.
8257 * Add signature printing for PSS. Add PSS OIDs.
8259 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8261 * Add algorithm specific signature printing. An individual ASN1 method
8262 can now print out signatures instead of the standard hex dump.
8264 More complex signatures (e.g. PSS) can print out more meaningful
8265 information. Include DSA version that prints out the signature
8270 * Password based recipient info support for CMS library: implementing
8275 * Split password based encryption into PBES2 and PBKDF2 functions. This
8276 neatly separates the code into cipher and PBE sections and is required
8277 for some algorithms that split PBES2 into separate pieces (such as
8278 password based CMS).
8282 * Session-handling fixes:
8283 - Fix handling of connections that are resuming with a session ID,
8284 but also support Session Tickets.
8285 - Fix a bug that suppressed issuing of a new ticket if the client
8286 presented a ticket with an expired session.
8287 - Try to set the ticket lifetime hint to something reasonable.
8288 - Make tickets shorter by excluding irrelevant information.
8289 - On the client side, don't ignore renewed tickets.
8291 *Adam Langley, Bodo Moeller (Google)*
8293 * Fix PSK session representation.
8297 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8299 This work was sponsored by Intel.
8303 * Add GCM support to TLS library. Some custom code is needed to split
8304 the IV between the fixed (from PRF) and explicit (from TLS record)
8305 portions. This adds all GCM ciphersuites supported by RFC5288 and
8306 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8307 add a special AESGCM string for GCM only.
8311 * Expand range of ctrls for AES GCM. Permit setting invocation
8312 field on decrypt and retrieval of invocation field only on encrypt.
8316 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8317 As required by RFC5289 these ciphersuites cannot be used if for
8318 versions of TLS earlier than 1.2.
8322 * For FIPS capable OpenSSL interpret a NULL default public key method
8323 as unset and return the appropriate default but do *not* set the default.
8324 This means we can return the appropriate method in applications that
8325 switch between FIPS and non-FIPS modes.
8329 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8330 ENGINE is used then we cannot handle that in the FIPS module so we
8331 keep original code iff non-FIPS operations are allowed.
8335 * Add -attime option to openssl utilities.
8337 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8339 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8343 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8344 FIPS EC methods unconditionally for now.
8348 * New build option no-ec2m to disable characteristic 2 code.
8352 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8353 all cases can be covered as some introduce binary incompatibilities.
8357 * Redirect RSA operations to FIPS module including keygen,
8358 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8362 * Add similar low-level API blocking to ciphers.
8366 * low-level digest APIs are not approved in FIPS mode: any attempt
8367 to use these will cause a fatal error. Applications that *really* want
8368 to use them can use the `private_*` version instead.
8372 * Redirect cipher operations to FIPS module for FIPS builds.
8376 * Redirect digest operations to FIPS module for FIPS builds.
8380 * Update build system to add "fips" flag which will link in fipscanister.o
8381 for static and shared library builds embedding a signature if needed.
8385 * Output TLS supported curves in preference order instead of numerical
8386 order. This is currently hardcoded for the highest order curves first.
8387 This should be configurable so applications can judge speed vs strength.
8391 * Add TLS v1.2 server support for client authentication.
8395 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8400 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8401 FIPS modules versions.
8405 * Add TLS v1.2 client side support for client authentication. Keep cache
8406 of handshake records longer as we don't know the hash algorithm to use
8407 until after the certificate request message is received.
8411 * Initial TLS v1.2 client support. Add a default signature algorithms
8412 extension including all the algorithms we support. Parse new signature
8413 format in client key exchange. Relax some ECC signing restrictions for
8414 TLS v1.2 as indicated in RFC5246.
8418 * Add server support for TLS v1.2 signature algorithms extension. Switch
8419 to new signature format when needed using client digest preference.
8420 All server ciphersuites should now work correctly in TLS v1.2. No client
8421 support yet and no support for client certificates.
8425 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8426 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8427 ciphersuites. At present only RSA key exchange ciphersuites work with
8428 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8429 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8430 and version checking.
8434 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8435 with this defined it will not be affected by any changes to ssl internal
8436 structures. Add several utility functions to allow openssl application
8437 to work with OPENSSL_NO_SSL_INTERN defined.
8441 * A long standing patch to add support for SRP from EdelWeb (Peter
8442 Sylvester and Christophe Renou) was integrated.
8443 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8444 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8447 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8451 * Permit abbreviated handshakes when renegotiating using the function
8452 SSL_renegotiate_abbreviated().
8454 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8456 * Add call to ENGINE_register_all_complete() to
8457 ENGINE_load_builtin_engines(), so some implementations get used
8458 automatically instead of needing explicit application support.
8462 * Add support for TLS key exporter as described in RFC5705.
8464 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8466 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8467 a few changes are required:
8469 Add SSL_OP_NO_TLSv1_1 flag.
8470 Add TLSv1_1 methods.
8471 Update version checking logic to handle version 1.1.
8472 Add explicit IV handling (ported from DTLS code).
8473 Add command line options to s_client/s_server.
8480 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8482 * X509_ATTRIBUTE memory leak
8484 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8485 memory. This structure is used by the PKCS#7 and CMS routines so any
8486 application which reads PKCS#7 or CMS data from untrusted sources is
8487 affected. SSL/TLS is not affected.
8489 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8495 * Race condition handling PSK identify hint
8497 If PSK identity hints are received by a multi-threaded client then
8498 the values are wrongly updated in the parent SSL_CTX structure. This can
8499 result in a race condition potentially leading to a double free of the
8505 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8507 * Malformed ECParameters causes infinite loop
8509 When processing an ECParameters structure OpenSSL enters an infinite loop
8510 if the curve specified is over a specially malformed binary polynomial
8513 This can be used to perform denial of service against any
8514 system which processes public keys, certificate requests or
8515 certificates. This includes TLS clients and TLS servers with
8516 client authentication enabled.
8518 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8523 * Exploitable out-of-bounds read in X509_cmp_time
8525 X509_cmp_time does not properly check the length of the ASN1_TIME
8526 string and can read a few bytes out of bounds. In addition,
8527 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8530 An attacker can use this to craft malformed certificates and CRLs of
8531 various sizes and potentially cause a segmentation fault, resulting in
8532 a DoS on applications that verify certificates or CRLs. TLS clients
8533 that verify CRLs are affected. TLS clients and servers with client
8534 authentication enabled may be affected if they use custom verification
8537 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8538 independently by Hanno Böck.
8543 * PKCS7 crash with missing EnvelopedContent
8545 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8546 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8547 with missing content and trigger a NULL pointer dereference on parsing.
8549 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8550 structures from untrusted sources are affected. OpenSSL clients and
8551 servers are not affected.
8553 This issue was reported to OpenSSL by Michal Zalewski (Google).
8558 * CMS verify infinite loop with unknown hash function
8560 When verifying a signedData message the CMS code can enter an infinite loop
8561 if presented with an unknown hash function OID. This can be used to perform
8562 denial of service against any system which verifies signedData messages using
8564 This issue was reported to OpenSSL by Johannes Bauer.
8569 * Race condition handling NewSessionTicket
8571 If a NewSessionTicket is received by a multi-threaded client when attempting to
8572 reuse a previous ticket then a race condition can occur potentially leading to
8573 a double free of the ticket data.
8578 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8580 * Segmentation fault in ASN1_TYPE_cmp fix
8582 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8583 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8584 certificate signature algorithm consistency this can be used to crash any
8585 certificate verification operation and exploited in a DoS attack. Any
8586 application which performs certificate verification is vulnerable including
8587 OpenSSL clients and servers which enable client authentication.
8592 * ASN.1 structure reuse memory corruption fix
8594 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8595 memory corruption via an invalid write. Such reuse is and has been
8596 strongly discouraged and is believed to be rare.
8598 Applications that parse structures containing CHOICE or ANY DEFINED BY
8599 components may be affected. Certificate parsing (d2i_X509 and related
8600 functions) are however not affected. OpenSSL clients and servers are
8606 * PKCS7 NULL pointer dereferences fix
8608 The PKCS#7 parsing code does not handle missing outer ContentInfo
8609 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8610 missing content and trigger a NULL pointer dereference on parsing.
8612 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8613 otherwise parse PKCS#7 structures from untrusted sources are
8614 affected. OpenSSL clients and servers are not affected.
8616 This issue was reported to OpenSSL by Michal Zalewski (Google).
8621 * DoS via reachable assert in SSLv2 servers fix
8623 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8624 servers that both support SSLv2 and enable export cipher suites by sending
8625 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8627 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8628 (OpenSSL development team).
8633 * Use After Free following d2i_ECPrivatekey error fix
8635 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8636 could cause a use after free condition. This, in turn, could cause a double
8637 free in several private key parsing functions (such as d2i_PrivateKey
8638 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8639 for applications that receive EC private keys from untrusted
8640 sources. This scenario is considered rare.
8642 This issue was discovered by the BoringSSL project and fixed in their
8648 * X509_to_X509_REQ NULL pointer deref fix
8650 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8651 the certificate key is invalid. This function is rarely used in practice.
8653 This issue was discovered by Brian Carpenter.
8658 * Removed the export ciphers from the DEFAULT ciphers
8662 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8664 * Build fixes for the Windows and OpenVMS platforms
8666 *Matt Caswell and Richard Levitte*
8668 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8670 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8671 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8672 dereference. This could lead to a Denial Of Service attack. Thanks to
8673 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8678 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8679 dtls1_buffer_record function under certain conditions. In particular this
8680 could occur if an attacker sent repeated DTLS records with the same
8681 sequence number but for the next epoch. The memory leak could be exploited
8682 by an attacker in a Denial of Service attack through memory exhaustion.
8683 Thanks to Chris Mueller for reporting this issue.
8688 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8689 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8690 method would be set to NULL which could later result in a NULL pointer
8691 dereference. Thanks to Frank Schmirler for reporting this issue.
8696 * Abort handshake if server key exchange message is omitted for ephemeral
8699 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8700 reporting this issue.
8705 * Remove non-export ephemeral RSA code on client and server. This code
8706 violated the TLS standard by allowing the use of temporary RSA keys in
8707 non-export ciphersuites and could be used by a server to effectively
8708 downgrade the RSA key length used to a value smaller than the server
8709 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8710 INRIA or reporting this issue.
8715 * Fixed issue where DH client certificates are accepted without verification.
8716 An OpenSSL server will accept a DH certificate for client authentication
8717 without the certificate verify message. This effectively allows a client to
8718 authenticate without the use of a private key. This only affects servers
8719 which trust a client certificate authority which issues certificates
8720 containing DH keys: these are extremely rare and hardly ever encountered.
8721 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8727 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8728 results on some platforms, including x86_64. This bug occurs at random
8729 with a very low probability, and is not known to be exploitable in any
8730 way, though its exact impact is difficult to determine. Thanks to Pieter
8731 Wuille (Blockstream) who reported this issue and also suggested an initial
8732 fix. Further analysis was conducted by the OpenSSL development team and
8733 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8734 the OpenSSL core team.
8739 * Fix various certificate fingerprint issues.
8741 By using non-DER or invalid encodings outside the signed portion of a
8742 certificate the fingerprint can be changed without breaking the signature.
8743 Although no details of the signed portion of the certificate can be changed
8744 this can cause problems with some applications: e.g. those using the
8745 certificate fingerprint for blacklists.
8747 1. Reject signatures with non zero unused bits.
8749 If the BIT STRING containing the signature has non zero unused bits reject
8750 the signature. All current signature algorithms require zero unused bits.
8752 2. Check certificate algorithm consistency.
8754 Check the AlgorithmIdentifier inside TBS matches the one in the
8755 certificate signature. NB: this will result in signature failure
8756 errors for some broken certificates.
8758 Thanks to Konrad Kraszewski from Google for reporting this issue.
8760 3. Check DSA/ECDSA signatures use DER.
8762 Reencode DSA/ECDSA signatures and compare with the original received
8763 signature. Return an error if there is a mismatch.
8765 This will reject various cases including garbage after signature
8766 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8767 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8768 (negative or with leading zeroes).
8770 Further analysis was conducted and fixes were developed by Stephen Henson
8771 of the OpenSSL core team.
8777 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8779 * Session Ticket Memory Leak.
8781 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8782 integrity of that ticket is first verified. In the event of a session
8783 ticket integrity check failing, OpenSSL will fail to free memory
8784 causing a memory leak. By sending a large number of invalid session
8785 tickets an attacker could exploit this issue in a Denial Of Service
8791 * Build option no-ssl3 is incomplete.
8793 When OpenSSL is configured with "no-ssl3" as a build option, servers
8794 could accept and complete an SSL 3.0 handshake, and clients could be
8795 configured to send them.
8798 *Akamai and the OpenSSL team*
8800 * Add support for TLS_FALLBACK_SCSV.
8801 Client applications doing fallback retries should call
8802 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8805 *Adam Langley, Bodo Moeller*
8807 * Add additional DigestInfo checks.
8809 Reencode DigestInto in DER and check against the original when
8810 verifying RSA signature: this will reject any improperly encoded
8811 DigestInfo structures.
8813 Note: this is a precautionary measure and no attacks are currently known.
8817 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8819 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8820 to a denial of service attack. A malicious server can crash the client
8821 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8822 ciphersuite and sending carefully crafted handshake messages.
8824 Thanks to Felix Gröbert (Google) for discovering and researching this
8830 * By sending carefully crafted DTLS packets an attacker could cause openssl
8831 to leak memory. This can be exploited through a Denial of Service attack.
8832 Thanks to Adam Langley for discovering and researching this issue.
8837 * An attacker can force openssl to consume large amounts of memory whilst
8838 processing DTLS handshake messages. This can be exploited through a
8839 Denial of Service attack.
8840 Thanks to Adam Langley for discovering and researching this issue.
8845 * An attacker can force an error condition which causes openssl to crash
8846 whilst processing DTLS packets due to memory being freed twice. This
8847 can be exploited through a Denial of Service attack.
8848 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8854 * If a multithreaded client connects to a malicious server using a resumed
8855 session and the server sends an ec point format extension it could write
8856 up to 255 bytes to freed memory.
8858 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8864 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8865 X509_name_oneline, X509_name_print_ex et al. to leak some information
8866 from the stack. Applications may be affected if they echo pretty printing
8867 output to the attacker.
8869 Thanks to Ivan Fratric (Google) for discovering this issue.
8872 *Emilia Käsper, and Steve Henson*
8874 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8875 for corner cases. (Certain input points at infinity could lead to
8876 bogus results, with non-infinity inputs mapped to infinity too.)
8880 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8882 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8883 handshake can force the use of weak keying material in OpenSSL
8884 SSL/TLS clients and servers.
8886 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8887 researching this issue. ([CVE-2014-0224])
8889 *KIKUCHI Masashi, Steve Henson*
8891 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8892 OpenSSL DTLS client the code can be made to recurse eventually crashing
8895 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8898 *Imre Rad, Steve Henson*
8900 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8901 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8902 client or server. This is potentially exploitable to run arbitrary
8903 code on a vulnerable client or server.
8905 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8907 *Jüri Aedla, Steve Henson*
8909 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8910 are subject to a denial of service attack.
8912 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8913 this issue. ([CVE-2014-3470])
8915 *Felix Gröbert, Ivan Fratric, Steve Henson*
8917 * Harmonize version and its documentation. -f flag is used to display
8920 *mancha <mancha1@zoho.com>*
8922 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8923 in i2d_ECPrivateKey.
8925 *mancha <mancha1@zoho.com>*
8927 * Fix some double frees. These are not thought to be exploitable.
8929 *mancha <mancha1@zoho.com>*
8931 * Fix for the attack described in the paper "Recovering OpenSSL
8932 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8933 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8934 <http://eprint.iacr.org/2014/140>
8936 Thanks to Yuval Yarom and Naomi Benger for discovering this
8937 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8939 *Yuval Yarom and Naomi Benger*
8941 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8943 * Keep original DTLS digest and encryption contexts in retransmission
8944 structures so we can use the previous session parameters if they need
8945 to be resent. ([CVE-2013-6450])
8949 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8950 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8951 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8952 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8953 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8954 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8956 *Rob Stradling, Adam Langley*
8958 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8960 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8962 This addresses the flaw in CBC record processing discovered by
8963 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8964 at: <http://www.isg.rhul.ac.uk/tls/>
8966 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8967 Security Group at Royal Holloway, University of London
8968 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8969 Emilia Käsper for the initial patch.
8972 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8974 * Return an error when checking OCSP signatures when key is NULL.
8975 This fixes a DoS attack. ([CVE-2013-0166])
8979 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8980 the right response is stapled. Also change SSL_get_certificate()
8981 so it returns the certificate actually sent.
8982 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8983 (This is a backport)
8985 *Rob Stradling <rob.stradling@comodo.com>*
8987 * Fix possible deadlock when decoding public keys.
8991 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8993 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8996 * Sanity check record length before skipping explicit IV in DTLS
8999 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
9000 fuzzing as a service testing platform.
9005 * Initialise tkeylen properly when encrypting CMS messages.
9006 Thanks to Solar Designer of Openwall for reporting this issue.
9010 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
9012 * Check for potentially exploitable overflows in asn1_d2i_read_bio
9013 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
9014 in CRYPTO_realloc_clean.
9016 Thanks to Tavis Ormandy, Google Security Team, for discovering this
9017 issue and to Adam Langley <agl@chromium.org> for fixing it.
9020 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
9022 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
9024 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
9025 in CMS and PKCS7 code. When RSA decryption fails use a random key for
9026 content decryption and always return the same error. Note: this attack
9027 needs on average 2^20 messages so it only affects automated senders. The
9028 old behaviour can be re-enabled in the CMS code by setting the
9029 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
9030 an MMA defence is not necessary.
9031 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
9032 this issue. ([CVE-2012-0884])
9036 * Fix CVE-2011-4619: make sure we really are receiving a
9037 client hello before rejecting multiple SGC restarts. Thanks to
9038 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
9042 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
9044 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
9045 Thanks to Antonio Martin, Enterprise Secure Access Research and
9046 Development, Cisco Systems, Inc. for discovering this bug and
9047 preparing a fix. ([CVE-2012-0050])
9051 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
9053 * Nadhem Alfardan and Kenny Paterson have discovered an extension
9054 of the Vaudenay padding oracle attack on CBC mode encryption
9055 which enables an efficient plaintext recovery attack against
9056 the OpenSSL implementation of DTLS. Their attack exploits timing
9057 differences arising during decryption processing. A research
9058 paper describing this attack can be found at:
9059 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
9060 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9061 Security Group at Royal Holloway, University of London
9062 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9063 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9064 for preparing the fix. ([CVE-2011-4108])
9066 *Robin Seggelmann, Michael Tuexen*
9068 * Clear bytes used for block padding of SSL 3.0 records.
9071 *Adam Langley (Google)*
9073 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9074 Kadianakis <desnacked@gmail.com> for discovering this issue and
9075 Adam Langley for preparing the fix. ([CVE-2011-4619])
9077 *Adam Langley (Google)*
9079 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9081 *Andrey Kulikov <amdeich@gmail.com>*
9083 * Prevent malformed RFC3779 data triggering an assertion failure.
9084 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9085 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9087 *Rob Austein <sra@hactrn.net>*
9089 * Improved PRNG seeding for VOS.
9091 *Paul Green <Paul.Green@stratus.com>*
9093 * Fix ssl_ciph.c set-up race.
9095 *Adam Langley (Google)*
9097 * Fix spurious failures in ecdsatest.c.
9099 *Emilia Käsper (Google)*
9101 * Fix the BIO_f_buffer() implementation (which was mixing different
9102 interpretations of the `..._len` fields).
9104 *Adam Langley (Google)*
9106 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9107 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9108 threads won't reuse the same blinding coefficients.
9110 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9111 lock to call BN_BLINDING_invert_ex, and avoids one use of
9112 BN_BLINDING_update for each BN_BLINDING structure (previously,
9113 the last update always remained unused).
9115 *Emilia Käsper (Google)*
9117 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9119 *Bob Buckholz (Google)*
9121 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9123 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9124 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9126 *Kaspar Brand <ossl@velox.ch>*
9128 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9129 for multi-threaded use of ECDH. ([CVE-2011-3210])
9131 *Adam Langley (Google)*
9133 * Fix x509_name_ex_d2i memory leak on bad inputs.
9137 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9138 signature public key algorithm by using OID xref utilities instead.
9139 Before this you could only use some ECC ciphersuites with SHA1 only.
9143 * Add protection against ECDSA timing attacks as mentioned in the paper
9144 by Billy Bob Brumley and Nicola Tuveri, see:
9145 <http://eprint.iacr.org/2011/232.pdf>
9147 *Billy Bob Brumley and Nicola Tuveri*
9149 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9151 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9153 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9155 * Fix bug in string printing code: if *any* escaping is enabled we must
9156 escape the escape character (backslash) or the resulting string is
9161 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9163 * Disable code workaround for ancient and obsolete Netscape browsers
9164 and servers: an attacker can use it in a ciphersuite downgrade attack.
9165 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9169 * Fixed J-PAKE implementation error, originally discovered by
9170 Sebastien Martini, further info and confirmation from Stefan
9171 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9175 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9177 * Fix extension code to avoid race conditions which can result in a buffer
9178 overrun vulnerability: resumed sessions must not be modified as they can
9179 be shared by multiple threads. CVE-2010-3864
9183 * Fix WIN32 build system to correctly link an ENGINE directory into
9188 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9190 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9193 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9195 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9197 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9198 context. The operation can be customised via the ctrl mechanism in
9199 case ENGINEs want to include additional functionality.
9203 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9207 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9208 output hashes compatible with older versions of OpenSSL.
9210 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9212 * Fix compression algorithm handling: if resuming a session use the
9213 compression algorithm of the resumed session instead of determining
9214 it from client hello again. Don't allow server to change algorithm.
9218 * Add load_crls() function to commands tidying load_certs() too. Add option
9219 to verify utility to allow additional CRLs to be included.
9223 * Update OCSP request code to permit adding custom headers to the request:
9224 some responders need this.
9228 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9231 *Julia Lawall <julia@diku.dk>*
9233 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9234 needlessly dereferenced structures, used obsolete functions and
9235 didn't handle all updated verify codes correctly.
9239 * Disable MD2 in the default configuration.
9243 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9244 indicate the initial BIO being pushed or popped. This makes it possible
9245 to determine whether the BIO is the one explicitly called or as a result
9246 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9247 it handles reference counts correctly and doesn't zero out the I/O bio
9248 when it is not being explicitly popped. WARNING: applications which
9249 included workarounds for the old buggy behaviour will need to be modified
9250 or they could free up already freed BIOs.
9254 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9255 renaming to all platforms (within the 0.9.8 branch, this was
9256 done conditionally on Netware platforms to avoid a name clash).
9258 *Guenter <lists@gknw.net>*
9260 * Add ECDHE and PSK support to DTLS.
9262 *Michael Tuexen <tuexen@fh-muenster.de>*
9264 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9269 * Add "missing" function EVP_MD_flags() (without this the only way to
9270 retrieve a digest flags is by accessing the structure directly. Update
9271 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9272 or cipher is registered as in the "from" argument. Print out all
9273 registered digests in the dgst usage message instead of manually
9274 attempting to work them out.
9278 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9279 this allows the use of compression and extensions. Change default cipher
9280 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9281 by default unless an application cipher string requests it.
9285 * Alter match criteria in PKCS12_parse(). It used to try to use local
9286 key ids to find matching certificates and keys but some PKCS#12 files
9287 don't follow the (somewhat unwritten) rules and this strategy fails.
9288 Now just gather all certificates together and the first private key
9289 then look for the first certificate that matches the key.
9293 * Support use of registered digest and cipher names for dgst and cipher
9294 commands instead of having to add each one as a special case. So now
9301 openssl dgst -sha256 foo
9303 and this works for ENGINE based algorithms too.
9307 * Update Gost ENGINE to support parameter files.
9309 *Victor B. Wagner <vitus@cryptocom.ru>*
9311 * Support GeneralizedTime in ca utility.
9313 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9315 * Enhance the hash format used for certificate directory links. The new
9316 form uses the canonical encoding (meaning equivalent names will work
9317 even if they aren't identical) and uses SHA1 instead of MD5. This form
9318 is incompatible with the older format and as a result c_rehash should
9319 be used to rebuild symbolic links.
9323 * Make PKCS#8 the default write format for private keys, replacing the
9324 traditional format. This form is standardised, more secure and doesn't
9325 include an implicit MD5 dependency.
9329 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9330 committed to OpenSSL should pass this lot as a minimum.
9334 * Add session ticket override functionality for use by EAP-FAST.
9336 *Jouni Malinen <j@w1.fi>*
9338 * Modify HMAC functions to return a value. Since these can be implemented
9339 in an ENGINE errors can occur.
9343 * Type-checked OBJ_bsearch_ex.
9347 * Type-checked OBJ_bsearch. Also some constification necessitated
9348 by type-checking. Still to come: TXT_DB, bsearch(?),
9349 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9354 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9355 seconds to a tm structure directly, instead of going through OS
9356 specific date routines. This avoids any issues with OS routines such
9357 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9358 and X509_time_adj_ex() to cover the extended range. The existing
9359 X509_time_adj() is still usable and will no longer have any date issues.
9363 * Delta CRL support. New use deltas option which will attempt to locate
9364 and search any appropriate delta CRLs available.
9366 This work was sponsored by Google.
9370 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9371 code and add additional score elements. Validate alternate CRL paths
9372 as part of the CRL checking and indicate a new error "CRL path validation
9373 error" in this case. Applications wanting additional details can use
9374 the verify callback and check the new "parent" field. If this is not
9375 NULL CRL path validation is taking place. Existing applications won't
9376 see this because it requires extended CRL support which is off by
9379 This work was sponsored by Google.
9383 * Support for freshest CRL extension.
9385 This work was sponsored by Google.
9389 * Initial indirect CRL support. Currently only supported in the CRLs
9390 passed directly and not via lookup. Process certificate issuer
9391 CRL entry extension and lookup CRL entries by bother issuer name
9392 and serial number. Check and process CRL issuer entry in IDP extension.
9394 This work was sponsored by Google.
9398 * Add support for distinct certificate and CRL paths. The CRL issuer
9399 certificate is validated separately in this case. Only enabled if
9400 an extended CRL support flag is set: this flag will enable additional
9401 CRL functionality in future.
9403 This work was sponsored by Google.
9407 * Add support for policy mappings extension.
9409 This work was sponsored by Google.
9413 * Fixes to pathlength constraint, self issued certificate handling,
9414 policy processing to align with RFC3280 and PKITS tests.
9416 This work was sponsored by Google.
9420 * Support for name constraints certificate extension. DN, email, DNS
9421 and URI types are currently supported.
9423 This work was sponsored by Google.
9427 * To cater for systems that provide a pointer-based thread ID rather
9428 than numeric, deprecate the current numeric thread ID mechanism and
9429 replace it with a structure and associated callback type. This
9430 mechanism allows a numeric "hash" to be extracted from a thread ID in
9431 either case, and on platforms where pointers are larger than 'long',
9432 mixing is done to help ensure the numeric 'hash' is usable even if it
9433 can't be guaranteed unique. The default mechanism is to use "&errno"
9434 as a pointer-based thread ID to distinguish between threads.
9436 Applications that want to provide their own thread IDs should now use
9437 CRYPTO_THREADID_set_callback() to register a callback that will call
9438 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9440 Note that ERR_remove_state() is now deprecated, because it is tied
9441 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9442 to free the current thread's error state should be replaced by
9443 ERR_remove_thread_state(NULL).
9445 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9446 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9447 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9448 application was previously providing a numeric thread callback that
9449 was inappropriate for distinguishing threads, then uniqueness might
9450 have been obtained with &errno that happened immediately in the
9451 intermediate development versions of OpenSSL; this is no longer the
9452 case, the numeric thread callback will now override the automatic use
9455 *Geoff Thorpe, with help from Bodo Moeller*
9457 * Initial support for different CRL issuing certificates. This covers a
9458 simple case where the self issued certificates in the chain exist and
9459 the real CRL issuer is higher in the existing chain.
9461 This work was sponsored by Google.
9465 * Removed effectively defunct crypto/store from the build.
9469 * Revamp of STACK to provide stronger type-checking. Still to come:
9470 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9471 ASN1_STRING, CONF_VALUE.
9475 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9476 RAM on SSL connections. This option can save about 34k per idle SSL.
9480 * Revamp of LHASH to provide stronger type-checking. Still to come:
9481 STACK, TXT_DB, bsearch, qsort.
9485 * Initial support for Cryptographic Message Syntax (aka CMS) based
9486 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9487 support for data, signedData, compressedData, digestedData and
9488 encryptedData, envelopedData types included. Scripts to check against
9489 RFC4134 examples draft and interop and consistency checks of many
9490 content types and variants.
9494 * Add options to enc utility to support use of zlib compression BIO.
9498 * Extend mk1mf to support importing of options and assembly language
9499 files from Configure script, currently only included in VC-WIN32.
9500 The assembly language rules can now optionally generate the source
9501 files from the associated perl scripts.
9505 * Implement remaining functionality needed to support GOST ciphersuites.
9506 Interop testing has been performed using CryptoPro implementations.
9508 *Victor B. Wagner <vitus@cryptocom.ru>*
9510 * s390x assembler pack.
9514 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9519 * Implement Opaque PRF Input TLS extension as specified in
9520 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9521 official specification yet and no extension type assignment by
9522 IANA exists, this extension (for now) will have to be explicitly
9523 enabled when building OpenSSL by providing the extension number
9524 to use. For example, specify an option
9526 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9528 to the "config" or "Configure" script to enable the extension,
9529 assuming extension number 0x9527 (which is a completely arbitrary
9530 and unofficial assignment based on the MD5 hash of the Internet
9531 Draft). Note that by doing so, you potentially lose
9532 interoperability with other TLS implementations since these might
9533 be using the same extension number for other purposes.
9535 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9536 opaque PRF input value to use in the handshake. This will create
9537 an internal copy of the length-'len' string at 'src', and will
9538 return non-zero for success.
9540 To get more control and flexibility, provide a callback function
9543 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9544 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9548 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9551 Callback function 'cb' will be called in handshakes, and is
9552 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9553 Argument 'arg' is for application purposes (the value as given to
9554 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9555 be provided to the callback function). The callback function
9556 has to return non-zero to report success: usually 1 to use opaque
9557 PRF input just if possible, or 2 to enforce use of the opaque PRF
9558 input. In the latter case, the library will abort the handshake
9559 if opaque PRF input is not successfully negotiated.
9561 Arguments 'peerinput' and 'len' given to the callback function
9562 will always be NULL and 0 in the case of a client. A server will
9563 see the client's opaque PRF input through these variables if
9564 available (NULL and 0 otherwise). Note that if the server
9565 provides an opaque PRF input, the length must be the same as the
9566 length of the client's opaque PRF input.
9568 Note that the callback function will only be called when creating
9569 a new session (session resumption can resume whatever was
9570 previously negotiated), and will not be called in SSL 2.0
9571 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9572 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9573 for applications that need to enforce opaque PRF input.
9577 * Update ssl code to support digests other than SHA1+MD5 for handshake
9580 *Victor B. Wagner <vitus@cryptocom.ru>*
9582 * Add RFC4507 support to OpenSSL. This includes the corrections in
9583 RFC4507bis. The encrypted ticket format is an encrypted encoded
9584 SSL_SESSION structure, that way new session features are automatically
9587 If a client application caches session in an SSL_SESSION structure
9588 support is transparent because tickets are now stored in the encoded
9591 The SSL_CTX structure automatically generates keys for ticket
9592 protection in servers so again support should be possible
9593 with no application modification.
9595 If a client or server wishes to disable RFC4507 support then the option
9596 SSL_OP_NO_TICKET can be set.
9598 Add a TLS extension debugging callback to allow the contents of any client
9599 or server extensions to be examined.
9601 This work was sponsored by Google.
9605 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9606 OpenSSL should now compile cleanly on gcc 4.2
9608 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9610 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9611 support including streaming MAC support: this is required for GOST
9612 ciphersuite support.
9614 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9616 * Add option -stream to use PKCS#7 streaming in smime utility. New
9617 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9618 to output in BER and PEM format.
9622 * Experimental support for use of HMAC via EVP_PKEY interface. This
9623 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9624 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9625 ENGINE support for HMAC keys which are unextractable. New -mac and
9626 -macopt options to dgst utility.
9630 * New option -sigopt to dgst utility. Update dgst to use
9631 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9632 alternative signing parameters such as X9.31 or PSS in the dgst
9637 * Change ssl_cipher_apply_rule(), the internal function that does
9638 the work each time a ciphersuite string requests enabling
9639 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9640 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9641 the order of disabled ciphersuites such that those ciphersuites
9642 that most recently went from enabled to disabled not only stay
9643 in order with respect to each other, but also have higher priority
9644 than other disabled ciphersuites the next time ciphersuites are
9647 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9648 the same ciphersuites as with "HIGH" alone, but in a specific
9649 order where the PSK ciphersuites come first (since they are the
9650 most recently disabled ciphersuites when "HIGH" is parsed).
9652 Also, change ssl_create_cipher_list() (using this new
9653 functionality) such that between otherwise identical
9654 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9659 * Change ssl_create_cipher_list() so that it automatically
9660 arranges the ciphersuites in reasonable order before starting
9661 to process the rule string. Thus, the definition for "DEFAULT"
9662 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9663 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9664 This makes it much easier to arrive at a reasonable default order
9665 in applications for which anonymous ciphers are OK (meaning
9666 that you can't actually use DEFAULT).
9668 *Bodo Moeller; suggested by Victor Duchovni*
9670 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9671 processing) into multiple integers instead of setting
9672 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9673 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9674 (These masks as well as the individual bit definitions are hidden
9675 away into the non-exported interface ssl/ssl_locl.h, so this
9676 change to the definition of the SSL_CIPHER structure shouldn't
9677 affect applications.) This give us more bits for each of these
9678 categories, so there is no longer a need to coagulate AES128 and
9679 AES256 into a single algorithm bit, and to coagulate Camellia128
9680 and Camellia256 into a single algorithm bit, which has led to all
9683 Thus, among other things, the kludge introduced in 0.9.7m and
9684 0.9.8e for masking out AES256 independently of AES128 or masking
9685 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9687 With the change, we also introduce new ciphersuite aliases that
9688 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9693 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9694 Use the leftmost N bytes of the signature input if the input is
9695 larger than the prime q (with N being the size in bytes of q).
9699 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9700 it yet and it is largely untested.
9704 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9708 * Initial incomplete changes to avoid need for function casts in OpenSSL
9709 some compilers (gcc 4.2 and later) reject their use. Safestack is
9710 reimplemented. Update ASN1 to avoid use of legacy functions.
9714 * Win32/64 targets are linked with Winsock2.
9718 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9719 to external functions. This can be used to increase CRL handling
9720 efficiency especially when CRLs are very large by (for example) storing
9721 the CRL revoked certificates in a database.
9725 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9726 new CRLs added to a directory can be used. New command line option
9727 -verify_return_error to s_client and s_server. This causes real errors
9728 to be returned by the verify callback instead of carrying on no matter
9729 what. This reflects the way a "real world" verify callback would behave.
9733 * GOST engine, supporting several GOST algorithms and public key formats.
9734 Kindly donated by Cryptocom.
9738 * Partial support for Issuing Distribution Point CRL extension. CRLs
9739 partitioned by DP are handled but no indirect CRL or reason partitioning
9740 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9741 selected via a scoring technique which handles IDP and AKID in CRLs.
9745 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9746 will ultimately be used for all verify operations: this will remove the
9747 X509_STORE dependency on certificate verification and allow alternative
9748 lookup methods. X509_STORE based implementations of these two callbacks.
9752 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9753 Modify get_crl() to find a valid (unexpired) CRL if possible.
9757 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9758 this would be called X509_CRL_cmp() but that name is already used by
9759 a function that just compares CRL issuer names. Cache several CRL
9760 extensions in X509_CRL structure and cache CRLDP in X509.
9764 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9765 this maps equivalent X509_NAME structures into a consistent structure.
9766 Name comparison can then be performed rapidly using memcmp().
9770 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9775 * Allow digests to supply their own micalg string for S/MIME type using
9776 the ctrl EVP_MD_CTRL_MICALG.
9780 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9781 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9782 ctrl. It can then customise the structure before and/or after signing
9787 * New function OBJ_add_sigid() to allow application defined signature OIDs
9788 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9789 to free up any added signature OIDs.
9793 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9794 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9795 digest and cipher tables. New options added to openssl utility:
9796 list-message-digest-algorithms and list-cipher-algorithms.
9800 * Change the array representation of binary polynomials: the list
9801 of degrees of non-zero coefficients is now terminated with -1.
9802 Previously it was terminated with 0, which was also part of the
9803 value; thus, the array representation was not applicable to
9804 polynomials where t^0 has coefficient zero. This change makes
9805 the array representation useful in a more general context.
9809 * Various modifications and fixes to SSL/TLS cipher string
9810 handling. For ECC, the code now distinguishes between fixed ECDH
9811 with RSA certificates on the one hand and with ECDSA certificates
9812 on the other hand, since these are separate ciphersuites. The
9813 unused code for Fortezza ciphersuites has been removed.
9815 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9816 (not "ECDHE"). For consistency with the code for DH
9817 certificates, use of ECDH certificates is now considered ECDH
9818 authentication, not RSA or ECDSA authentication (the latter is
9819 merely the CA's signing algorithm and not actively used in the
9822 The temporary ciphersuite alias "ECCdraft" is no longer
9823 available, and ECC ciphersuites are no longer excluded from "ALL"
9824 and "DEFAULT". The following aliases now exist for RFC 4492
9825 ciphersuites, most of these by analogy with the DH case:
9827 kECDHr - ECDH cert, signed with RSA
9828 kECDHe - ECDH cert, signed with ECDSA
9829 kECDH - ECDH cert (signed with either RSA or ECDSA)
9830 kEECDH - ephemeral ECDH
9831 ECDH - ECDH cert or ephemeral ECDH
9837 AECDH - anonymous ECDH
9838 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9842 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9843 Use correct micalg parameters depending on digest(s) in signed message.
9847 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9848 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9852 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9853 an engine to register a method. Add ENGINE lookups for methods and
9854 functional reference processing.
9858 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9859 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9864 * New -resign option to smime utility. This adds one or more signers
9865 to an existing PKCS#7 signedData structure. Also -md option to use an
9866 alternative message digest algorithm for signing.
9870 * Tidy up PKCS#7 routines and add new functions to make it easier to
9871 create PKCS7 structures containing multiple signers. Update smime
9872 application to support multiple signers.
9876 * New -macalg option to pkcs12 utility to allow setting of an alternative
9881 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9882 Reorganize PBE internals to lookup from a static table using NIDs,
9883 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9884 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9885 PRF which will be automatically used with PBES2.
9889 * Replace the algorithm specific calls to generate keys in "req" with the
9894 * Update PKCS#7 enveloped data routines to use new API. This is now
9895 supported by any public key method supporting the encrypt operation. A
9896 ctrl is added to allow the public key algorithm to examine or modify
9897 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9902 * Add a ctrl to asn1 method to allow a public key algorithm to express
9903 a default digest type to use. In most cases this will be SHA1 but some
9904 algorithms (such as GOST) need to specify an alternative digest. The
9905 return value indicates how strong the preference is 1 means optional and
9906 2 is mandatory (that is it is the only supported type). Modify
9907 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9908 use the default md. Update openssl utilities to use the default digest
9909 type for signing if it is not explicitly indicated.
9913 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9914 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9915 signing method from the key type. This effectively removes the link
9916 between digests and public key types.
9920 * Add an OID cross reference table and utility functions. Its purpose is to
9921 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9922 rsaEncryption. This will allow some of the algorithm specific hackery
9923 needed to use the correct OID to be removed.
9927 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9928 structures for PKCS7_sign(). They are now set up by the relevant public
9933 * Add provisional EC pkey method with support for ECDSA and ECDH.
9937 * Add support for key derivation (agreement) in the API, DH method and
9942 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9943 public and private key formats. As a side effect these add additional
9944 command line functionality not previously available: DSA signatures can be
9945 generated and verified using pkeyutl and DH key support and generation in
9952 *Oliver Tappe <zooey@hirschkaefer.de>*
9954 * New make target "install_html_docs" installs HTML renditions of the
9957 *Oliver Tappe <zooey@hirschkaefer.de>*
9959 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9960 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9961 support key and parameter generation and add initial key generation
9962 functionality for RSA.
9966 * Add functions for main EVP_PKEY_method operations. The undocumented
9967 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9968 `EVP_PKEY_{encrypt,decrypt}_old`.
9972 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9973 key API, doesn't do much yet.
9977 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9978 public key algorithms. New option to openssl utility:
9979 "list-public-key-algorithms" to print out info.
9983 * Implement the Supported Elliptic Curves Extension for
9984 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9988 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9989 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9993 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9994 utilities such as rsa, dsa, dsaparam etc except they process any key
9999 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
10000 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
10001 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
10006 * Initial support for pluggable public key ASN1.
10007 De-spaghettify the public key ASN1 handling. Move public and private
10008 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
10009 algorithm specific handling to a single module within the relevant
10010 algorithm directory. Add functions to allow (near) opaque processing
10011 of public and private key structures.
10015 * Implement the Supported Point Formats Extension for
10016 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10020 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
10021 for the psk identity [hint] and the psk callback functions to the
10022 SSL_SESSION, SSL and SSL_CTX structure.
10025 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
10029 SSL_CTX_use_psk_identity_hint
10030 SSL_get_psk_identity_hint
10031 SSL_get_psk_identity
10032 SSL_use_psk_identity_hint
10034 *Mika Kousa and Pasi Eronen of Nokia Corporation*
10036 * Add RFC 3161 compliant time stamp request creation, response generation
10037 and response verification functionality.
10039 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
10041 * Add initial support for TLS extensions, specifically for the server_name
10042 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10043 have new members for a hostname. The SSL data structure has an
10044 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10045 stored in that context to allow for session resumption, even after the
10046 SSL has been switched to a new SSL_CTX in reaction to a client's
10047 server_name extension.
10049 New functions (subject to change):
10051 SSL_get_servername()
10052 SSL_get_servername_type()
10055 New CTRL codes and macros (subject to change):
10057 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10058 - SSL_CTX_set_tlsext_servername_callback()
10059 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10060 - SSL_CTX_set_tlsext_servername_arg()
10061 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10063 openssl s_client has a new '-servername ...' option.
10065 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10066 '-key2 ...', '-servername_fatal' (subject to change). This allows
10067 testing the HostName extension for a specific single hostname ('-cert'
10068 and '-key' remain fallbacks for handshakes without HostName
10069 negotiation). If the unrecognized_name alert has to be sent, this by
10070 default is a warning; it becomes fatal with the '-servername_fatal'
10073 *Peter Sylvester, Remy Allais, Christophe Renou*
10075 * Whirlpool hash implementation is added.
10079 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10080 bn(64,32). Because of instruction set limitations it doesn't have
10081 any negative impact on performance. This was done mostly in order
10082 to make it possible to share assembler modules, such as bn_mul_mont
10083 implementations, between 32- and 64-bit builds without hassle.
10087 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10088 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10093 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10094 dedicated Montgomery multiplication procedure, is introduced.
10095 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10096 "64-bit" performance on certain 32-bit targets.
10100 * New option SSL_OP_NO_COMP to disable use of compression selectively
10101 in SSL structures. New SSL ctrl to set maximum send fragment size.
10102 Save memory by setting the I/O buffer sizes dynamically instead of
10103 using the maximum available value.
10107 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10108 in addition to the text details.
10112 * Very, very preliminary EXPERIMENTAL support for printing of general
10113 ASN1 structures. This currently produces rather ugly output and doesn't
10114 handle several customised structures at all.
10118 * Integrated support for PVK file format and some related formats such
10119 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10120 these in the 'rsa' and 'dsa' utilities.
10124 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10128 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10129 place for the (very old) "NETSCAPE" format certificates which are now
10130 handled using new ASN1 code equivalents.
10134 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10135 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10136 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10140 * Modify CRL distribution points extension code to print out previously
10141 unsupported fields. Enhance extension setting code to allow setting of
10146 * Add print and set support for Issuing Distribution Point CRL extension.
10150 * Change 'Configure' script to enable Camellia by default.
10157 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10159 * When rejecting SSL/TLS records due to an incorrect version number, never
10160 update s->server with a new major version number. As of
10161 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10162 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10163 the previous behavior could result in a read attempt at NULL when
10164 receiving specific incorrect SSL/TLS records once record payload
10165 protection is active. ([CVE-2010-0740])
10167 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10169 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10170 could be crashed if the relevant tables were not present (e.g. chrooted).
10172 *Tomas Hoger <thoger@redhat.com>*
10174 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10176 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10178 *Martin Olsson, Neel Mehta*
10180 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10181 accommodate for stack sorting, always a write lock!).
10185 * On some versions of WIN32 Heap32Next is very slow. This can cause
10186 excessive delays in the RAND_poll(): over a minute. As a workaround
10187 include a time check in the inner Heap32Next loop too.
10191 * The code that handled flushing of data in SSL/TLS originally used the
10192 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10193 the problem outlined in PR#1949. The fix suggested there however can
10194 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10195 of Apache). So instead simplify the code to flush unconditionally.
10196 This should be fine since flushing with no data to flush is a no op.
10200 * Handle TLS versions 2.0 and later properly and correctly use the
10201 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10202 off ancient servers have a habit of sticking around for a while...
10206 * Modify compression code so it frees up structures without using the
10207 ex_data callbacks. This works around a problem where some applications
10208 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10209 restarting) then use compression (e.g. SSL with compression) later.
10210 This results in significant per-connection memory leaks and
10211 has caused some security issues including CVE-2008-1678 and
10216 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10217 change when encrypting or decrypting.
10221 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10222 connect and renegotiate with servers which do not support RI.
10223 Until RI is more widely deployed this option is enabled by default.
10227 * Add "missing" ssl ctrls to clear options and mode.
10231 * If client attempts to renegotiate and doesn't support RI respond with
10232 a no_renegotiation alert as required by RFC5746. Some renegotiating
10233 TLS clients will continue a connection gracefully when they receive
10234 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10235 waiting for a server hello which it will never receive. Now we treat a
10236 received no_renegotiation alert as a fatal error. This is because
10237 applications requesting a renegotiation might well expect it to succeed
10238 and would have no code in place to handle the server denying it so the
10239 only safe thing to do is to terminate the connection.
10243 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10244 peer supports secure renegotiation and 0 otherwise. Print out peer
10245 renegotiation support in s_client/s_server.
10249 * Replace the highly broken and deprecated SPKAC certification method with
10250 the updated NID creation version. This should correctly handle UTF8.
10254 * Implement RFC5746. Re-enable renegotiation but require the extension
10255 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10256 turns out to be a bad idea. It has been replaced by
10257 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10258 SSL_CTX_set_options(). This is really not recommended unless you
10259 know what you are doing.
10261 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10263 * Fixes to stateless session resumption handling. Use initial_ctx when
10264 issuing and attempting to decrypt tickets in case it has changed during
10265 servername handling. Use a non-zero length session ID when attempting
10266 stateless session resumption: this makes it possible to determine if
10267 a resumption has occurred immediately after receiving server hello
10268 (several places in OpenSSL subtly assume this) instead of later in
10273 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10274 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10275 fixes for a few places where the return code is not checked
10278 *Julia Lawall <julia@diku.dk>*
10280 * Add --strict-warnings option to Configure script to include devteam
10281 warnings in other configurations.
10285 * Add support for --libdir option and LIBDIR variable in makefiles. This
10286 makes it possible to install openssl libraries in locations which
10287 have names other than "lib", for example "/usr/lib64" which some
10290 *Steve Henson, based on patch from Jeremy Utley*
10292 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10293 X690 8.9.12 and can produce some misleading textual output of OIDs.
10295 *Steve Henson, reported by Dan Kaminsky*
10297 * Delete MD2 from algorithm tables. This follows the recommendation in
10298 several standards that it is not used in new applications due to
10299 several cryptographic weaknesses. For binary compatibility reasons
10300 the MD2 API is still compiled in by default.
10304 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10309 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10310 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10313 *Guenter <lists@gknw.net>*
10315 * Fix the server certificate chain building code to use X509_verify_cert(),
10316 it used to have an ad-hoc builder which was unable to cope with anything
10317 other than a simple chain.
10319 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10321 * Don't check self signed certificate signatures in X509_verify_cert()
10322 by default (a flag can override this): it just wastes time without
10323 adding any security. As a useful side effect self signed root CAs
10324 with non-FIPS digests are now usable in FIPS mode.
10328 * In dtls1_process_out_of_seq_message() the check if the current message
10329 is already buffered was missing. For every new message was memory
10330 allocated, allowing an attacker to perform an denial of service attack
10331 with sending out of seq handshake messages until there is no memory
10332 left. Additionally every future message was buffered, even if the
10333 sequence number made no sense and would be part of another handshake.
10334 So only messages with sequence numbers less than 10 in advance will be
10335 buffered. ([CVE-2009-1378])
10337 *Robin Seggelmann, discovered by Daniel Mentz*
10339 * Records are buffered if they arrive with a future epoch to be
10340 processed after finishing the corresponding handshake. There is
10341 currently no limitation to this buffer allowing an attacker to perform
10342 a DOS attack with sending records with future epochs until there is no
10343 memory left. This patch adds the pqueue_size() function to determine
10344 the size of a buffer and limits the record buffer to 100 entries.
10347 *Robin Seggelmann, discovered by Daniel Mentz*
10349 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10350 parent structure is freed. ([CVE-2009-1379])
10354 * Handle non-blocking I/O properly in SSL_shutdown() call.
10356 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10358 * Add `2.5.4.*` OIDs
10360 *Ilya O. <vrghost@gmail.com>*
10362 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10364 * Disable renegotiation completely - this fixes a severe security
10365 problem ([CVE-2009-3555]) at the cost of breaking all
10366 renegotiation. Renegotiation can be re-enabled by setting
10367 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10368 run-time. This is really not recommended unless you know what
10373 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10375 * Don't set val to NULL when freeing up structures, it is freed up by
10376 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10377 zeroing past the valid field. ([CVE-2009-0789])
10379 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10381 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10382 checked correctly. This would allow some invalid signed attributes to
10383 appear to verify correctly. ([CVE-2009-0591])
10385 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10387 * Reject UniversalString and BMPString types with invalid lengths. This
10388 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10389 a legal length. ([CVE-2009-0590])
10393 * Set S/MIME signing as the default purpose rather than setting it
10394 unconditionally. This allows applications to override it at the store
10399 * Permit restricted recursion of ASN1 strings. This is needed in practice
10400 to handle some structures.
10404 * Improve efficiency of mem_gets: don't search whole buffer each time
10407 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10409 * New -hex option for openssl rand.
10413 * Print out UTF8String and NumericString when parsing ASN1.
10417 * Support NumericString type for name components.
10421 * Allow CC in the environment to override the automatically chosen
10422 compiler. Note that nothing is done to ensure flags work with the
10427 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10429 * Properly check EVP_VerifyFinal() and similar return values
10432 *Ben Laurie, Bodo Moeller, Google Security Team*
10434 * Enable TLS extensions by default.
10438 * Allow the CHIL engine to be loaded, whether the application is
10439 multithreaded or not. (This does not release the developer from the
10440 obligation to set up the dynamic locking callbacks.)
10442 *Sander Temme <sander@temme.net>*
10444 * Use correct exit code if there is an error in dgst command.
10446 *Steve Henson; problem pointed out by Roland Dirlewanger*
10448 * Tweak Configure so that you need to say "experimental-jpake" to enable
10449 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10453 * Add experimental JPAKE support, including demo authentication in
10454 s_client and s_server.
10458 * Set the comparison function in v3_addr_canonize().
10460 *Rob Austein <sra@hactrn.net>*
10462 * Add support for XMPP STARTTLS in s_client.
10464 *Philip Paeps <philip@freebsd.org>*
10466 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10467 to ensure that even with this option, only ciphersuites in the
10468 server's preference list will be accepted. (Note that the option
10469 applies only when resuming a session, so the earlier behavior was
10470 just about the algorithm choice for symmetric cryptography.)
10474 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10476 * Fix NULL pointer dereference if a DTLS server received
10477 ChangeCipherSpec as first record ([CVE-2009-1386]).
10481 * Fix a state transition in s3_srvr.c and d1_srvr.c
10482 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10484 *Nagendra Modadugu*
10486 * The fix in 0.9.8c that supposedly got rid of unsafe
10487 double-checked locking was incomplete for RSA blinding,
10488 addressing just one layer of what turns out to have been
10489 doubly unsafe triple-checked locking.
10491 So now fix this for real by retiring the MONT_HELPER macro
10492 in crypto/rsa/rsa_eay.c.
10494 *Bodo Moeller; problem pointed out by Marius Schilder*
10496 * Various precautionary measures:
10498 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10500 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10501 (NB: This would require knowledge of the secret session ticket key
10502 to exploit, in which case you'd be SOL either way.)
10504 - Change bn_nist.c so that it will properly handle input BIGNUMs
10505 outside the expected range.
10507 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10510 *Neel Mehta, Bodo Moeller*
10512 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10513 the load fails. Useful for distros.
10515 *Ben Laurie and the FreeBSD team*
10517 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10521 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10525 * Expand ENGINE to support engine supplied SSL client certificate functions.
10527 This work was sponsored by Logica.
10531 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10532 keystores. Support for SSL/TLS client authentication too.
10533 Not compiled unless enable-capieng specified to Configure.
10535 This work was sponsored by Logica.
10539 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10540 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10541 attribute creation routines such as certificate requests and PKCS#12
10546 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10548 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10549 handshake which could lead to a client crash as found using the
10550 Codenomicon TLS test suite ([CVE-2008-1672])
10552 *Steve Henson, Mark Cox*
10554 * Fix double free in TLS server name extensions which could lead to
10555 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10559 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10561 Clear the error queue to ensure that error entries left from
10562 older function calls do not interfere with the correct operation.
10564 *Lutz Jaenicke, Erik de Castro Lopo*
10566 * Remove root CA certificates of commercial CAs:
10568 The OpenSSL project does not recommend any specific CA and does not
10569 have any policy with respect to including or excluding any CA.
10570 Therefore, it does not make any sense to ship an arbitrary selection
10571 of root CA certificates with the OpenSSL software.
10575 * RSA OAEP patches to fix two separate invalid memory reads.
10576 The first one involves inputs when 'lzero' is greater than
10577 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10578 before the beginning of from). The second one involves inputs where
10579 the 'db' section contains nothing but zeroes (there is a one-byte
10580 invalid read after the end of 'db').
10582 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10584 * Partial backport from 0.9.9-dev:
10586 Introduce bn_mul_mont (dedicated Montgomery multiplication
10587 procedure) as a candidate for BIGNUM assembler implementation.
10588 While 0.9.9-dev uses assembler for various architectures, only
10589 x86_64 is available by default here in the 0.9.8 branch, and
10590 32-bit x86 is available through a compile-time setting.
10592 To try the 32-bit x86 assembler implementation, use Configure
10593 option "enable-montasm" (which exists only for this backport).
10595 As "enable-montasm" for 32-bit x86 disclaims code stability
10596 anyway, in this constellation we activate additional code
10597 backported from 0.9.9-dev for further performance improvements,
10598 namely BN_from_montgomery_word. (To enable this otherwise,
10599 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10601 *Andy Polyakov (backport partially by Bodo Moeller)*
10603 * Add TLS session ticket callback. This allows an application to set
10604 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10605 values. This is useful for key rollover for example where several key
10606 sets may exist with different names.
10610 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10611 This was broken until now in 0.9.8 releases, such that the only way
10612 a registered ENGINE could be used (assuming it initialises
10613 successfully on the host) was to explicitly set it as the default
10614 for the relevant algorithms. This is in contradiction with 0.9.7
10615 behaviour and the documentation. With this fix, when an ENGINE is
10616 registered into a given algorithm's table of implementations, the
10617 'uptodate' flag is reset so that auto-discovery will be used next
10618 time a new context for that algorithm attempts to select an
10621 *Ian Lister (tweaked by Geoff Thorpe)*
10623 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10624 implementation in the following ways:
10626 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10629 Lack of BER streaming support means one pass streaming processing is
10630 only supported if data is detached: setting the streaming flag is
10631 ignored for embedded content.
10633 CMS support is disabled by default and must be explicitly enabled
10634 with the enable-cms configuration option.
10638 * Update the GMP engine glue to do direct copies between BIGNUM and
10639 mpz_t when openssl and GMP use the same limb size. Otherwise the
10640 existing "conversion via a text string export" trick is still used.
10642 *Paul Sheer <paulsheer@gmail.com>*
10644 * Zlib compression BIO. This is a filter BIO which compressed and
10645 uncompresses any data passed through it.
10649 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10650 RFC3394 compatible AES key wrapping.
10654 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10655 sets string data without copying. X509_ALGOR_set0() and
10656 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10657 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10658 from an X509_ATTRIBUTE structure optionally checking it occurs only
10659 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10664 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10665 to get the expected BN_FLG_CONSTTIME behavior.
10667 *Bodo Moeller (Google)*
10671 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10672 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10673 - added some more tests to do_tests.pl
10674 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10675 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10676 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10677 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10678 - various changes to netware.pl to enable gcc-cross builds on Win32
10680 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10681 - various changes to fix missing prototype warnings
10682 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10683 - added AES, WHIRLPOOL and CPUID assembler code to build files
10684 - added missing AES assembler make rules to mk1mf.pl
10685 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10687 *Guenter Knauf <eflash@gmx.net>*
10689 * Implement certificate status request TLS extension defined in RFC3546.
10690 A client can set the appropriate parameters and receive the encoded
10691 OCSP response via a callback. A server can query the supplied parameters
10692 and set the encoded OCSP response in the callback. Add simplified examples
10693 to s_client and s_server.
10697 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10699 * Fix various bugs:
10700 + Binary incompatibility of ssl_ctx_st structure
10701 + DTLS interoperation with non-compliant servers
10702 + Don't call get_session_cb() without proposed session
10703 + Fix ia64 assembler code
10705 *Andy Polyakov, Steve Henson*
10707 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10709 * DTLS Handshake overhaul. There were longstanding issues with
10710 OpenSSL DTLS implementation, which were making it impossible for
10711 RFC 4347 compliant client to communicate with OpenSSL server.
10712 Unfortunately just fixing these incompatibilities would "cut off"
10713 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10714 server keeps tolerating non RFC compliant syntax. The opposite is
10715 not true, 0.9.8f client can not communicate with earlier server.
10716 This update even addresses CVE-2007-4995.
10720 * Changes to avoid need for function casts in OpenSSL: some compilers
10721 (gcc 4.2 and later) reject their use.
10722 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10725 * Add RFC4507 support to OpenSSL. This includes the corrections in
10726 RFC4507bis. The encrypted ticket format is an encrypted encoded
10727 SSL_SESSION structure, that way new session features are automatically
10730 If a client application caches session in an SSL_SESSION structure
10731 support is transparent because tickets are now stored in the encoded
10734 The SSL_CTX structure automatically generates keys for ticket
10735 protection in servers so again support should be possible
10736 with no application modification.
10738 If a client or server wishes to disable RFC4507 support then the option
10739 SSL_OP_NO_TICKET can be set.
10741 Add a TLS extension debugging callback to allow the contents of any client
10742 or server extensions to be examined.
10744 This work was sponsored by Google.
10748 * Add initial support for TLS extensions, specifically for the server_name
10749 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10750 have new members for a hostname. The SSL data structure has an
10751 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10752 stored in that context to allow for session resumption, even after the
10753 SSL has been switched to a new SSL_CTX in reaction to a client's
10754 server_name extension.
10756 New functions (subject to change):
10758 SSL_get_servername()
10759 SSL_get_servername_type()
10762 New CTRL codes and macros (subject to change):
10764 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10765 - SSL_CTX_set_tlsext_servername_callback()
10766 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10767 - SSL_CTX_set_tlsext_servername_arg()
10768 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10770 openssl s_client has a new '-servername ...' option.
10772 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10773 '-key2 ...', '-servername_fatal' (subject to change). This allows
10774 testing the HostName extension for a specific single hostname ('-cert'
10775 and '-key' remain fallbacks for handshakes without HostName
10776 negotiation). If the unrecognized_name alert has to be sent, this by
10777 default is a warning; it becomes fatal with the '-servername_fatal'
10780 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10782 * Add AES and SSE2 assembly language support to VC++ build.
10786 * Mitigate attack on final subtraction in Montgomery reduction.
10790 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10791 (which previously caused an internal error).
10795 * Squeeze another 10% out of IGE mode when in != out.
10799 * AES IGE mode speedup.
10801 *Dean Gaudet (Google)*
10803 * Add the Korean symmetric 128-bit cipher SEED (see
10804 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10805 add SEED ciphersuites from RFC 4162:
10807 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10808 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10809 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10810 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10812 To minimize changes between patchlevels in the OpenSSL 0.9.8
10813 series, SEED remains excluded from compilation unless OpenSSL
10814 is configured with 'enable-seed'.
10816 *KISA, Bodo Moeller*
10818 * Mitigate branch prediction attacks, which can be practical if a
10819 single processor is shared, allowing a spy process to extract
10820 information. For detailed background information, see
10821 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10822 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10823 and Necessary Software Countermeasures"). The core of the change
10824 are new versions BN_div_no_branch() and
10825 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10826 respectively, which are slower, but avoid the security-relevant
10827 conditional branches. These are automatically called by BN_div()
10828 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10829 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10830 remove a conditional branch.
10832 BN_FLG_CONSTTIME is the new name for the previous
10833 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10834 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10835 in the exponent causes BN_mod_exp_mont() to use the alternative
10836 implementation in BN_mod_exp_mont_consttime().) The old name
10837 remains as a deprecated alias.
10839 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10840 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10841 constant-time implementations for more than just exponentiation.
10842 Here too the old name is kept as a deprecated alias.
10844 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10845 the BN_BLINDING structure gets an independent copy of the
10846 modulus. This means that the previous `BIGNUM *m` argument to
10847 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10848 essentially becomes `const BIGNUM *m`, although we can't actually
10849 change this in the header file before 0.9.9. It allows
10850 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10851 enable BN_FLG_CONSTTIME.
10853 *Matthew D Wood (Intel Corp)*
10855 * In the SSL/TLS server implementation, be strict about session ID
10856 context matching (which matters if an application uses a single
10857 external cache for different purposes). Previously,
10858 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10859 set. This did ensure strict client verification, but meant that,
10860 with applications using a single external cache for quite
10861 different requirements, clients could circumvent ciphersuite
10862 restrictions for a given session ID context by starting a session
10863 in a different context.
10867 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10868 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10869 authentication-only ciphersuites.
10873 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10874 not complete and could lead to a possible single byte overflow
10875 ([CVE-2007-5135]) [Ben Laurie]
10877 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10879 * Since AES128 and AES256 (and similarly Camellia128 and
10880 Camellia256) share a single mask bit in the logic of
10881 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10882 kludge to work properly if AES128 is available and AES256 isn't
10883 (or if Camellia128 is available and Camellia256 isn't).
10887 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10888 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10889 When a point or a seed is encoded in a BIT STRING, we need to
10890 prevent the removal of trailing zero bits to get the proper DER
10891 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10892 of a NamedBitList, for which trailing 0 bits need to be removed.)
10896 * Have SSL/TLS server implementation tolerate "mismatched" record
10897 protocol version while receiving ClientHello even if the
10898 ClientHello is fragmented. (The server can't insist on the
10899 particular protocol version it has chosen before the ServerHello
10900 message has informed the client about his choice.)
10904 * Add RFC 3779 support.
10906 *Rob Austein for ARIN, Ben Laurie*
10908 * Load error codes if they are not already present instead of using a
10909 static variable. This allows them to be cleanly unloaded and reloaded.
10910 Improve header file function name parsing.
10914 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10915 or CAPABILITY handshake as required by RFCs.
10917 *Goetz Babin-Ebell*
10919 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10921 * Introduce limits to prevent malicious keys being able to
10922 cause a denial of service. ([CVE-2006-2940])
10924 *Steve Henson, Bodo Moeller*
10926 * Fix ASN.1 parsing of certain invalid structures that can result
10927 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10929 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10930 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10932 * Fix SSL client code which could crash if connecting to a
10933 malicious SSLv2 server. ([CVE-2006-4343])
10935 *Tavis Ormandy and Will Drewry, Google Security Team*
10937 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10938 match only those. Before that, "AES256-SHA" would be interpreted
10939 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10940 the same strength classification in 0.9.7h) as we currently only
10941 have a single AES bit in the ciphersuite description bitmap.
10942 That change, however, also applied to ciphersuite strings such as
10943 "RC4-MD5" that intentionally matched multiple ciphersuites --
10944 namely, SSL 2.0 ciphersuites in addition to the more common ones
10945 from SSL 3.0/TLS 1.0.
10947 So we change the selection algorithm again: Naming an explicit
10948 ciphersuite selects this one ciphersuite, and any other similar
10949 ciphersuite (same bitmap) from *other* protocol versions.
10950 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10951 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10953 Since SSL 2.0 does not have any ciphersuites for which the
10954 128/256 bit distinction would be relevant, this works for now.
10955 The proper fix will be to use different bits for AES128 and
10956 AES256, which would have avoided the problems from the beginning;
10957 however, bits are scarce, so we can only do this in a new release
10958 (not just a patchlevel) when we can change the SSL_CIPHER
10959 definition to split the single 'unsigned long mask' bitmap into
10960 multiple values to extend the available space.
10964 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10966 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10967 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10969 * Add AES IGE and biIGE modes.
10973 * Change the Unix randomness entropy gathering to use poll() when
10974 possible instead of select(), since the latter has some
10975 undesirable limitations.
10977 *Darryl Miles via Richard Levitte and Bodo Moeller*
10979 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10980 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10981 cannot be implicitly activated as part of, e.g., the "AES" alias.
10982 However, please upgrade to OpenSSL 0.9.9[-dev] for
10983 non-experimental use of the ECC ciphersuites to get TLS extension
10984 support, which is required for curve and point format negotiation
10985 to avoid potential handshake problems.
10989 * Disable rogue ciphersuites:
10991 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10992 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10993 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10995 The latter two were purportedly from
10996 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10999 Also deactivate the remaining ciphersuites from
11000 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11001 unofficial, and the ID has long expired.
11005 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11006 dual-core machines) and other potential thread-safety issues.
11010 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
11011 versions), which is now available for royalty-free use
11012 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
11013 Also, add Camellia TLS ciphersuites from RFC 4132.
11015 To minimize changes between patchlevels in the OpenSSL 0.9.8
11016 series, Camellia remains excluded from compilation unless OpenSSL
11017 is configured with 'enable-camellia'.
11021 * Disable the padding bug check when compression is in use. The padding
11022 bug check assumes the first packet is of even length, this is not
11023 necessarily true if compression is enabled and can result in false
11024 positives causing handshake failure. The actual bug test is ancient
11025 code so it is hoped that implementations will either have fixed it by
11026 now or any which still have the bug do not support compression.
11030 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
11032 * When applying a cipher rule check to see if string match is an explicit
11033 cipher suite and only match that one cipher suite if it is.
11037 * Link in manifests for VC++ if needed.
11039 *Austin Ziegler <halostatue@gmail.com>*
11041 * Update support for ECC-based TLS ciphersuites according to
11042 draft-ietf-tls-ecc-12.txt with proposed changes (but without
11043 TLS extensions, which are supported starting with the 0.9.9
11044 branch, not in the OpenSSL 0.9.8 branch).
11048 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
11049 opaque EVP_CIPHER_CTX handling.
11053 * Fixes and enhancements to zlib compression code. We now only use
11054 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
11055 to conform with the standards mentioned here:
11056 <http://www.zlib.net/DLL_FAQ.txt>
11057 Static zlib linking now works on Windows and the new --with-zlib-include
11058 --with-zlib-lib options to Configure can be used to supply the location
11059 of the headers and library. Gracefully handle case where zlib library
11064 * Several fixes and enhancements to the OID generation code. The old code
11065 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11066 handle numbers larger than ULONG_MAX, truncated printing and had a
11067 non standard OBJ_obj2txt() behaviour.
11071 * Add support for building of engines under engine/ as shared libraries
11072 under VC++ build system.
11076 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11077 Hopefully, we will not see any false combination of paths any more.
11081 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11083 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11084 (part of SSL_OP_ALL). This option used to disable the
11085 countermeasure against man-in-the-middle protocol-version
11086 rollback in the SSL 2.0 server implementation, which is a bad
11087 idea. ([CVE-2005-2969])
11089 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11090 for Information Security, National Institute of Advanced Industrial
11091 Science and Technology [AIST], Japan)*
11093 * Add two function to clear and return the verify parameter flags.
11097 * Keep cipherlists sorted in the source instead of sorting them at
11098 runtime, thus removing the need for a lock.
11102 * Avoid some small subgroup attacks in Diffie-Hellman.
11104 *Nick Mathewson and Ben Laurie*
11106 * Add functions for well-known primes.
11110 * Extended Windows CE support.
11112 *Satoshi Nakamura and Andy Polyakov*
11114 * Initialize SSL_METHOD structures at compile time instead of during
11115 runtime, thus removing the need for a lock.
11119 * Make PKCS7_decrypt() work even if no certificate is supplied by
11120 attempting to decrypt each encrypted key in turn. Add support to
11125 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11127 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11130 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11134 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11135 key into the same file any more.
11139 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11143 * Add -utf8 command line and config file option to 'ca'.
11145 *Stefan <stf@udoma.org*
11147 * Removed the macro des_crypt(), as it seems to conflict with some
11148 libraries. Use DES_crypt().
11152 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11153 involves renaming the source and generated shared-libs for
11154 both. The engines will accept the corrected or legacy ids
11155 ('ncipher' and '4758_cca' respectively) when binding. NB,
11156 this only applies when building 'shared'.
11158 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11160 * Add attribute functions to EVP_PKEY structure. Modify
11161 PKCS12_create() to recognize a CSP name attribute and
11162 use it. Make -CSP option work again in pkcs12 utility.
11166 * Add new functionality to the bn blinding code:
11167 - automatic re-creation of the BN_BLINDING parameters after
11168 a fixed number of uses (currently 32)
11169 - add new function for parameter creation
11170 - introduce flags to control the update behaviour of the
11171 BN_BLINDING parameters
11172 - hide BN_BLINDING structure
11173 Add a second BN_BLINDING slot to the RSA structure to improve
11174 performance when a single RSA object is shared among several
11179 * Add support for DTLS.
11181 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11183 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11184 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11188 * Remove buggy and incomplete DH cert support from
11189 ssl/ssl_rsa.c and ssl/s3_both.c
11193 * Use SHA-1 instead of MD5 as the default digest algorithm for
11194 the `apps/openssl` commands.
11198 * Compile clean with "-Wall -Wmissing-prototypes
11199 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11200 DEBUG_SAFESTACK must also be set.
11204 * Change ./Configure so that certain algorithms can be disabled by default.
11205 The new counterpiece to "no-xxx" is "enable-xxx".
11207 The patented RC5 and MDC2 algorithms will now be disabled unless
11208 "enable-rc5" and "enable-mdc2", respectively, are specified.
11210 (IDEA remains enabled despite being patented. This is because IDEA
11211 is frequently required for interoperability, and there is no license
11212 fee for non-commercial use. As before, "no-idea" can be used to
11213 avoid this algorithm.)
11217 * Add processing of proxy certificates (see RFC 3820). This work was
11218 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11219 EGEE (Enabling Grids for E-science in Europe).
11223 * RC4 performance overhaul on modern architectures/implementations, such
11224 as Intel P4, IA-64 and AMD64.
11228 * New utility extract-section.pl. This can be used specify an alternative
11229 section number in a pod file instead of having to treat each file as
11230 a separate case in Makefile. This can be done by adding two lines to the
11233 =for comment openssl_section:XXX
11235 The blank line is mandatory.
11239 * New arguments -certform, -keyform and -pass for s_client and s_server
11240 to allow alternative format key and certificate files and passphrase
11245 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11246 update associated structures and add various utility functions.
11248 Add new policy related verify parameters, include policy checking in
11249 standard verify code. Enhance 'smime' application with extra parameters
11250 to support policy checking and print out.
11254 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11255 Nehemiah processors. These extensions support AES encryption in hardware
11256 as well as RNG (though RNG support is currently disabled).
11258 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11260 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11264 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11266 *Andy Polyakov and a number of other people*
11268 * Improved PowerPC platform support. Most notably BIGNUM assembler
11269 implementation contributed by IBM.
11271 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11273 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11274 exponent rather than 'unsigned long'. There is a corresponding change to
11275 the new 'rsa_keygen' element of the RSA_METHOD structure.
11277 *Jelte Jansen, Geoff Thorpe*
11279 * Functionality for creating the initial serial number file is now
11280 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11282 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11283 number file to 1, which is bound to cause problems. To avoid
11284 the problems while respecting compatibility between different 0.9.7
11285 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11286 CA.pl for serial number initialization. With the new release 0.9.8,
11287 we can fix the problem directly in the 'ca' utility.)
11291 * Reduced header interdependencies by declaring more opaque objects in
11292 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11293 give fewer recursive includes, which could break lazy source code - so
11294 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11295 developers should define this symbol when building and using openssl to
11296 ensure they track the recommended behaviour, interfaces, [etc], but
11297 backwards-compatible behaviour prevails when this isn't defined.
11301 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11305 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11306 This will generate a random key of the appropriate length based on the
11307 cipher context. The EVP_CIPHER can provide its own random key generation
11308 routine to support keys of a specific form. This is used in the des and
11309 3des routines to generate a key of the correct parity. Update S/MIME
11310 code to use new functions and hence generate correct parity DES keys.
11311 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11312 valid (weak or incorrect parity).
11316 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11317 as looking them up. This is useful when the verified structure may contain
11318 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11319 present unless the new PKCS7_NO_CRL flag is asserted.
11323 * Extend ASN1 oid configuration module. It now additionally accepts the
11326 shortName = some long name, 1.2.3.4
11330 * Reimplemented the BN_CTX implementation. There is now no more static
11331 limitation on the number of variables it can handle nor the depth of the
11332 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11333 information can now expand as required, and rather than having a single
11334 static array of bignums, BN_CTX now uses a linked-list of such arrays
11335 allowing it to expand on demand whilst maintaining the usefulness of
11336 BN_CTX's "bundling".
11340 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11341 to allow all RSA operations to function using a single BN_CTX.
11345 * Preliminary support for certificate policy evaluation and checking. This
11346 is initially intended to pass the tests outlined in "Conformance Testing
11347 of Relying Party Client Certificate Path Processing Logic" v1.07.
11351 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11352 remained unused and not that useful. A variety of other little bignum
11353 tweaks and fixes have also been made continuing on from the audit (see
11358 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11359 associated ASN1, EVP and SSL functions and old ASN1 macros.
11363 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11364 and this should never fail. So the return value from the use of
11365 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11366 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11370 * BN_CTX_get() should return zero-valued bignums, providing the same
11371 initialised value as BN_new().
11373 *Geoff Thorpe, suggested by Ulf Möller*
11375 * Support for inhibitAnyPolicy certificate extension.
11379 * An audit of the BIGNUM code is underway, for which debugging code is
11380 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11381 is considered valid when processing BIGNUMs, and causes execution to
11382 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11383 further steps are taken to deliberately pollute unused data in BIGNUM
11384 structures to try and expose faulty code further on. For now, openssl will
11385 (in its default mode of operation) continue to tolerate the inconsistent
11386 forms that it has tolerated in the past, but authors and packagers should
11387 consider trying openssl and their own applications when compiled with
11388 these debugging symbols defined. It will help highlight potential bugs in
11389 their own code, and will improve the test coverage for OpenSSL itself. At
11390 some point, these tighter rules will become openssl's default to improve
11391 maintainability, though the assert()s and other overheads will remain only
11392 in debugging configurations. See bn.h for more details.
11394 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11396 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11397 that can only be obtained through BN_CTX_new() (which implicitly
11398 initialises it). The presence of this function only made it possible
11399 to overwrite an existing structure (and cause memory leaks).
11403 * Because of the callback-based approach for implementing LHASH as a
11404 template type, lh_insert() adds opaque objects to hash-tables and
11405 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11406 to clean up those corresponding objects before destroying the hash table
11407 (and losing the object pointers). So some over-zealous constifications in
11408 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11409 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11410 prototyped to have "const" restrictions on the object pointers they are
11411 given (and so aren't required to cast them away any more).
11415 * The tmdiff.h API was so ugly and minimal that our own timing utility
11416 (speed) prefers to use its own implementation. The two implementations
11417 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11418 its object type properly exposed (MS_TM) instead of casting to/from
11419 `char *`. This may still change yet if someone realises MS_TM and
11421 aren't necessarily the greatest nomenclatures - but this is what was used
11422 internally to the implementation so I've used that for now.
11426 * Ensure that deprecated functions do not get compiled when
11427 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11428 the self-tests were still using deprecated key-generation functions so
11429 these have been updated also.
11433 * Reorganise PKCS#7 code to separate the digest location functionality
11434 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11435 New function PKCS7_set_digest() to set the digest type for PKCS#7
11436 digestedData type. Add additional code to correctly generate the
11437 digestedData type and add support for this type in PKCS7 initialization
11442 * New function PKCS7_set0_type_other() this initializes a PKCS7
11443 structure of type "other".
11447 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11448 sure the loop does correctly stop and breaking ("division by zero")
11449 modulus operations are not performed. The (pre-generated) prime
11450 table crypto/bn/bn_prime.h was already correct, but it could not be
11451 re-generated on some platforms because of the "division by zero"
11452 situation in the script.
11454 *Ralf S. Engelschall*
11456 * Update support for ECC-based TLS ciphersuites according to
11457 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11458 SHA-1 now is only used for "small" curves (where the
11459 representation of a field element takes up to 24 bytes); for
11460 larger curves, the field element resulting from ECDH is directly
11461 used as premaster secret.
11463 *Douglas Stebila (Sun Microsystems Laboratories)*
11465 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11466 curve secp160r1 to the tests.
11468 *Douglas Stebila (Sun Microsystems Laboratories)*
11470 * Add the possibility to load symbols globally with DSO.
11472 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11474 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11475 control of the error stack.
11479 * Add support for STORE in ENGINE.
11483 * Add the STORE type. The intention is to provide a common interface
11484 to certificate and key stores, be they simple file-based stores, or
11485 HSM-type store, or LDAP stores, or...
11486 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11490 * Add a generic structure called OPENSSL_ITEM. This can be used to
11491 pass a list of arguments to any function as well as provide a way
11492 for a function to pass data back to the caller.
11496 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11497 works like BUF_strdup() but can be used to duplicate a portion of
11498 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11503 * Add the function sk_find_ex() which works like sk_find(), but will
11504 return an index to an element even if an exact match couldn't be
11505 found. The index is guaranteed to point at the element where the
11506 searched-for key would be inserted to preserve sorting order.
11510 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11511 takes an extra flags argument for optional functionality. Currently,
11512 the following flags are defined:
11514 OBJ_BSEARCH_VALUE_ON_NOMATCH
11515 This one gets OBJ_bsearch_ex() to return a pointer to the first
11516 element where the comparing function returns a negative or zero
11519 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11520 This one gets OBJ_bsearch_ex() to return a pointer to the first
11521 element where the comparing function returns zero. This is useful
11522 if there are more than one element where the comparing function
11527 * Make it possible to create self-signed certificates with 'openssl ca'
11528 in such a way that the self-signed certificate becomes part of the
11529 CA database and uses the same mechanisms for serial number generation
11530 as all other certificate signing. The new flag '-selfsign' enables
11531 this functionality. Adapt CA.sh and CA.pl.in.
11535 * Add functionality to check the public key of a certificate request
11536 against a given private. This is useful to check that a certificate
11537 request can be signed by that key (self-signing).
11541 * Make it possible to have multiple active certificates with the same
11542 subject in the CA index file. This is done only if the keyword
11543 'unique_subject' is set to 'no' in the main CA section (default
11544 if 'CA_default') of the configuration file. The value is saved
11545 with the database itself in a separate index attribute file,
11546 named like the index file with '.attr' appended to the name.
11550 * Generate multi-valued AVAs using '+' notation in config files for
11555 * Support for nameConstraints certificate extension.
11559 * Support for policyConstraints certificate extension.
11563 * Support for policyMappings certificate extension.
11567 * Make sure the default DSA_METHOD implementation only uses its
11568 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11569 and change its own handlers to be NULL so as to remove unnecessary
11570 indirection. This lets alternative implementations fallback to the
11571 default implementation more easily.
11575 * Support for directoryName in GeneralName related extensions
11580 * Make it possible to link applications using Makefile.shared.
11581 Make that possible even when linking against static libraries!
11585 * Support for single pass processing for S/MIME signing. This now
11586 means that S/MIME signing can be done from a pipe, in addition
11587 cleartext signing (multipart/signed type) is effectively streaming
11588 and the signed data does not need to be all held in memory.
11590 This is done with a new flag PKCS7_STREAM. When this flag is set
11591 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11592 is done after the data is output (and digests calculated) in
11593 SMIME_write_PKCS7().
11597 * Add full support for -rpath/-R, both in shared libraries and
11598 applications, at least on the platforms where it's known how
11603 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11604 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11605 will now compute a table of multiples of the generator that
11606 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11607 faster (notably in the case of a single point multiplication,
11608 scalar * generator).
11610 *Nils Larsch, Bodo Moeller*
11612 * IPv6 support for certificate extensions. The various extensions
11613 which use the IP:a.b.c.d can now take IPv6 addresses using the
11614 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11619 * Added an ENGINE that implements RSA by performing private key
11620 exponentiations with the GMP library. The conversions to and from
11621 GMP's mpz_t format aren't optimised nor are any montgomery forms
11622 cached, and on x86 it appears OpenSSL's own performance has caught up.
11623 However there are likely to be other architectures where GMP could
11624 provide a boost. This ENGINE is not built in by default, but it can be
11625 specified at Configure time and should be accompanied by the necessary
11626 linker additions, eg;
11627 ./config -DOPENSSL_USE_GMP -lgmp
11631 * "openssl engine" will not display ENGINE/DSO load failure errors when
11632 testing availability of engines with "-t" - the old behaviour is
11633 produced by increasing the feature's verbosity with "-tt".
11637 * ECDSA routines: under certain error conditions uninitialized BN objects
11638 could be freed. Solution: make sure initialization is performed early
11639 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11644 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11645 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11646 software implementations. For DSA and DH, parameter generation can
11647 also be overridden by providing the appropriate method callbacks.
11651 * Change the "progress" mechanism used in key-generation and
11652 primality testing to functions that take a new BN_GENCB pointer in
11653 place of callback/argument pairs. The new API functions have `_ex`
11654 postfixes and the older functions are reimplemented as wrappers for
11655 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11656 declarations of the old functions to help (graceful) attempts to
11657 migrate to the new functions. Also, the new key-generation API
11658 functions operate on a caller-supplied key-structure and return
11659 success/failure rather than returning a key or NULL - this is to
11660 help make "keygen" another member function of RSA_METHOD etc.
11662 Example for using the new callback interface:
11664 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11665 void *my_arg = ...;
11668 BN_GENCB_set(&my_cb, my_callback, my_arg);
11670 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11671 /* For the meaning of a, b in calls to my_callback(), see the
11672 * documentation of the function that calls the callback.
11673 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11674 * my_callback should return 1 if it wants BN_is_prime_ex()
11675 * to continue, or 0 to stop.
11680 * Change the ZLIB compression method to be stateful, and make it
11681 available to TLS with the number defined in
11682 draft-ietf-tls-compression-04.txt.
11686 * Add the ASN.1 structures and functions for CertificatePair, which
11687 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11689 CertificatePair ::= SEQUENCE {
11690 forward [0] Certificate OPTIONAL,
11691 reverse [1] Certificate OPTIONAL,
11692 -- at least one of the pair shall be present -- }
11694 Also implement the PEM functions to read and write certificate
11695 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11697 This needed to be defined, mostly for the sake of the LDAP
11698 attribute crossCertificatePair, but may prove useful elsewhere as
11703 * Make it possible to inhibit symlinking of shared libraries in
11704 Makefile.shared, for Cygwin's sake.
11708 * Extend the BIGNUM API by creating a function
11709 void BN_set_negative(BIGNUM *a, int neg);
11710 and a macro that behave like
11711 int BN_is_negative(const BIGNUM *a);
11713 to avoid the need to access 'a->neg' directly in applications.
11717 * Implement fast modular reduction for pseudo-Mersenne primes
11718 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11719 EC_GROUP_new_curve_GFp() will now automatically use this
11722 *Nils Larsch <nla@trustcenter.de>*
11724 * Add new lock type (CRYPTO_LOCK_BN).
11728 * Change the ENGINE framework to automatically load engines
11729 dynamically from specific directories unless they could be
11730 found to already be built in or loaded. Move all the
11731 current engines except for the cryptodev one to a new
11732 directory engines/.
11733 The engines in engines/ are built as shared libraries if
11734 the "shared" options was given to ./Configure or ./config.
11735 Otherwise, they are inserted in libcrypto.a.
11736 /usr/local/ssl/engines is the default directory for dynamic
11737 engines, but that can be overridden at configure time through
11738 the usual use of --prefix and/or --openssldir, and at run
11739 time with the environment variable OPENSSL_ENGINES.
11741 *Geoff Thorpe and Richard Levitte*
11743 * Add Makefile.shared, a helper makefile to build shared
11744 libraries. Adapt Makefile.org.
11748 * Add version info to Win32 DLLs.
11750 *Peter 'Luna' Runestig" <peter@runestig.com>*
11752 * Add new 'medium level' PKCS#12 API. Certificates and keys
11753 can be added using this API to created arbitrary PKCS#12
11754 files while avoiding the low-level API.
11756 New options to PKCS12_create(), key or cert can be NULL and
11757 will then be omitted from the output file. The encryption
11758 algorithm NIDs can be set to -1 for no encryption, the mac
11759 iteration count can be set to 0 to omit the mac.
11761 Enhance pkcs12 utility by making the -nokeys and -nocerts
11762 options work when creating a PKCS#12 file. New option -nomac
11763 to omit the mac, NONE can be set for an encryption algorithm.
11764 New code is modified to use the enhanced PKCS12_create()
11765 instead of the low-level API.
11769 * Extend ASN1 encoder to support indefinite length constructed
11770 encoding. This can output sequences tags and octet strings in
11771 this form. Modify pk7_asn1.c to support indefinite length
11772 encoding. This is experimental and needs additional code to
11773 be useful, such as an ASN1 bio and some enhanced streaming
11776 Extend template encode functionality so that tagging is passed
11777 down to the template encoder.
11781 * Let 'openssl req' fail if an argument to '-newkey' is not
11782 recognized instead of using RSA as a default.
11786 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11787 As these are not official, they are not included in "ALL";
11788 the "ECCdraft" ciphersuite group alias can be used to select them.
11790 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11792 * Add ECDH engine support.
11794 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11796 * Add ECDH in new directory crypto/ecdh/.
11798 *Douglas Stebila (Sun Microsystems Laboratories)*
11800 * Let BN_rand_range() abort with an error after 100 iterations
11801 without success (which indicates a broken PRNG).
11805 * Change BN_mod_sqrt() so that it verifies that the input value
11806 is really the square of the return value. (Previously,
11807 BN_mod_sqrt would show GIGO behaviour.)
11811 * Add named elliptic curves over binary fields from X9.62, SECG,
11812 and WAP/WTLS; add OIDs that were still missing.
11814 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11816 * Extend the EC library for elliptic curves over binary fields
11817 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11820 EC_GF2m_simple_method
11824 EC_GROUP_new_curve_GF2m
11825 EC_GROUP_set_curve_GF2m
11826 EC_GROUP_get_curve_GF2m
11827 EC_POINT_set_affine_coordinates_GF2m
11828 EC_POINT_get_affine_coordinates_GF2m
11829 EC_POINT_set_compressed_coordinates_GF2m
11831 Point compression for binary fields is disabled by default for
11832 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11835 As binary polynomials are represented as BIGNUMs, various members
11836 of the EC_GROUP and EC_POINT data structures can be shared
11837 between the implementations for prime fields and binary fields;
11838 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11839 are essentially identical to their `..._GFp` counterparts.
11840 (For simplicity, the `..._GFp` prefix has been dropped from
11841 various internal method names.)
11843 An internal 'field_div' method (similar to 'field_mul' and
11844 'field_sqr') has been added; this is used only for binary fields.
11846 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11848 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11849 through methods ('mul', 'precompute_mult').
11851 The generic implementations (now internally called 'ec_wNAF_mul'
11852 and 'ec_wNAF_precomputed_mult') remain the default if these
11853 methods are undefined.
11855 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11857 * New function EC_GROUP_get_degree, which is defined through
11858 EC_METHOD. For curves over prime fields, this returns the bit
11859 length of the modulus.
11861 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11863 * New functions EC_GROUP_dup, EC_POINT_dup.
11864 (These simply call ..._new and ..._copy).
11866 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11868 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11869 Polynomials are represented as BIGNUMs (where the sign bit is not
11870 used) in the following functions [macros]:
11873 BN_GF2m_sub [= BN_GF2m_add]
11874 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11875 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11876 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11878 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11879 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11880 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11881 BN_GF2m_cmp [= BN_ucmp]
11883 (Note that only the 'mod' functions are actually for fields GF(2^m).
11884 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11886 For some functions, an the irreducible polynomial defining a
11887 field can be given as an 'unsigned int[]' with strictly
11888 decreasing elements giving the indices of those bits that are set;
11889 i.e., p[] represents the polynomial
11890 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11892 p[0] > p[1] > ... > p[k] = 0.
11893 This applies to the following functions:
11896 BN_GF2m_mod_mul_arr
11897 BN_GF2m_mod_sqr_arr
11898 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11899 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11900 BN_GF2m_mod_exp_arr
11901 BN_GF2m_mod_sqrt_arr
11902 BN_GF2m_mod_solve_quad_arr
11906 Conversion can be performed by the following functions:
11911 bntest.c has additional tests for binary polynomial arithmetic.
11913 Two implementations for BN_GF2m_mod_div() are available.
11914 The default algorithm simply uses BN_GF2m_mod_inv() and
11915 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11916 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11917 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11919 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11921 * Add new error code 'ERR_R_DISABLED' that can be used when some
11922 functionality is disabled at compile-time.
11924 *Douglas Stebila <douglas.stebila@sun.com>*
11926 * Change default behaviour of 'openssl asn1parse' so that more
11927 information is visible when viewing, e.g., a certificate:
11929 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11930 mode the content of non-printable OCTET STRINGs is output in a
11931 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11932 avoid the appearance of a printable string.
11934 *Nils Larsch <nla@trustcenter.de>*
11936 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11938 EC_GROUP_set_asn1_flag()
11939 EC_GROUP_get_asn1_flag()
11940 EC_GROUP_set_point_conversion_form()
11941 EC_GROUP_get_point_conversion_form()
11942 These control ASN1 encoding details:
11943 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11944 has been set to OPENSSL_EC_NAMED_CURVE.
11945 - Points are encoded in uncompressed form by default; options for
11946 asn1_for are as for point2oct, namely
11947 POINT_CONVERSION_COMPRESSED
11948 POINT_CONVERSION_UNCOMPRESSED
11949 POINT_CONVERSION_HYBRID
11951 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11953 EC_GROUP_set_seed()
11954 EC_GROUP_get0_seed()
11955 EC_GROUP_get_seed_len()
11956 This is used only for ASN1 purposes (so far).
11958 *Nils Larsch <nla@trustcenter.de>*
11960 * Add 'field_type' member to EC_METHOD, which holds the NID
11961 of the appropriate field type OID. The new function
11962 EC_METHOD_get_field_type() returns this value.
11964 *Nils Larsch <nla@trustcenter.de>*
11967 EC_POINT_point2bn()
11968 EC_POINT_bn2point()
11969 EC_POINT_point2hex()
11970 EC_POINT_hex2point()
11971 providing useful interfaces to EC_POINT_point2oct() and
11972 EC_POINT_oct2point().
11974 *Nils Larsch <nla@trustcenter.de>*
11976 * Change internals of the EC library so that the functions
11977 EC_GROUP_set_generator()
11978 EC_GROUP_get_generator()
11979 EC_GROUP_get_order()
11980 EC_GROUP_get_cofactor()
11981 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11982 to methods, which would lead to unnecessary code duplication when
11983 adding different types of curves.
11985 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11987 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11988 arithmetic, and such that modified wNAFs are generated
11989 (which avoid length expansion in many cases).
11993 * Add a function EC_GROUP_check_discriminant() (defined via
11994 EC_METHOD) that verifies that the curve discriminant is non-zero.
11996 Add a function EC_GROUP_check() that makes some sanity tests
11997 on a EC_GROUP, its generator and order. This includes
11998 EC_GROUP_check_discriminant().
12000 *Nils Larsch <nla@trustcenter.de>*
12002 * Add ECDSA in new directory crypto/ecdsa/.
12004 Add applications 'openssl ecparam' and 'openssl ecdsa'
12005 (these are based on 'openssl dsaparam' and 'openssl dsa').
12007 ECDSA support is also included in various other files across the
12008 library. Most notably,
12009 - 'openssl req' now has a '-newkey ecdsa:file' option;
12010 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
12011 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
12012 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
12013 them suitable for ECDSA where domain parameters must be
12014 extracted before the specific public key;
12015 - ECDSA engine support has been added.
12017 *Nils Larsch <nla@trustcenter.de>*
12019 * Include some named elliptic curves, and add OIDs from X9.62,
12020 SECG, and WAP/WTLS. Each curve can be obtained from the new
12022 EC_GROUP_new_by_curve_name(),
12023 and the list of available named curves can be obtained with
12024 EC_get_builtin_curves().
12025 Also add a 'curve_name' member to EC_GROUP objects, which can be
12027 EC_GROUP_set_curve_name()
12028 EC_GROUP_get_curve_name()
12030 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
12032 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12033 was actually never needed) and in BN_mul(). The removal in BN_mul()
12034 required a small change in bn_mul_part_recursive() and the addition
12035 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12036 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12037 bn_sub_words() and bn_add_words() except they take arrays with
12042 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
12044 * Cleanse PEM buffers before freeing them since they may contain
12047 *Benjamin Bennett <ben@psc.edu>*
12049 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
12050 a ciphersuite string such as "DEFAULT:RSA" cannot enable
12051 authentication-only ciphersuites.
12055 * Since AES128 and AES256 share a single mask bit in the logic of
12056 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
12057 kludge to work properly if AES128 is available and AES256 isn't.
12061 * Expand security boundary to match 1.1.1 module.
12065 * Remove redundant features: hash file source, editing of test vectors
12066 modify fipsld to use external fips_premain.c signature.
12070 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12071 run algorithm test programs.
12075 * Make algorithm test programs more tolerant of whitespace.
12079 * Have SSL/TLS server implementation tolerate "mismatched" record
12080 protocol version while receiving ClientHello even if the
12081 ClientHello is fragmented. (The server can't insist on the
12082 particular protocol version it has chosen before the ServerHello
12083 message has informed the client about his choice.)
12087 * Load error codes if they are not already present instead of using a
12088 static variable. This allows them to be cleanly unloaded and reloaded.
12092 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12094 * Introduce limits to prevent malicious keys being able to
12095 cause a denial of service. ([CVE-2006-2940])
12097 *Steve Henson, Bodo Moeller*
12099 * Fix ASN.1 parsing of certain invalid structures that can result
12100 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12102 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12103 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12105 * Fix SSL client code which could crash if connecting to a
12106 malicious SSLv2 server. ([CVE-2006-4343])
12108 *Tavis Ormandy and Will Drewry, Google Security Team*
12110 * Change ciphersuite string processing so that an explicit
12111 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12112 will no longer include "AES128-SHA"), and any other similar
12113 ciphersuite (same bitmap) from *other* protocol versions (so that
12114 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12115 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12116 changes from 0.9.8b and 0.9.8d.
12120 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12122 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12123 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12125 * Change the Unix randomness entropy gathering to use poll() when
12126 possible instead of select(), since the latter has some
12127 undesirable limitations.
12129 *Darryl Miles via Richard Levitte and Bodo Moeller*
12131 * Disable rogue ciphersuites:
12133 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12134 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12135 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12137 The latter two were purportedly from
12138 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12141 Also deactivate the remaining ciphersuites from
12142 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12143 unofficial, and the ID has long expired.
12147 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12148 dual-core machines) and other potential thread-safety issues.
12152 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12154 * Adapt fipsld and the build system to link against the validated FIPS
12155 module in FIPS mode.
12159 * Fixes for VC++ 2005 build under Windows.
12163 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12164 from a Windows bash shell such as MSYS. It is autodetected from the
12165 "config" script when run from a VC++ environment. Modify standard VC++
12166 build to use fipscanister.o from the GNU make build.
12170 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12172 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12173 The value now differs depending on if you build for FIPS or not.
12174 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12175 safely run with a non-FIPSed libcrypto, as it may crash because of
12176 the difference induced by this change.
12180 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12182 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12183 (part of SSL_OP_ALL). This option used to disable the
12184 countermeasure against man-in-the-middle protocol-version
12185 rollback in the SSL 2.0 server implementation, which is a bad
12186 idea. ([CVE-2005-2969])
12188 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12189 for Information Security, National Institute of Advanced Industrial
12190 Science and Technology [AIST, Japan)]*
12192 * Minimal support for X9.31 signatures and PSS padding modes. This is
12193 mainly for FIPS compliance and not fully integrated at this stage.
12197 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12198 the exponentiation using a fixed-length exponent. (Otherwise,
12199 the information leaked through timing could expose the secret key
12200 after many signatures; cf. Bleichenbacher's attack on DSA with
12205 * Make a new fixed-window mod_exp implementation the default for
12206 RSA, DSA, and DH private-key operations so that the sequence of
12207 squares and multiplies and the memory access pattern are
12208 independent of the particular secret key. This will mitigate
12209 cache-timing and potential related attacks.
12211 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12212 and this is automatically used by BN_mod_exp_mont() if the new flag
12213 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12214 will use this BN flag for private exponents unless the flag
12215 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12216 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12218 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12220 * Change the client implementation for SSLv23_method() and
12221 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12222 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12223 (Previously, the SSL 2.0 backwards compatible Client Hello
12224 message format would be used even with SSL_OP_NO_SSLv2.)
12228 * Add support for smime-type MIME parameter in S/MIME messages which some
12233 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12234 a threadsafe manner. Modify rsa code to use new function and add calls
12235 to dsa and dh code (which had race conditions before).
12239 * Include the fixed error library code in the C error file definitions
12240 instead of fixing them up at runtime. This keeps the error code
12241 structures constant.
12245 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12247 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12250 * Fixes for newer kerberos headers. NB: the casts are needed because
12251 the 'length' field is signed on one version and unsigned on another
12252 with no (?) obvious way to tell the difference, without these VC++
12253 complains. Also the "definition" of FAR (blank) is no longer included
12254 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12255 some needed definitions.
12259 * Undo Cygwin change.
12263 * Added support for proxy certificates according to RFC 3820.
12264 Because they may be a security thread to unaware applications,
12265 they must be explicitly allowed in run-time. See
12266 docs/HOWTO/proxy_certificates.txt for further information.
12270 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12272 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12273 server and client random values. Previously
12274 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12275 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12277 This change has negligible security impact because:
12279 1. Server and client random values still have 24 bytes of pseudo random
12282 2. Server and client random values are sent in the clear in the initial
12285 3. The master secret is derived using the premaster secret (48 bytes in
12286 size for static RSA ciphersuites) as well as client server and random
12289 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12292 *Stephen Henson, reported by UK NISCC*
12294 * Use Windows randomness collection on Cygwin.
12298 * Fix hang in EGD/PRNGD query when communication socket is closed
12299 prematurely by EGD/PRNGD.
12301 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12303 * Prompt for pass phrases when appropriate for PKCS12 input format.
12307 * Back-port of selected performance improvements from development
12308 branch, as well as improved support for PowerPC platforms.
12312 * Add lots of checks for memory allocation failure, error codes to indicate
12313 failure and freeing up memory if a failure occurs.
12315 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12317 * Add new -passin argument to dgst.
12321 * Perform some character comparisons of different types in X509_NAME_cmp:
12322 this is needed for some certificates that re-encode DNs into UTF8Strings
12323 (in violation of RFC3280) and can't or won't issue name rollover
12328 * Make an explicit check during certificate validation to see that
12329 the CA setting in each certificate on the chain is correct. As a
12330 side effect always do the following basic checks on extensions,
12331 not just when there's an associated purpose to the check:
12333 - if there is an unhandled critical extension (unless the user
12334 has chosen to ignore this fault)
12335 - if the path length has been exceeded (if one is set at all)
12336 - that certain extensions fit the associated purpose (if one has
12341 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12343 * Avoid a race condition when CRLs are checked in a multi threaded
12344 environment. This would happen due to the reordering of the revoked
12345 entries during signature checking and serial number lookup. Now the
12346 encoding is cached and the serial number sort performed under a lock.
12347 Add new STACK function sk_is_sorted().
12351 * Add Delta CRL to the extension code.
12355 * Various fixes to s3_pkt.c so alerts are sent properly.
12357 *David Holmes <d.holmes@f5.com>*
12359 * Reduce the chances of duplicate issuer name and serial numbers (in
12360 violation of RFC3280) using the OpenSSL certificate creation utilities.
12361 This is done by creating a random 64 bit value for the initial serial
12362 number when a serial number file is created or when a self signed
12363 certificate is created using 'openssl req -x509'. The initial serial
12364 number file is created using 'openssl x509 -next_serial' in CA.pl
12365 rather than being initialized to 1.
12369 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12371 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12372 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12374 *Joe Orton, Steve Henson*
12376 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12379 *Joe Orton, Steve Henson*
12381 * Make it possible to have multiple active certificates with the same
12382 subject in the CA index file. This is done only if the keyword
12383 'unique_subject' is set to 'no' in the main CA section (default
12384 if 'CA_default') of the configuration file. The value is saved
12385 with the database itself in a separate index attribute file,
12386 named like the index file with '.attr' appended to the name.
12390 * X509 verify fixes. Disable broken certificate workarounds when
12391 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12392 keyUsage extension present. Don't accept CRLs with unhandled critical
12393 extensions: since verify currently doesn't process CRL extensions this
12394 rejects a CRL with *any* critical extensions. Add new verify error codes
12399 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12400 A clarification of RFC2560 will require the use of OCTET STRINGs and
12401 some implementations cannot handle the current raw format. Since OpenSSL
12402 copies and compares OCSP nonces as opaque blobs without any attempt at
12403 parsing them this should not create any compatibility issues.
12407 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12408 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12409 this HMAC (and other) operations are several times slower than OpenSSL
12414 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12416 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12418 * Use the correct content when signing type "other".
12422 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12424 * Fix various bugs revealed by running the NISCC test suite:
12426 Stop out of bounds reads in the ASN1 code when presented with
12427 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12429 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12431 If verify callback ignores invalid public key errors don't try to check
12432 certificate signature with the NULL public key.
12436 * New -ignore_err option in ocsp application to stop the server
12437 exiting on the first error in a request.
12441 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12442 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12447 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12448 extra data after the compression methods not only for TLS 1.0
12449 but also for SSL 3.0 (as required by the specification).
12451 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12453 * Change X509_certificate_type() to mark the key as exported/exportable
12454 when it's 512 *bits* long, not 512 bytes.
12458 * Change AES_cbc_encrypt() so it outputs exact multiple of
12459 blocks during encryption.
12463 * Various fixes to base64 BIO and non blocking I/O. On write
12464 flushes were not handled properly if the BIO retried. On read
12465 data was not being buffered properly and had various logic bugs.
12466 This also affects blocking I/O when the data being decoded is a
12471 * Various S/MIME bugfixes and compatibility changes:
12472 output correct application/pkcs7 MIME type if
12473 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12474 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12475 of files as .eml work). Correctly handle very long lines in MIME
12480 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12482 * Countermeasure against the Klima-Pokorny-Rosa extension of
12483 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12484 a protocol version number mismatch like a decryption error
12485 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12489 * Turn on RSA blinding by default in the default implementation
12490 to avoid a timing attack. Applications that don't want it can call
12491 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12492 They would be ill-advised to do so in most cases.
12494 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12496 * Change RSA blinding code so that it works when the PRNG is not
12497 seeded (in this case, the secret RSA exponent is abused as
12498 an unpredictable seed -- if it is not unpredictable, there
12499 is no point in blinding anyway). Make RSA blinding thread-safe
12500 by remembering the creator's thread ID in rsa->blinding and
12501 having all other threads use local one-time blinding factors
12502 (this requires more computation than sharing rsa->blinding, but
12503 avoids excessive locking; and if an RSA object is not shared
12504 between threads, blinding will still be very fast).
12508 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12509 ENGINE as defaults for all supported algorithms irrespective of
12510 the 'flags' parameter. 'flags' is now honoured, so applications
12511 should make sure they are passing it correctly.
12515 * Target "mingw" now allows native Windows code to be generated in
12516 the Cygwin environment as well as with the MinGW compiler.
12520 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12522 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12523 via timing by performing a MAC computation even if incorrect
12524 block cipher padding has been found. This is a countermeasure
12525 against active attacks where the attacker has to distinguish
12526 between bad padding and a MAC verification error. ([CVE-2003-0078])
12528 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12529 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12530 Martin Vuagnoux (EPFL, Ilion)*
12532 * Make the no-err option work as intended. The intention with no-err
12533 is not to have the whole error stack handling routines removed from
12534 libcrypto, it's only intended to remove all the function name and
12535 reason texts, thereby removing some of the footprint that may not
12536 be interesting if those errors aren't displayed anyway.
12538 NOTE: it's still possible for any application or module to have its
12539 own set of error texts inserted. The routines are there, just not
12540 used by default when no-err is given.
12544 * Add support for FreeBSD on IA64.
12546 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12548 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12549 Kerberos function mit_des_cbc_cksum(). Before this change,
12550 the value returned by DES_cbc_cksum() was like the one from
12551 mit_des_cbc_cksum(), except the bytes were swapped.
12553 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12555 * Allow an application to disable the automatic SSL chain building.
12556 Before this a rather primitive chain build was always performed in
12557 ssl3_output_cert_chain(): an application had no way to send the
12558 correct chain if the automatic operation produced an incorrect result.
12560 Now the chain builder is disabled if either:
12562 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12564 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12566 The reasoning behind this is that an application would not want the
12567 auto chain building to take place if extra chain certificates are
12568 present and it might also want a means of sending no additional
12569 certificates (for example the chain has two certificates and the
12574 * Add the possibility to build without the ENGINE framework.
12576 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12578 * Under Win32 gmtime() can return NULL: check return value in
12579 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12583 * DSA routines: under certain error conditions uninitialized BN objects
12584 could be freed. Solution: make sure initialization is performed early
12585 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12586 Nils Larsch <nla@trustcenter.de> via PR#459)
12590 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12591 checked on reconnect on the client side, therefore session resumption
12592 could still fail with a "ssl session id is different" error. This
12593 behaviour is masked when SSL_OP_ALL is used due to
12594 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12595 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12596 followup to PR #377.
12600 * IA-32 assembler support enhancements: unified ELF targets, support
12601 for SCO/Caldera platforms, fix for Cygwin shared build.
12605 * Add support for FreeBSD on sparc64. As a consequence, support for
12606 FreeBSD on non-x86 processors is separate from x86 processors on
12607 the config script, much like the NetBSD support.
12609 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12611 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12613 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12616 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12617 code (06) was taken as the first octet of the session ID and the last
12618 octet was ignored consequently. As a result SSLv2 client side session
12619 caching could not have worked due to the session ID mismatch between
12621 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12626 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12627 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12632 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12633 seems that in spite of existing for more than a year, many application
12634 author have done nothing to provide the necessary callbacks, which
12635 means that this particular engine will not work properly anywhere.
12636 This is a very unfortunate situation which forces us, in the name
12637 of usability, to give the hw_ncipher.c a static lock, which is part
12639 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12640 appear in 0.9.8 or later. We EXPECT application authors to have
12641 dealt properly with this when 0.9.8 is released (unless we actually
12642 make such changes in the libcrypto locking code that changes will
12643 have to be made anyway).
12647 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12648 octets have been read, EOF or an error occurs. Without this change
12649 some truncated ASN1 structures will not produce an error.
12653 * Disable Heimdal support, since it hasn't been fully implemented.
12654 Still give the possibility to force the use of Heimdal, but with
12655 warnings and a request that patches get sent to openssl-dev.
12659 * Add the VC-CE target, introduce the WINCE sysname, and add
12660 INSTALL.WCE and appropriate conditionals to make it build.
12662 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12664 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12665 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12666 edit numbers of the version.
12668 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12670 * Introduce safe string copy and catenation functions
12671 (BUF_strlcpy() and BUF_strlcat()).
12673 *Ben Laurie (CHATS) and Richard Levitte*
12675 * Avoid using fixed-size buffers for one-line DNs.
12677 *Ben Laurie (CHATS)*
12679 * Add BUF_MEM_grow_clean() to avoid information leakage when
12680 resizing buffers containing secrets, and use where appropriate.
12682 *Ben Laurie (CHATS)*
12684 * Avoid using fixed size buffers for configuration file location.
12686 *Ben Laurie (CHATS)*
12688 * Avoid filename truncation for various CA files.
12690 *Ben Laurie (CHATS)*
12692 * Use sizeof in preference to magic numbers.
12694 *Ben Laurie (CHATS)*
12696 * Avoid filename truncation in cert requests.
12698 *Ben Laurie (CHATS)*
12700 * Add assertions to check for (supposedly impossible) buffer
12703 *Ben Laurie (CHATS)*
12705 * Don't cache truncated DNS entries in the local cache (this could
12706 potentially lead to a spoofing attack).
12708 *Ben Laurie (CHATS)*
12710 * Fix various buffers to be large enough for hex/decimal
12711 representations in a platform independent manner.
12713 *Ben Laurie (CHATS)*
12715 * Add CRYPTO_realloc_clean() to avoid information leakage when
12716 resizing buffers containing secrets, and use where appropriate.
12718 *Ben Laurie (CHATS)*
12720 * Add BIO_indent() to avoid much slightly worrying code to do
12723 *Ben Laurie (CHATS)*
12725 * Convert sprintf()/BIO_puts() to BIO_printf().
12727 *Ben Laurie (CHATS)*
12729 * buffer_gets() could terminate with the buffer only half
12732 *Ben Laurie (CHATS)*
12734 * Add assertions to prevent user-supplied crypto functions from
12735 overflowing internal buffers by having large block sizes, etc.
12737 *Ben Laurie (CHATS)*
12739 * New OPENSSL_assert() macro (similar to assert(), but enabled
12742 *Ben Laurie (CHATS)*
12744 * Eliminate unused copy of key in RC4.
12746 *Ben Laurie (CHATS)*
12748 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12750 *Ben Laurie (CHATS)*
12752 * Fix off-by-one error in EGD path.
12754 *Ben Laurie (CHATS)*
12756 * If RANDFILE path is too long, ignore instead of truncating.
12758 *Ben Laurie (CHATS)*
12760 * Eliminate unused and incorrectly sized X.509 structure
12763 *Ben Laurie (CHATS)*
12765 * Eliminate unused and dangerous function knumber().
12767 *Ben Laurie (CHATS)*
12769 * Eliminate unused and dangerous structure, KSSL_ERR.
12771 *Ben Laurie (CHATS)*
12773 * Protect against overlong session ID context length in an encoded
12774 session object. Since these are local, this does not appear to be
12777 *Ben Laurie (CHATS)*
12779 * Change from security patch (see 0.9.6e below) that did not affect
12780 the 0.9.6 release series:
12782 Remote buffer overflow in SSL3 protocol - an attacker could
12783 supply an oversized master key in Kerberos-enabled versions.
12786 *Ben Laurie (CHATS)*
12788 * Change the SSL kerb5 codes to match RFC 2712.
12792 * Make -nameopt work fully for req and add -reqopt switch.
12794 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12796 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12798 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12800 * Make sure tests can be performed even if the corresponding algorithms
12801 have been removed entirely. This was also the last step to make
12802 OpenSSL compilable with DJGPP under all reasonable conditions.
12804 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12806 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12807 to allow version independent disabling of normally unselected ciphers,
12808 which may be activated as a side-effect of selecting a single cipher.
12810 (E.g., cipher list string "RSA" enables ciphersuites that are left
12811 out of "ALL" because they do not provide symmetric encryption.
12812 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12814 *Lutz Jaenicke, Bodo Moeller*
12816 * Add appropriate support for separate platform-dependent build
12817 directories. The recommended way to make a platform-dependent
12818 build directory is the following (tested on Linux), maybe with
12821 # Place yourself outside of the OpenSSL source tree. In
12822 # this example, the environment variable OPENSSL_SOURCE
12823 # is assumed to contain the absolute OpenSSL source directory.
12824 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12825 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12826 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12827 mkdir -p `dirname $F`
12828 ln -s $OPENSSL_SOURCE/$F $F
12831 To be absolutely sure not to disturb the source tree, a "make clean"
12832 is a good thing. If it isn't successful, don't worry about it,
12833 it probably means the source directory is very clean.
12837 * Make sure any ENGINE control commands make local copies of string
12838 pointers passed to them whenever necessary. Otherwise it is possible
12839 the caller may have overwritten (or deallocated) the original string
12840 data when a later ENGINE operation tries to use the stored values.
12842 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12844 * Improve diagnostics in file reading and command-line digests.
12846 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12848 * Add AES modes CFB and OFB to the object database. Correct an
12849 error in AES-CFB decryption.
12853 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12854 allows existing EVP_CIPHER_CTX structures to be reused after
12855 calling `EVP_*Final()`. This behaviour is used by encryption
12856 BIOs and some applications. This has the side effect that
12857 applications must explicitly clean up cipher contexts with
12858 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12862 * Check the values of dna and dnb in bn_mul_recursive before calling
12863 bn_mul_comba (a non zero value means the a or b arrays do not contain
12864 n2 elements) and fallback to bn_mul_normal if either is not zero.
12868 * Fix escaping of non-ASCII characters when using the -subj option
12869 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12873 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12874 form for "surname", serialNumber has no short form.
12875 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12876 therefore remove "mail" short name for "internet 7".
12877 The OID for unique identifiers in X509 certificates is
12878 x500UniqueIdentifier, not uniqueIdentifier.
12879 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12883 * Add an "init" command to the ENGINE config module and auto initialize
12884 ENGINEs. Without any "init" command the ENGINE will be initialized
12885 after all ctrl commands have been executed on it. If init=1 the
12886 ENGINE is initialized at that point (ctrls before that point are run
12887 on the uninitialized ENGINE and after on the initialized one). If
12888 init=0 then the ENGINE will not be initialized at all.
12892 * Fix the 'app_verify_callback' interface so that the user-defined
12893 argument is actually passed to the callback: In the
12894 SSL_CTX_set_cert_verify_callback() prototype, the callback
12895 declaration has been changed from
12898 int (*cb)(X509_STORE_CTX *,void *);
12899 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12900 i=s->ctx->app_verify_callback(&ctx)
12901 has been changed into
12902 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12904 To update applications using SSL_CTX_set_cert_verify_callback(),
12905 a dummy argument can be added to their callback functions.
12907 *D. K. Smetters <smetters@parc.xerox.com>*
12909 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12911 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12913 * Add and OPENSSL_LOAD_CONF define which will cause
12914 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12915 This allows older applications to transparently support certain
12916 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12917 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12918 load the config file and OPENSSL_add_all_algorithms_conf() which will
12919 always load it have also been added.
12923 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12924 Adjust NIDs and EVP layer.
12926 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12928 * Config modules support in openssl utility.
12930 Most commands now load modules from the config file,
12931 though in a few (such as version) this isn't done
12932 because it couldn't be used for anything.
12934 In the case of ca and req the config file used is
12935 the same as the utility itself: that is the -config
12936 command line option can be used to specify an
12941 * Move default behaviour from OPENSSL_config(). If appname is NULL
12942 use "openssl_conf" if filename is NULL use default openssl config file.
12946 * Add an argument to OPENSSL_config() to allow the use of an alternative
12947 config section name. Add a new flag to tolerate a missing config file
12948 and move code to CONF_modules_load_file().
12952 * Support for crypto accelerator cards from Accelerated Encryption
12953 Processing, www.aep.ie. (Use engine 'aep')
12954 The support was copied from 0.9.6c [engine] and adapted/corrected
12955 to work with the new engine framework.
12957 *AEP Inc. and Richard Levitte*
12959 * Support for SureWare crypto accelerator cards from Baltimore
12960 Technologies. (Use engine 'sureware')
12961 The support was copied from 0.9.6c [engine] and adapted
12962 to work with the new engine framework.
12966 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12967 make the newer ENGINE framework commands for the CHIL engine work.
12969 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12971 * Make it possible to produce shared libraries on ReliantUNIX.
12973 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12975 * Add the configuration target debug-linux-ppro.
12976 Make 'openssl rsa' use the general key loading routines
12977 implemented in `apps.c`, and make those routines able to
12978 handle the key format FORMAT_NETSCAPE and the variant
12981 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12983 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12985 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12987 * Add -keyform to rsautl, and document -engine.
12989 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12991 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12992 BIO_R_NO_SUCH_FILE error code rather than the generic
12993 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12997 * Add new functions
12998 ERR_peek_last_error
12999 ERR_peek_last_error_line
13000 ERR_peek_last_error_line_data.
13001 These are similar to
13003 ERR_peek_error_line
13004 ERR_peek_error_line_data,
13005 but report on the latest error recorded rather than the first one
13006 still in the error queue.
13008 *Ben Laurie, Bodo Moeller*
13010 * default_algorithms option in ENGINE config module. This allows things
13012 default_algorithms = ALL
13013 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
13017 * Preliminary ENGINE config module.
13021 * New experimental application configuration code.
13025 * Change the AES code to follow the same name structure as all other
13026 symmetric ciphers, and behave the same way. Move everything to
13027 the directory crypto/aes, thereby obsoleting crypto/rijndael.
13029 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13031 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
13033 *Ben Laurie and Theo de Raadt*
13035 * Add option to output public keys in req command.
13037 *Massimiliano Pala madwolf@openca.org*
13039 * Use wNAFs in EC_POINTs_mul() for improved efficiency
13040 (up to about 10% better than before for P-192 and P-224).
13044 * New functions/macros
13046 SSL_CTX_set_msg_callback(ctx, cb)
13047 SSL_CTX_set_msg_callback_arg(ctx, arg)
13048 SSL_set_msg_callback(ssl, cb)
13049 SSL_set_msg_callback_arg(ssl, arg)
13051 to request calling a callback function
13053 void cb(int write_p, int version, int content_type,
13054 const void *buf, size_t len, SSL *ssl, void *arg)
13056 whenever a protocol message has been completely received
13057 (write_p == 0) or sent (write_p == 1). Here 'version' is the
13058 protocol version according to which the SSL library interprets
13059 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
13060 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
13061 the content type as defined in the SSL 3.0/TLS 1.0 protocol
13062 specification (change_cipher_spec(20), alert(21), handshake(22)).
13063 'buf' and 'len' point to the actual message, 'ssl' to the
13064 SSL object, and 'arg' is the application-defined value set by
13065 SSL[_CTX]_set_msg_callback_arg().
13067 'openssl s_client' and 'openssl s_server' have new '-msg' options
13068 to enable a callback that displays all protocol messages.
13072 * Change the shared library support so shared libraries are built as
13073 soon as the corresponding static library is finished, and thereby get
13074 openssl and the test programs linked against the shared library.
13075 This still only happens when the keyword "shard" has been given to
13076 the configuration scripts.
13078 NOTE: shared library support is still an experimental thing, and
13079 backward binary compatibility is still not guaranteed.
13081 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13083 * Add support for Subject Information Access extension.
13085 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13087 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13088 additional bytes when new memory had to be allocated, not just
13089 when reusing an existing buffer.
13093 * New command line and configuration option 'utf8' for the req command.
13094 This allows field values to be specified as UTF8 strings.
13098 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13099 runs for the former and machine-readable output for the latter.
13103 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13104 of the e-mail address in the DN (i.e., it will go into a certificate
13105 extension only). The new configuration file option 'email_in_dn = no'
13106 has the same effect.
13108 *Massimiliano Pala madwolf@openca.org*
13110 * Change all functions with names starting with `des_` to be starting
13111 with `DES_` instead. Add wrappers that are compatible with libdes,
13112 but are named `_ossl_old_des_*`. Finally, add macros that map the
13113 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13114 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13115 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13118 Since we provide two compatibility mappings, the user needs to
13119 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13120 compatibility is desired. The default (i.e., when that macro
13121 isn't defined) is OpenSSL 0.9.6c compatibility.
13123 There are also macros that enable and disable the support of old
13124 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13125 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13126 are defined, the default will apply: to support the old des routines.
13128 In either case, one must include openssl/des.h to get the correct
13129 definitions. Do not try to just include openssl/des_old.h, that
13132 NOTE: This is a major break of an old API into a new one. Software
13133 authors are encouraged to switch to the `DES_` style functions. Some
13134 time in the future, des_old.h and the libdes compatibility functions
13135 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13136 default), and then completely removed.
13140 * Test for certificates which contain unsupported critical extensions.
13141 If such a certificate is found during a verify operation it is
13142 rejected by default: this behaviour can be overridden by either
13143 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13144 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13145 X509_supported_extension() has also been added which returns 1 if a
13146 particular extension is supported.
13150 * Modify the behaviour of EVP cipher functions in similar way to digests
13151 to retain compatibility with existing code.
13155 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13156 compatibility with existing code. In particular the 'ctx' parameter does
13157 not have to be to be initialized before the call to EVP_DigestInit() and
13158 it is tidied up after a call to EVP_DigestFinal(). New function
13159 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13160 EVP_MD_CTX_copy() changed to not require the destination to be
13161 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13162 requires the destination to be valid.
13164 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13165 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13169 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13170 so that complete 'Handshake' protocol structures are kept in memory
13171 instead of overwriting 'msg_type' and 'length' with 'body' data.
13175 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13177 *Massimo Santin via Richard Levitte*
13179 * Major restructuring to the underlying ENGINE code. This includes
13180 reduction of linker bloat, separation of pure "ENGINE" manipulation
13181 (initialisation, etc) from functionality dealing with implementations
13182 of specific crypto interfaces. This change also introduces integrated
13183 support for symmetric ciphers and digest implementations - so ENGINEs
13184 can now accelerate these by providing EVP_CIPHER and EVP_MD
13185 implementations of their own. This is detailed in
13186 [crypto/engine/README.md](crypto/engine/README.md)
13187 as it couldn't be adequately described here. However, there are a few
13188 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13189 were changed in the original introduction of ENGINE code have now
13190 reverted back - the hooking from this code to ENGINE is now a good
13191 deal more passive and at run-time, operations deal directly with
13192 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13193 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13194 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13195 they were not being used by the framework as there is no concept of a
13196 BIGNUM_METHOD and they could not be generalised to the new
13197 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13198 ENGINE_cpy() has been removed as it cannot be consistently defined in
13203 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13207 * Change mkdef.pl to sort symbols that get the same entry number,
13208 and make sure the automatically generated functions `ERR_load_*`
13209 become part of libeay.num as well.
13213 * New function SSL_renegotiate_pending(). This returns true once
13214 renegotiation has been requested (either SSL_renegotiate() call
13215 or HelloRequest/ClientHello received from the peer) and becomes
13216 false once a handshake has been completed.
13217 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13218 sends a HelloRequest, but does not ensure that a handshake takes
13219 place. SSL_renegotiate_pending() is useful for checking if the
13220 client has followed the request.)
13224 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13225 By default, clients may request session resumption even during
13226 renegotiation (if session ID contexts permit); with this option,
13227 session resumption is possible only in the first handshake.
13229 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13230 more bits available for options that should not be part of
13231 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13235 * Add some demos for certificate and certificate request creation.
13239 * Make maximum certificate chain size accepted from the peer application
13240 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13241 "Douglas E. Engert" <deengert@anl.gov>.
13245 * Add support for shared libraries for Unixware-7
13246 (Boyd Lynn Gerber <gerberb@zenez.com>).
13250 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13251 be done prior to destruction. Use this to unload error strings from
13252 ENGINEs that load their own error strings. NB: This adds two new API
13253 functions to "get" and "set" this destroy handler in an ENGINE.
13257 * Alter all existing ENGINE implementations (except "openssl" and
13258 "openbsd") to dynamically instantiate their own error strings. This
13259 makes them more flexible to be built both as statically-linked ENGINEs
13260 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13261 Also, add stub code to each that makes building them as self-contained
13262 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13266 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13267 implementations into applications that are completely implemented in
13268 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13269 commands that can be used to configure what shared-library to load and
13270 to control aspects of the way it is handled. Also, made an update to
13271 the [README-Engine.md](README-Engine.md) file
13272 that brings its information up-to-date and
13273 provides some information and instructions on the "dynamic" ENGINE
13274 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13278 * Make it possible to unload ranges of ERR strings with a new
13279 "ERR_unload_strings" function.
13283 * Add a copy() function to EVP_MD.
13287 * Make EVP_MD routines take a context pointer instead of just the
13288 md_data void pointer.
13292 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13293 that the digest can only process a single chunk of data
13294 (typically because it is provided by a piece of
13295 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13296 is only going to provide a single chunk of data, and hence the
13297 framework needn't accumulate the data for oneshot drivers.
13301 * As with "ERR", make it possible to replace the underlying "ex_data"
13302 functions. This change also alters the storage and management of global
13303 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13304 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13305 index counters. The API functions that use this state have been changed
13306 to take a "class_index" rather than pointers to the class's local STACK
13307 and counter, and there is now an API function to dynamically create new
13308 classes. This centralisation allows us to (a) plug a lot of the
13309 thread-safety problems that existed, and (b) makes it possible to clean
13310 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13311 such data would previously have always leaked in application code and
13312 workarounds were in place to make the memory debugging turn a blind eye
13313 to it. Application code that doesn't use this new function will still
13314 leak as before, but their memory debugging output will announce it now
13315 rather than letting it slide.
13317 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13318 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13319 has a return value to indicate success or failure.
13323 * Make it possible to replace the underlying "ERR" functions such that the
13324 global state (2 LHASH tables and 2 locks) is only used by the "default"
13325 implementation. This change also adds two functions to "get" and "set"
13326 the implementation prior to it being automatically set the first time
13327 any other ERR function takes place. Ie. an application can call "get",
13328 pass the return value to a module it has just loaded, and that module
13329 can call its own "set" function using that value. This means the
13330 module's "ERR" operations will use (and modify) the error state in the
13331 application and not in its own statically linked copy of OpenSSL code.
13335 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13336 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13337 the operation, and provides a more encapsulated way for external code
13338 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13339 to use these functions rather than manually incrementing the counts.
13341 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13345 * Add EVP test program.
13349 * Add symmetric cipher support to ENGINE. Expect the API to change!
13353 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13354 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13355 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13356 These allow a CRL to be built without having to access X509_CRL fields
13357 directly. Modify 'ca' application to use new functions.
13361 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13362 bug workarounds. Rollback attack detection is a security feature.
13363 The problem will only arise on OpenSSL servers when TLSv1 is not
13364 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13365 Software authors not wanting to support TLSv1 will have special reasons
13366 for their choice and can explicitly enable this option.
13368 *Bodo Moeller, Lutz Jaenicke*
13370 * Rationalise EVP so it can be extended: don't include a union of
13371 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13372 (similar to those existing for EVP_CIPHER_CTX).
13377 EVP_MD_CTX_init(&md); /* new function call */
13378 EVP_DigestInit(&md, EVP_sha1());
13379 EVP_DigestUpdate(&md, in, len);
13380 EVP_DigestFinal(&md, out, NULL);
13381 EVP_MD_CTX_cleanup(&md); /* new function call */
13385 * Make DES key schedule conform to the usual scheme, as well as
13386 correcting its structure. This means that calls to DES functions
13387 now have to pass a pointer to a des_key_schedule instead of a
13388 plain des_key_schedule (which was actually always a pointer
13391 des_key_schedule ks;
13393 des_set_key_checked(..., &ks);
13394 des_ncbc_encrypt(..., &ks, ...);
13396 (Note that a later change renames 'des_...' into 'DES_...'.)
13400 * Initial reduction of linker bloat: the use of some functions, such as
13401 PEM causes large amounts of unused functions to be linked in due to
13402 poor organisation. For example pem_all.c contains every PEM function
13403 which has a knock on effect of linking in large amounts of (unused)
13404 ASN1 code. Grouping together similar functions and splitting unrelated
13405 functions prevents this.
13409 * Cleanup of EVP macros.
13413 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13414 correct `_ecb suffix`.
13418 * Add initial OCSP responder support to ocsp application. The
13419 revocation information is handled using the text based index
13420 use by the ca application. The responder can either handle
13421 requests generated internally, supplied in files (for example
13422 via a CGI script) or using an internal minimal server.
13426 * Add configuration choices to get zlib compression for TLS.
13430 * Changes to Kerberos SSL for RFC 2712 compliance:
13431 1. Implemented real KerberosWrapper, instead of just using
13432 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13433 2. Implemented optional authenticator field of KerberosWrapper.
13435 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13436 and authenticator structs; see crypto/krb5/.
13438 Generalized Kerberos calls to support multiple Kerberos libraries.
13439 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13440 via Richard Levitte*
13442 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13443 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13444 values for each of the key sizes rather than having just
13445 parameters (and 'speed' generating keys each time).
13449 * Speed up EVP routines.
13452 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13453 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13454 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13455 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13457 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13458 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13459 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13462 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13464 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13468 * Added the OS2-EMX target.
13470 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13472 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13473 New functions to support `NCONF` routines in extension code.
13474 New function `CONF_set_nconf()`
13475 to allow functions which take an `NCONF` to also handle the old `LHASH`
13476 structure: this means that the old `CONF` compatible routines can be
13477 retained (in particular w.rt. extensions) without having to duplicate the
13478 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13482 * Enhance the general user interface with mechanisms for inner control
13483 and with possibilities to have yes/no kind of prompts.
13487 * Change all calls to low-level digest routines in the library and
13488 applications to use EVP. Add missing calls to HMAC_cleanup() and
13489 don't assume HMAC_CTX can be copied using memcpy().
13491 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13493 * Add the possibility to control engines through control names but with
13494 arbitrary arguments instead of just a string.
13495 Change the key loaders to take a UI_METHOD instead of a callback
13496 function pointer. NOTE: this breaks binary compatibility with earlier
13497 versions of OpenSSL [engine].
13498 Adapt the nCipher code for these new conditions and add a card insertion
13503 * Enhance the general user interface with mechanisms to better support
13504 dialog box interfaces, application-defined prompts, the possibility
13505 to use defaults (for example default passwords from somewhere else)
13506 and interrupts/cancellations.
13510 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13511 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13515 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13516 tidy up some unnecessarily weird code in 'sk_new()').
13518 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13520 * Change the key loading routines for ENGINEs to use the same kind
13521 callback (pem_password_cb) as all other routines that need this
13526 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13527 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13528 than this minimum value is recommended.
13532 * New random seeder for OpenVMS, using the system process statistics
13533 that are easily reachable.
13537 * Windows apparently can't transparently handle global
13538 variables defined in DLLs. Initialisations such as:
13540 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13542 won't compile. This is used by the any applications that need to
13543 declare their own ASN1 modules. This was fixed by adding the option
13544 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13545 needed for static libraries under Win32.
13549 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13550 setting of purpose and trust fields. New X509_STORE trust and
13551 purpose functions and tidy up setting in other SSL functions.
13555 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13556 structure. These are inherited by X509_STORE_CTX when it is
13557 initialised. This allows various defaults to be set in the
13558 X509_STORE structure (such as flags for CRL checking and custom
13559 purpose or trust settings) for functions which only use X509_STORE_CTX
13560 internally such as S/MIME.
13562 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13563 trust settings if they are not set in X509_STORE. This allows X509_STORE
13564 purposes and trust (in S/MIME for example) to override any set by default.
13566 Add command line options for CRL checking to smime, s_client and s_server
13571 * Initial CRL based revocation checking. If the CRL checking flag(s)
13572 are set then the CRL is looked up in the X509_STORE structure and
13573 its validity and signature checked, then if the certificate is found
13574 in the CRL the verify fails with a revoked error.
13576 Various new CRL related callbacks added to X509_STORE_CTX structure.
13578 Command line options added to 'verify' application to support this.
13580 This needs some additional work, such as being able to handle multiple
13581 CRLs with different times, extension based lookup (rather than just
13582 by subject name) and ultimately more complete V2 CRL extension
13587 * Add a general user interface API (crypto/ui/). This is designed
13588 to replace things like des_read_password and friends (backward
13589 compatibility functions using this new API are provided).
13590 The purpose is to remove prompting functions from the DES code
13591 section as well as provide for prompting through dialog boxes in
13592 a window system and the like.
13596 * Add "ex_data" support to ENGINE so implementations can add state at a
13597 per-structure level rather than having to store it globally.
13601 * Make it possible for ENGINE structures to be copied when retrieved by
13602 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13603 This causes the "original" ENGINE structure to act like a template,
13604 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13605 operational state can be localised to each ENGINE structure, despite the
13606 fact they all share the same "methods". New ENGINE structures returned in
13607 this case have no functional references and the return value is the single
13608 structural reference. This matches the single structural reference returned
13609 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13614 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13615 needs to match any other type at all we need to manually clear the
13620 * Changes to the "openssl engine" utility to include;
13621 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13622 about an ENGINE's available control commands.
13623 - executing control commands from command line arguments using the
13624 '-pre' and '-post' switches. '-post' is only used if '-t' is
13625 specified and the ENGINE is successfully initialised. The syntax for
13626 the individual commands are colon-separated, for example;
13627 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13631 * New dynamic control command support for ENGINEs. ENGINEs can now
13632 declare their own commands (numbers), names (strings), descriptions,
13633 and input types for run-time discovery by calling applications. A
13634 subset of these commands are implicitly classed as "executable"
13635 depending on their input type, and only these can be invoked through
13636 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13637 can be based on user input, config files, etc). The distinction is
13638 that "executable" commands cannot return anything other than a boolean
13639 result and can only support numeric or string input, whereas some
13640 discoverable commands may only be for direct use through
13641 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13642 pointers, or other custom uses. The "executable" commands are to
13643 support parameterisations of ENGINE behaviour that can be
13644 unambiguously defined by ENGINEs and used consistently across any
13645 OpenSSL-based application. Commands have been added to all the
13646 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13647 control over shared-library paths without source code alterations.
13651 * Changed all ENGINE implementations to dynamically allocate their
13652 ENGINEs rather than declaring them statically. Apart from this being
13653 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13654 this also allows the implementations to compile without using the
13655 internal engine_int.h header.
13659 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13660 'const' value. Any code that should be able to modify a RAND_METHOD
13661 should already have non-const pointers to it (ie. they should only
13662 modify their own ones).
13666 * Made a variety of little tweaks to the ENGINE code.
13667 - "atalla" and "ubsec" string definitions were moved from header files
13668 to C code. "nuron" string definitions were placed in variables
13669 rather than hard-coded - allowing parameterisation of these values
13670 later on via ctrl() commands.
13671 - Removed unused "#if 0"'d code.
13672 - Fixed engine list iteration code so it uses ENGINE_free() to release
13673 structural references.
13674 - Constified the RAND_METHOD element of ENGINE structures.
13675 - Constified various get/set functions as appropriate and added
13676 missing functions (including a catch-all ENGINE_cpy that duplicates
13677 all ENGINE values onto a new ENGINE except reference counts/state).
13678 - Removed NULL parameter checks in get/set functions. Setting a method
13679 or function to NULL is a way of cancelling out a previously set
13680 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13681 and doesn't justify the extra error symbols and code.
13682 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13683 flags from engine_int.h to engine.h.
13684 - Changed prototypes for ENGINE handler functions (init(), finish(),
13685 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13689 * Implement binary inversion algorithm for BN_mod_inverse in addition
13690 to the algorithm using long division. The binary algorithm can be
13691 used only if the modulus is odd. On 32-bit systems, it is faster
13692 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13693 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13694 up to 450 bits. In 64-bit environments, the binary algorithm
13695 appears to be advantageous for much longer moduli; here we use it
13696 for moduli up to 2048 bits.
13700 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13701 could not support the combine flag in choice fields.
13705 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13706 extensions from a certificate request to the certificate.
13710 * Allow multiple 'certopt' and 'nameopt' options to be separated
13711 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13712 file: this allows the display of the certificate about to be
13713 signed to be customised, to allow certain fields to be included
13714 or excluded and extension details. The old system didn't display
13715 multicharacter strings properly, omitted fields not in the policy
13716 and couldn't display additional details such as extensions.
13720 * Function EC_POINTs_mul for multiple scalar multiplication
13721 of an arbitrary number of elliptic curve points
13722 \sum scalars[i]*points[i],
13723 optionally including the generator defined for the EC_GROUP:
13724 scalar*generator + \sum scalars[i]*points[i].
13726 EC_POINT_mul is a simple wrapper function for the typical case
13727 that the point list has just one item (besides the optional
13732 * First EC_METHODs for curves over GF(p):
13734 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13735 operations and provides various method functions that can also
13736 operate with faster implementations of modular arithmetic.
13738 EC_GFp_mont_method() reuses most functions that are part of
13739 EC_GFp_simple_method, but uses Montgomery arithmetic.
13741 *Bodo Moeller; point addition and point doubling
13742 implementation directly derived from source code provided by
13743 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13745 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13746 crypto/ec/ec_lib.c):
13748 Curves are EC_GROUP objects (with an optional group generator)
13749 based on EC_METHODs that are built into the library.
13751 Points are EC_POINT objects based on EC_GROUP objects.
13753 Most of the framework would be able to handle curves over arbitrary
13754 finite fields, but as there are no obvious types for fields other
13755 than GF(p), some functions are limited to that for now.
13759 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13760 that the file contains a complete HTTP response.
13764 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13765 change the def and num file printf format specifier from "%-40sXXX"
13766 to "%-39s XXX". The latter will always guarantee a space after the
13767 field while the former will cause them to run together if the field
13768 is 40 of more characters long.
13772 * Constify the cipher and digest 'method' functions and structures
13773 and modify related functions to take constant EVP_MD and EVP_CIPHER
13778 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13779 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13783 * Modify `EVP_Digest*()` routines so they now return values. Although the
13784 internal software routines can never fail additional hardware versions
13789 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13791 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13792 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13795 ERR_R_NESTED_ASN1_ERROR
13797 ERR_R_MISSING_ASN1_EOS
13798 were 4 .. 9, conflicting with
13799 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13801 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13802 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13804 Add new error code 'ERR_R_INTERNAL_ERROR'.
13808 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13813 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13814 sets the subject name for a new request or supersedes the
13815 subject name in a given request. Formats that can be parsed are
13816 'CN=Some Name, OU=myOU, C=IT'
13818 'CN=Some Name/OU=myOU/C=IT'.
13820 Add options '-batch' and '-verbose' to 'openssl req'.
13822 *Massimiliano Pala <madwolf@hackmasters.net>*
13824 * Introduce the possibility to access global variables through
13825 functions on platform were that's the best way to handle exporting
13826 global variables in shared libraries. To enable this functionality,
13827 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13828 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13829 is normally done by Configure or something similar).
13831 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13832 in the source file (foo.c) like this:
13834 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13835 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13837 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13838 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13840 OPENSSL_DECLARE_GLOBAL(int,foo);
13841 #define foo OPENSSL_GLOBAL_REF(foo)
13842 OPENSSL_DECLARE_GLOBAL(double,bar);
13843 #define bar OPENSSL_GLOBAL_REF(bar)
13845 The #defines are very important, and therefore so is including the
13846 header file everywhere where the defined globals are used.
13848 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13849 of ASN.1 items, but that structure is a bit different.
13851 The largest change is in util/mkdef.pl which has been enhanced with
13852 better and easier to understand logic to choose which symbols should
13853 go into the Windows .def files as well as a number of fixes and code
13854 cleanup (among others, algorithm keywords are now sorted
13855 lexicographically to avoid constant rewrites).
13859 * In BN_div() keep a copy of the sign of 'num' before writing the
13860 result to 'rm' because if rm==num the value will be overwritten
13861 and produce the wrong result if 'num' is negative: this caused
13862 problems with BN_mod() and BN_nnmod().
13866 * Function OCSP_request_verify(). This checks the signature on an
13867 OCSP request and verifies the signer certificate. The signer
13868 certificate is just checked for a generic purpose and OCSP request
13873 * Add OCSP_check_validity() function to check the validity of OCSP
13874 responses. OCSP responses are prepared in real time and may only
13875 be a few seconds old. Simply checking that the current time lies
13876 between thisUpdate and nextUpdate max reject otherwise valid responses
13877 caused by either OCSP responder or client clock inaccuracy. Instead
13878 we allow thisUpdate and nextUpdate to fall within a certain period of
13879 the current time. The age of the response can also optionally be
13880 checked. Two new options -validity_period and -status_age added to
13885 * If signature or public key algorithm is unrecognized print out its
13886 OID rather that just UNKNOWN.
13890 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13891 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13892 ID to be generated from the issuer certificate alone which can then be
13893 passed to OCSP_id_issuer_cmp().
13897 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13898 ASN1 modules to export functions returning ASN1_ITEM pointers
13899 instead of the ASN1_ITEM structures themselves. This adds several
13900 new macros which allow the underlying ASN1 function/structure to
13901 be accessed transparently. As a result code should not use ASN1_ITEM
13902 references directly (such as &X509_it) but instead use the relevant
13903 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13904 use of the new ASN1 code on platforms where exporting structures
13905 is problematical (for example in shared libraries) but exporting
13906 functions returning pointers to structures is not.
13910 * Add support for overriding the generation of SSL/TLS session IDs.
13911 These callbacks can be registered either in an SSL_CTX or per SSL.
13912 The purpose of this is to allow applications to control, if they wish,
13913 the arbitrary values chosen for use as session IDs, particularly as it
13914 can be useful for session caching in multiple-server environments. A
13915 command-line switch for testing this (and any client code that wishes
13916 to use such a feature) has been added to "s_server".
13918 *Geoff Thorpe, Lutz Jaenicke*
13920 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13921 of the form `#if defined(...) || defined(...) || ...` and
13922 `#if !defined(...) && !defined(...) && ...`. This also avoids
13923 the growing number of special cases it was previously handling.
13927 * Make all configuration macros available for application by making
13928 sure they are available in opensslconf.h, by giving them names starting
13929 with `OPENSSL_` to avoid conflicts with other packages and by making
13930 sure e_os2.h will cover all platform-specific cases together with
13932 Additionally, it is now possible to define configuration/platform-
13933 specific names (called "system identities"). In the C code, these
13934 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13935 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13936 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13941 * New option -set_serial to 'req' and 'x509' this allows the serial
13942 number to use to be specified on the command line. Previously self
13943 signed certificates were hard coded with serial number 0 and the
13944 CA options of 'x509' had to use a serial number in a file which was
13949 * New options to 'ca' utility to support V2 CRL entry extensions.
13950 Currently CRL reason, invalidity date and hold instruction are
13951 supported. Add new CRL extensions to V3 code and some new objects.
13955 * New function EVP_CIPHER_CTX_set_padding() this is used to
13956 disable standard block padding (aka PKCS#5 padding) in the EVP
13957 API, which was previously mandatory. This means that the data is
13958 not padded in any way and so the total length much be a multiple
13959 of the block size, otherwise an error occurs.
13963 * Initial (incomplete) OCSP SSL support.
13967 * New function OCSP_parse_url(). This splits up a URL into its host,
13968 port and path components: primarily to parse OCSP URLs. New -url
13969 option to ocsp utility.
13973 * New nonce behavior. The return value of OCSP_check_nonce() now
13974 reflects the various checks performed. Applications can decide
13975 whether to tolerate certain situations such as an absent nonce
13976 in a response when one was present in a request: the ocsp application
13977 just prints out a warning. New function OCSP_add1_basic_nonce()
13978 this is to allow responders to include a nonce in a response even if
13979 the request is nonce-less.
13983 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13984 skipped when using openssl x509 multiple times on a single input file,
13985 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13989 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13990 set string type: to handle setting ASN1_TIME structures. Fix ca
13991 utility to correctly initialize revocation date of CRLs.
13995 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13996 the clients preferred ciphersuites and rather use its own preferences.
13997 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13998 Internet Explorer by ensuring unchanged hash method during stepup.
13999 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
14003 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
14004 to aes and add a new 'exist' option to print out symbols that don't
14009 * Additional options to ocsp utility to allow flags to be set and
14010 additional certificates supplied.
14014 * Add the option -VAfile to 'openssl ocsp', so the user can give the
14015 OCSP client a number of certificate to only verify the response
14020 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
14021 handle the new API. Currently only ECB, CBC modes supported. Add new
14024 Add TLS AES ciphersuites as described in RFC3268, "Advanced
14025 Encryption Standard (AES) Ciphersuites for Transport Layer
14026 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
14027 not enabled by default and were not part of the "ALL" ciphersuite
14028 alias because they were not yet official; they could be
14029 explicitly requested by specifying the "AESdraft" ciphersuite
14030 group alias. In the final release of OpenSSL 0.9.7, the group
14031 alias is called "AES" and is part of "ALL".)
14033 *Ben Laurie, Steve Henson, Bodo Moeller*
14035 * New function OCSP_copy_nonce() to copy nonce value (if present) from
14036 request to response.
14040 * Functions for OCSP responders. OCSP_request_onereq_count(),
14041 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
14042 extract information from a certificate request. OCSP_response_create()
14043 creates a response and optionally adds a basic response structure.
14044 OCSP_basic_add1_status() adds a complete single response to a basic
14045 response and returns the OCSP_SINGLERESP structure just added (to allow
14046 extensions to be included for example). OCSP_basic_add1_cert() adds a
14047 certificate to a basic response and OCSP_basic_sign() signs a basic
14048 response with various flags. New helper functions ASN1_TIME_check()
14049 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
14050 (converts ASN1_TIME to GeneralizedTime).
14054 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
14055 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
14056 structure from a certificate. X509_pubkey_digest() digests the public_key
14057 contents: this is used in various key identifiers.
14061 * Make sk_sort() tolerate a NULL argument.
14063 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14065 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14066 passed by the function are trusted implicitly. If any of them signed the
14067 response then it is assumed to be valid and is not verified.
14071 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14072 to data. This was previously part of the PKCS7 ASN1 code. This
14073 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14074 *Steve Henson, reported by Kenneth R. Robinette
14075 <support@securenetterm.com>*
14077 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14078 routines: without these tracing memory leaks is very painful.
14079 Fix leaks in PKCS12 and PKCS7 routines.
14083 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14084 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14085 effectively meant GeneralizedTime would never be used. Now it
14086 is initialised to -1 but X509_time_adj() now has to check the value
14087 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14088 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14089 *Steve Henson, reported by Kenneth R. Robinette
14090 <support@securenetterm.com>*
14092 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14093 result in a zero length in the ASN1_INTEGER structure which was
14094 not consistent with the structure when d2i_ASN1_INTEGER() was used
14095 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14096 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14097 where it did not print out a minus for negative ASN1_INTEGER.
14101 * Add summary printout to ocsp utility. The various functions which
14102 convert status values to strings have been renamed to:
14103 OCSP_response_status_str(), OCSP_cert_status_str() and
14104 OCSP_crl_reason_str() and are no longer static. New options
14105 to verify nonce values and to disable verification. OCSP response
14106 printout format cleaned up.
14110 * Add additional OCSP certificate checks. These are those specified
14111 in RFC2560. This consists of two separate checks: the CA of the
14112 certificate being checked must either be the OCSP signer certificate
14113 or the issuer of the OCSP signer certificate. In the latter case the
14114 OCSP signer certificate must contain the OCSP signing extended key
14115 usage. This check is performed by attempting to match the OCSP
14116 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14117 in the OCSP_CERTID structures of the response.
14121 * Initial OCSP certificate verification added to OCSP_basic_verify()
14122 and related routines. This uses the standard OpenSSL certificate
14123 verify routines to perform initial checks (just CA validity) and
14124 to obtain the certificate chain. Then additional checks will be
14125 performed on the chain. Currently the root CA is checked to see
14126 if it is explicitly trusted for OCSP signing. This is used to set
14127 a root CA as a global signing root: that is any certificate that
14128 chains to that CA is an acceptable OCSP signing certificate.
14132 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14133 extensions from a separate configuration file.
14134 As when reading extensions from the main configuration file,
14135 the '-extensions ...' option may be used for specifying the
14138 *Massimiliano Pala <madwolf@comune.modena.it>*
14140 * New OCSP utility. Allows OCSP requests to be generated or
14141 read. The request can be sent to a responder and the output
14142 parsed, outputted or printed in text form. Not complete yet:
14143 still needs to check the OCSP response validity.
14147 * New subcommands for 'openssl ca':
14148 `openssl ca -status <serial>` prints the status of the cert with
14149 the given serial number (according to the index file).
14150 `openssl ca -updatedb` updates the expiry status of certificates
14153 *Massimiliano Pala <madwolf@comune.modena.it>*
14155 * New '-newreq-nodes' command option to CA.pl. This is like
14156 '-newreq', but calls 'openssl req' with the '-nodes' option
14157 so that the resulting key is not encrypted.
14159 *Damien Miller <djm@mindrot.org>*
14161 * New configuration for the GNU Hurd.
14163 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14165 * Initial code to implement OCSP basic response verify. This
14166 is currently incomplete. Currently just finds the signer's
14167 certificate and verifies the signature on the response.
14171 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14172 value of OPENSSLDIR. This is available via the new '-d' option
14173 to 'openssl version', and is also included in 'openssl version -a'.
14177 * Allowing defining memory allocation callbacks that will be given
14178 file name and line number information in additional arguments
14179 (a `const char*` and an int). The basic functionality remains, as
14180 well as the original possibility to just replace malloc(),
14181 realloc() and free() by functions that do not know about these
14182 additional arguments. To register and find out the current
14183 settings for extended allocation functions, the following
14184 functions are provided:
14186 CRYPTO_set_mem_ex_functions
14187 CRYPTO_set_locked_mem_ex_functions
14188 CRYPTO_get_mem_ex_functions
14189 CRYPTO_get_locked_mem_ex_functions
14191 These work the same way as CRYPTO_set_mem_functions and friends.
14192 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14193 extended allocation function is enabled.
14194 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14195 a conventional allocation function is enabled.
14197 *Richard Levitte, Bodo Moeller*
14199 * Finish off removing the remaining LHASH function pointer casts.
14200 There should no longer be any prototype-casting required when using
14201 the LHASH abstraction, and any casts that remain are "bugs". See
14202 the callback types and macros at the head of lhash.h for details
14203 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14207 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14208 If /dev/[u]random devices are not available or do not return enough
14209 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14211 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14212 /etc/entropy will be queried once each in this sequence, querying stops
14213 when enough entropy was collected without querying more sockets.
14217 * Change the Unix RAND_poll() variant to be able to poll several
14218 random devices, as specified by DEVRANDOM, until a sufficient amount
14219 of data has been collected. We spend at most 10 ms on each file
14220 (select timeout) and read in non-blocking mode. DEVRANDOM now
14221 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14222 (previously it was just the string "/dev/urandom"), so on typical
14223 platforms the 10 ms delay will never occur.
14224 Also separate out the Unix variant to its own file, rand_unix.c.
14225 For VMS, there's a currently-empty rand_vms.c.
14229 * Move OCSP client related routines to ocsp_cl.c. These
14230 provide utility functions which an application needing
14231 to issue a request to an OCSP responder and analyse the
14232 response will typically need: as opposed to those which an
14233 OCSP responder itself would need which will be added later.
14235 OCSP_request_sign() signs an OCSP request with an API similar
14236 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14237 response. OCSP_response_get1_basic() extracts basic response
14238 from response. OCSP_resp_find_status(): finds and extracts status
14239 information from an OCSP_CERTID structure (which will be created
14240 when the request structure is built). These are built from lower
14241 level functions which work on OCSP_SINGLERESP structures but
14242 won't normally be used unless the application wishes to examine
14243 extensions in the OCSP response for example.
14245 Replace nonce routines with a pair of functions.
14246 OCSP_request_add1_nonce() adds a nonce value and optionally
14247 generates a random value. OCSP_check_nonce() checks the
14248 validity of the nonce in an OCSP response.
14252 * Change function OCSP_request_add() to OCSP_request_add0_id().
14253 This doesn't copy the supplied OCSP_CERTID and avoids the
14254 need to free up the newly created id. Change return type
14255 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14256 This can then be used to add extensions to the request.
14257 Deleted OCSP_request_new(), since most of its functionality
14258 is now in OCSP_REQUEST_new() (and the case insensitive name
14259 clash) apart from the ability to set the request name which
14260 will be added elsewhere.
14264 * Update OCSP API. Remove obsolete extensions argument from
14265 various functions. Extensions are now handled using the new
14266 OCSP extension code. New simple OCSP HTTP function which
14267 can be used to send requests and parse the response.
14271 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14272 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14273 uses the special reorder version of SET OF to sort the attributes
14274 and reorder them to match the encoded order. This resolves a long
14275 standing problem: a verify on a PKCS7 structure just after signing
14276 it used to fail because the attribute order did not match the
14277 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14278 it uses the received order. This is necessary to tolerate some broken
14279 software that does not order SET OF. This is handled by encoding
14280 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14281 to produce the required SET OF.
14285 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14286 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14287 files to get correct declarations of the ASN.1 item variables.
14291 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14292 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14293 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14294 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14295 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14296 ASN1_ITEM and no wrapper functions.
14300 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14301 replace the old function pointer based I/O routines. Change most of
14302 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14306 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14307 lines, recognize more "algorithms" that can be deselected, and make
14308 it complain about algorithm deselection that isn't recognised.
14312 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14313 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14314 to use new functions. Add NO_ASN1_OLD which can be set to remove
14315 some old style ASN1 functions: this can be used to determine if old
14316 code will still work when these eventually go away.
14320 * New extension functions for OCSP structures, these follow the
14321 same conventions as certificates and CRLs.
14325 * New function X509V3_add1_i2d(). This automatically encodes and
14326 adds an extension. Its behaviour can be customised with various
14327 flags to append, replace or delete. Various wrappers added for
14328 certificates and CRLs.
14332 * Fix to avoid calling the underlying ASN1 print routine when
14333 an extension cannot be parsed. Correct a typo in the
14334 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14338 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14339 entries for variables.
14343 * Add functionality to `apps/openssl.c` for detecting locking
14344 problems: As the program is single-threaded, all we have
14345 to do is register a locking callback using an array for
14346 storing which locks are currently held by the program.
14350 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14351 SSL_get_ex_data_X509_STORE_idx(), which is used in
14352 ssl_verify_cert_chain() and thus can be called at any time
14353 during TLS/SSL handshakes so that thread-safety is essential.
14354 Unfortunately, the ex_data design is not at all suited
14355 for multi-threaded use, so it probably should be abolished.
14359 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14361 *Broadcom, tweaked and integrated by Geoff Thorpe*
14363 * Move common extension printing code to new function
14364 X509V3_print_extensions(). Reorganise OCSP print routines and
14365 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14369 * New function X509_signature_print() to remove duplication in some
14374 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14375 set (this was treated exactly the same as SET OF previously). This
14376 is used to reorder the STACK representing the structure to match the
14377 encoding. This will be used to get round a problem where a PKCS7
14378 structure which was signed could not be verified because the STACK
14379 order did not reflect the encoded order.
14383 * Reimplement the OCSP ASN1 module using the new code.
14387 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14388 for its ASN1 operations. The old style function pointers still exist
14389 for now but they will eventually go away.
14393 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14394 completely replaces the old ASN1 functionality with a table driven
14395 encoder and decoder which interprets an ASN1_ITEM structure describing
14396 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14397 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14398 has also been converted to the new form.
14402 * Change BN_mod_exp_recp so that negative moduli are tolerated
14403 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14404 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14405 for negative moduli.
14409 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14410 of not touching the result's sign bit.
14414 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14419 * Changed the LHASH code to use prototypes for callbacks, and created
14420 macros to declare and implement thin (optionally static) functions
14421 that provide type-safety and avoid function pointer casting for the
14422 type-specific callbacks.
14426 * Added Kerberos Cipher Suites to be used with TLS, as written in
14428 *Veers Staats <staatsvr@asc.hpc.mil>,
14429 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14431 * Reformat the FAQ so the different questions and answers can be divided
14432 in sections depending on the subject.
14436 * Have the zlib compression code load ZLIB.DLL dynamically under
14441 * New function BN_mod_sqrt for computing square roots modulo a prime
14442 (using the probabilistic Tonelli-Shanks algorithm unless
14443 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14444 be handled deterministically).
14446 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14448 * Make BN_mod_inverse faster by explicitly handling small quotients
14449 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14450 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14454 * New function BN_kronecker.
14458 * Fix BN_gcd so that it works on negative inputs; the result is
14459 positive unless both parameters are zero.
14460 Previously something reasonably close to an infinite loop was
14461 possible because numbers could be growing instead of shrinking
14462 in the implementation of Euclid's algorithm.
14466 * Fix BN_is_word() and BN_is_one() macros to take into account the
14467 sign of the number in question.
14469 Fix BN_is_word(a,w) to work correctly for w == 0.
14471 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14472 because its test if the absolute value of 'a' equals 'w'.
14473 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14474 it exists mostly for use in the implementations of BN_is_zero(),
14475 BN_is_one(), and BN_is_word().
14479 * New function BN_swap.
14483 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14484 the exponentiation functions are more likely to produce reasonable
14485 results on negative inputs.
14489 * Change BN_mod_mul so that the result is always non-negative.
14490 Previously, it could be negative if one of the factors was negative;
14491 I don't think anyone really wanted that behaviour.
14495 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14496 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14497 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14498 and add new functions:
14507 BN_mod_lshift1_quick
14509 BN_mod_lshift_quick
14511 These functions always generate non-negative results.
14513 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14514 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14516 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14517 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14518 be reduced modulo `m`.
14520 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14523 The following entry accidentally appeared in the CHANGES file
14524 distributed with OpenSSL 0.9.7. The modifications described in
14525 it do *not* apply to OpenSSL 0.9.7.
14527 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14528 was actually never needed) and in BN_mul(). The removal in BN_mul()
14529 required a small change in bn_mul_part_recursive() and the addition
14530 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14531 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14532 bn_sub_words() and bn_add_words() except they take arrays with
14538 * In 'openssl passwd', verify passwords read from the terminal
14539 unless the '-salt' option is used (which usually means that
14540 verification would just waste user's time since the resulting
14541 hash is going to be compared with some given password hash)
14542 or the new '-noverify' option is used.
14544 This is an incompatible change, but it does not affect
14545 non-interactive use of 'openssl passwd' (passwords on the command
14546 line, '-stdin' option, '-in ...' option) and thus should not
14547 cause any problems.
14551 * Remove all references to RSAref, since there's no more need for it.
14555 * Make DSO load along a path given through an environment variable
14556 (SHLIB_PATH) with shl_load().
14560 * Constify the ENGINE code as a result of BIGNUM constification.
14561 Also constify the RSA code and most things related to it. In a
14562 few places, most notable in the depth of the ASN.1 code, ugly
14563 casts back to non-const were required (to be solved at a later
14568 * Make it so the openssl application has all engines loaded by default.
14572 * Constify the BIGNUM routines a little more.
14576 * Add the following functions:
14578 ENGINE_load_cswift()
14580 ENGINE_load_atalla()
14581 ENGINE_load_nuron()
14582 ENGINE_load_builtin_engines()
14584 That way, an application can itself choose if external engines that
14585 are built-in in OpenSSL shall ever be used or not. The benefit is
14586 that applications won't have to be linked with libdl or other dso
14587 libraries unless it's really needed.
14589 Changed 'openssl engine' to load all engines on demand.
14590 Changed the engine header files to avoid the duplication of some
14591 declarations (they differed!).
14595 * 'openssl engine' can now list capabilities.
14599 * Better error reporting in 'openssl engine'.
14603 * Never call load_dh_param(NULL) in s_server.
14607 * Add engine application. It can currently list engines by name and
14608 identity, and test if they are actually available.
14612 * Improve RPM specification file by forcing symbolic linking and making
14613 sure the installed documentation is also owned by root.root.
14615 *Damien Miller <djm@mindrot.org>*
14617 * Give the OpenSSL applications more possibilities to make use of
14618 keys (public as well as private) handled by engines.
14622 * Add OCSP code that comes from CertCo.
14626 * Add VMS support for the Rijndael code.
14630 * Added untested support for Nuron crypto accelerator.
14634 * Add support for external cryptographic devices. This code was
14635 previously distributed separately as the "engine" branch.
14637 *Geoff Thorpe, Richard Levitte*
14639 * Rework the filename-translation in the DSO code. It is now possible to
14640 have far greater control over how a "name" is turned into a filename
14641 depending on the operating environment and any oddities about the
14642 different shared library filenames on each system.
14646 * Support threads on FreeBSD-elf in Configure.
14650 * Fix for SHA1 assembly problem with MASM: it produces
14651 warnings about corrupt line number information when assembling
14652 with debugging information. This is caused by the overlapping
14655 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14658 NCONF_get_number() has no error checking at all. As a replacement,
14659 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14660 promoted strongly. The old NCONF_get_number is kept around for
14661 binary backward compatibility.
14662 Make it possible for methods to load from something other than a BIO,
14663 by providing a function pointer that is given a name instead of a BIO.
14664 For example, this could be used to load configuration data from an
14669 * Fix for non blocking accept BIOs. Added new I/O special reason
14670 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14671 with non blocking I/O was not possible because no retry code was
14672 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14677 * Added the beginnings of Rijndael support.
14681 * Fix for bug in DirectoryString mask setting. Add support for
14682 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14683 to allow certificate printing to more controllable, additional
14684 'certopt' option to 'x509' to allow new printing options to be
14689 * Clean old EAY MD5 hack from e_os.h.
14693 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14695 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14696 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14698 *Joe Orton, Steve Henson*
14700 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14702 * Fix additional bug revealed by the NISCC test suite:
14704 Stop bug triggering large recursion when presented with
14705 certain ASN.1 tags ([CVE-2003-0851])
14709 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14711 * Fix various bugs revealed by running the NISCC test suite:
14713 Stop out of bounds reads in the ASN1 code when presented with
14714 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14716 If verify callback ignores invalid public key errors don't try to check
14717 certificate signature with the NULL public key.
14721 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14722 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14727 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14728 extra data after the compression methods not only for TLS 1.0
14729 but also for SSL 3.0 (as required by the specification).
14731 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14733 * Change X509_certificate_type() to mark the key as exported/exportable
14734 when it's 512 *bits* long, not 512 bytes.
14738 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14740 * Countermeasure against the Klima-Pokorny-Rosa extension of
14741 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14742 a protocol version number mismatch like a decryption error
14743 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14747 * Turn on RSA blinding by default in the default implementation
14748 to avoid a timing attack. Applications that don't want it can call
14749 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14750 They would be ill-advised to do so in most cases.
14752 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14754 * Change RSA blinding code so that it works when the PRNG is not
14755 seeded (in this case, the secret RSA exponent is abused as
14756 an unpredictable seed -- if it is not unpredictable, there
14757 is no point in blinding anyway). Make RSA blinding thread-safe
14758 by remembering the creator's thread ID in rsa->blinding and
14759 having all other threads use local one-time blinding factors
14760 (this requires more computation than sharing rsa->blinding, but
14761 avoids excessive locking; and if an RSA object is not shared
14762 between threads, blinding will still be very fast).
14766 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14768 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14769 via timing by performing a MAC computation even if incorrect
14770 block cipher padding has been found. This is a countermeasure
14771 against active attacks where the attacker has to distinguish
14772 between bad padding and a MAC verification error. ([CVE-2003-0078])
14774 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14775 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14776 Martin Vuagnoux (EPFL, Ilion)*
14778 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14780 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14781 memory from its contents. This is done with a counter that will
14782 place alternating values in each byte. This can be used to solve
14783 two issues: 1) the removal of calls to memset() by highly optimizing
14784 compilers, and 2) cleansing with other values than 0, since those can
14785 be read through on certain media, for example a swap space on disk.
14789 * Bugfix: client side session caching did not work with external caching,
14790 because the session->cipher setting was not restored when reloading
14791 from the external cache. This problem was masked, when
14792 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14793 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14797 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14798 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14800 *Zeev Lieber <zeev-l@yahoo.com>*
14802 * Undo an undocumented change introduced in 0.9.6e which caused
14803 repeated calls to OpenSSL_add_all_ciphers() and
14804 OpenSSL_add_all_digests() to be ignored, even after calling
14809 * Change the default configuration reader to deal with last line not
14810 being properly terminated.
14814 * Change X509_NAME_cmp() so it applies the special rules on handling
14815 DN values that are of type PrintableString, as well as RDNs of type
14816 emailAddress where the value has the type ia5String.
14818 *stefank@valicert.com via Richard Levitte*
14820 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14821 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14822 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14823 the bitwise-OR of the two for use by the majority of applications
14824 wanting this behaviour, and update the docs. The documented
14825 behaviour and actual behaviour were inconsistent and had been
14826 changing anyway, so this is more a bug-fix than a behavioural
14829 *Geoff Thorpe, diagnosed by Nadav Har'El*
14831 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14832 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14836 * Fix initialization code race conditions in
14837 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14838 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14839 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14840 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14841 ssl2_get_cipher_by_char(),
14842 ssl3_get_cipher_by_char().
14844 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14846 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14847 the cached sessions are flushed, as the remove_cb() might use ex_data
14848 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14849 (see [openssl.org #212]).
14851 *Geoff Thorpe, Lutz Jaenicke*
14853 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14854 length, instead of the encoding length to d2i_ASN1_OBJECT.
14858 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14860 * [In 0.9.6g-engine release:]
14861 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14863 *Lynn Gazis <lgazis@rainbow.com>*
14865 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14867 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14868 and get fix the header length calculation.
14869 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14870 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14872 * Use proper error handling instead of 'assertions' in buffer
14873 overflow checks added in 0.9.6e. This prevents DoS (the
14874 assertions could call abort()).
14876 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14878 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14880 * Add various sanity checks to asn1_get_length() to reject
14881 the ASN1 length bytes if they exceed sizeof(long), will appear
14882 negative or the content length exceeds the length of the
14885 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14887 * Fix cipher selection routines: ciphers without encryption had no flags
14888 for the cipher strength set and where therefore not handled correctly
14889 by the selection routines (PR #130).
14893 * Fix EVP_dsa_sha macro.
14898 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14899 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14900 that was added in OpenSSL 0.9.6d.
14902 As the countermeasure turned out to be incompatible with some
14903 broken SSL implementations, the new option is part of SSL_OP_ALL.
14904 SSL_OP_ALL is usually employed when compatibility with weird SSL
14905 implementations is desired (e.g. '-bugs' option to 's_client' and
14906 's_server'), so the new option is automatically set in many
14911 * Changes in security patch:
14913 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14914 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14915 Air Force Materiel Command, USAF, under agreement number
14918 * Add various sanity checks to asn1_get_length() to reject
14919 the ASN1 length bytes if they exceed sizeof(long), will appear
14920 negative or the content length exceeds the length of the
14921 supplied buffer. ([CVE-2002-0659])
14923 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14925 * Assertions for various potential buffer overflows, not known to
14926 happen in practice.
14928 *Ben Laurie (CHATS)*
14930 * Various temporary buffers to hold ASCII versions of integers were
14931 too small for 64 bit platforms. ([CVE-2002-0655])
14932 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14934 * Remote buffer overflow in SSL3 protocol - an attacker could
14935 supply an oversized session ID to a client. ([CVE-2002-0656])
14937 *Ben Laurie (CHATS)*
14939 * Remote buffer overflow in SSL2 protocol - an attacker could
14940 supply an oversized client master key. ([CVE-2002-0656])
14942 *Ben Laurie (CHATS)*
14944 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14946 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14947 encoded as NULL) with id-dsa-with-sha1.
14949 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14951 * Check various `X509_...()` return values in `apps/req.c`.
14953 *Nils Larsch <nla@trustcenter.de>*
14955 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14956 an end-of-file condition would erroneously be flagged, when the CRLF
14957 was just at the end of a processed block. The bug was discovered when
14958 processing data through a buffering memory BIO handing the data to a
14959 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14960 <ptsekov@syntrex.com> and Nedelcho Stanev.
14964 * Implement a countermeasure against a vulnerability recently found
14965 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14966 before application data chunks to avoid the use of known IVs
14967 with data potentially chosen by the attacker.
14971 * Fix length checks in ssl3_get_client_hello().
14975 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14976 to prevent ssl3_read_internal() from incorrectly assuming that
14977 ssl3_read_bytes() found application data while handshake
14978 processing was enabled when in fact s->s3->in_read_app_data was
14979 merely automatically cleared during the initial handshake.
14981 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14983 * Fix object definitions for Private and Enterprise: they were not
14984 recognized in their shortname (=lowercase) representation. Extend
14985 obj_dat.pl to issue an error when using undefined keywords instead
14986 of silently ignoring the problem (Svenning Sorensen
14987 <sss@sss.dnsalias.net>).
14991 * Fix DH_generate_parameters() so that it works for 'non-standard'
14992 generators, i.e. generators other than 2 and 5. (Previously, the
14993 code did not properly initialise the 'add' and 'rem' values to
14994 BN_generate_prime().)
14996 In the new general case, we do not insist that 'generator' is
14997 actually a primitive root: This requirement is rather pointless;
14998 a generator of the order-q subgroup is just as good, if not
15003 * Map new X509 verification errors to alerts. Discovered and submitted by
15004 Tom Wu <tom@arcot.com>.
15008 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
15009 returning non-zero before the data has been completely received
15010 when using non-blocking I/O.
15012 *Bodo Moeller; problem pointed out by John Hughes*
15014 * Some of the ciphers missed the strength entry (SSL_LOW etc).
15016 *Ben Laurie, Lutz Jaenicke*
15018 * Fix bug in SSL_clear(): bad sessions were not removed (found by
15019 Yoram Zahavi <YoramZ@gilian.com>).
15023 * Add information about CygWin 1.3 and on, and preserve proper
15024 configuration for the versions before that.
15026 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
15028 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
15029 check whether we deal with a copy of a session and do not delete from
15030 the cache in this case. Problem reported by "Izhar Shoshani Levi"
15031 <izhar@checkpoint.com>.
15035 * Do not store session data into the internal session cache, if it
15036 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
15037 flag is set). Proposed by Aslam <aslam@funk.com>.
15041 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
15046 * [In 0.9.6d-engine release:]
15047 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
15049 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
15051 * Add the configuration target linux-s390x.
15053 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
15055 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
15056 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
15057 variable as an indication that a ClientHello message has been
15058 received. As the flag value will be lost between multiple
15059 invocations of ssl3_accept when using non-blocking I/O, the
15060 function may not be aware that a handshake has actually taken
15061 place, thus preventing a new session from being added to the
15064 To avoid this problem, we now set s->new_session to 2 instead of
15065 using a local variable.
15067 *Lutz Jaenicke, Bodo Moeller*
15069 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15070 if the SSL_R_LENGTH_MISMATCH error is detected.
15072 *Geoff Thorpe, Bodo Moeller*
15074 * New 'shared_ldflag' column in Configure platform table.
15078 * Fix EVP_CIPHER_mode macro.
15080 *"Dan S. Camper" <dan@bti.net>*
15082 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15083 type, we must throw them away by setting rr->length to 0.
15085 *D P Chang <dpc@qualys.com>*
15087 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15089 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15090 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15091 worked incorrectly for those cases where range = `10..._2` and
15092 `3*range` is two bits longer than range.)
15096 * Only add signing time to PKCS7 structures if it is not already
15101 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15102 OBJ_ld_ce should be OBJ_id_ce.
15103 Also some ip-pda OIDs in crypto/objects/objects.txt were
15104 incorrect (cf. RFC 3039).
15106 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15108 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15109 returns early because it has nothing to do.
15111 *Andy Schneider <andy.schneider@bjss.co.uk>*
15113 * [In 0.9.6c-engine release:]
15114 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15116 *Andy Schneider <andy.schneider@bjss.co.uk>*
15118 * [In 0.9.6c-engine release:]
15119 Add support for Cryptographic Appliance's keyserver technology.
15120 (Use engine 'keyclient')
15122 *Cryptographic Appliances and Geoff Thorpe*
15124 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15125 is called via tools/c89.sh because arguments have to be
15126 rearranged (all '-L' options must appear before the first object
15129 *Richard Shapiro <rshapiro@abinitio.com>*
15131 * [In 0.9.6c-engine release:]
15132 Add support for Broadcom crypto accelerator cards, backported
15135 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15137 * [In 0.9.6c-engine release:]
15138 Add support for SureWare crypto accelerator cards from
15139 Baltimore Technologies. (Use engine 'sureware')
15141 *Baltimore Technologies and Mark Cox*
15143 * [In 0.9.6c-engine release:]
15144 Add support for crypto accelerator cards from Accelerated
15145 Encryption Processing, www.aep.ie. (Use engine 'aep')
15147 *AEP Inc. and Mark Cox*
15149 * Add a configuration entry for gcc on UnixWare.
15151 *Gary Benson <gbenson@redhat.com>*
15153 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15154 messages are stored in a single piece (fixed-length part and
15155 variable-length part combined) and fix various bugs found on the way.
15159 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15160 instead. BIO_gethostbyname() does not know what timeouts are
15161 appropriate, so entries would stay in cache even when they have
15163 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15165 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15166 faced with a pathologically small ClientHello fragment that does
15167 not contain client_version: Instead of aborting with an error,
15168 simply choose the highest available protocol version (i.e.,
15169 TLS 1.0 unless it is disabled). In practice, ClientHello
15170 messages are never sent like this, but this change gives us
15171 strictly correct behaviour at least for TLS.
15175 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15176 never resets s->method to s->ctx->method when called from within
15177 one of the SSL handshake functions.
15179 *Bodo Moeller; problem pointed out by Niko Baric*
15181 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15182 (sent using the client's version number) if client_version is
15183 smaller than the protocol version in use. Also change
15184 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15185 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15186 the client will at least see that alert.
15190 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15195 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15196 client receives HelloRequest while in a handshake.
15198 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15200 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15201 should end in 'break', not 'goto end' which circumvents various
15202 cleanups done in state SSL_ST_OK. But session related stuff
15203 must be disabled for SSL_ST_OK in the case that we just sent a
15206 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15207 before just sending a HelloRequest.
15209 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15211 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15212 reveal whether illegal block cipher padding was found or a MAC
15213 verification error occurred. (Neither SSLerr() codes nor alerts
15214 are directly visible to potential attackers, but the information
15215 may leak via logfiles.)
15217 Similar changes are not required for the SSL 2.0 implementation
15218 because the number of padding bytes is sent in clear for SSL 2.0,
15219 and the extra bytes are just ignored. However ssl/s2_pkt.c
15220 failed to verify that the purported number of padding bytes is in
15225 * Add OpenUNIX-8 support including shared libraries
15226 (Boyd Lynn Gerber <gerberb@zenez.com>).
15230 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15231 'wristwatch attack' using huge encoding parameters (cf.
15232 James H. Manger's CRYPTO 2001 paper). Note that the
15233 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15234 encoding parameters and hence was not vulnerable.
15238 * BN_sqr() bug fix.
15240 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15242 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15243 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15244 followed by modular reduction.
15246 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15248 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15249 equivalent based on BN_pseudo_rand() instead of BN_rand().
15253 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15254 This function was broken, as the check for a new client hello message
15255 to handle SGC did not allow these large messages.
15256 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15260 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15264 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15265 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15269 * Rework the configuration and shared library support for Tru64 Unix.
15270 The configuration part makes use of modern compiler features and
15271 still retains old compiler behavior for those that run older versions
15272 of the OS. The shared library support part includes a variant that
15273 uses the RPATH feature, and is available through the special
15274 configuration target "alpha-cc-rpath", which will never be selected
15277 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15279 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15280 with the same message size as in ssl3_get_certificate_request().
15281 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15282 messages might inadvertently be reject as too long.
15284 *Petr Lampa <lampa@fee.vutbr.cz>*
15286 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15290 * Modified SSL library such that the verify_callback that has been set
15291 specifically for an SSL object with SSL_set_verify() is actually being
15292 used. Before the change, a verify_callback set with this function was
15293 ignored and the verify_callback() set in the SSL_CTX at the time of
15294 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15295 to allow the necessary settings.
15299 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15300 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15301 done automatically (in contradiction to the requirements of the C
15302 standard). This made problems when used from OpenSSH.
15306 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15307 dh->length and always used
15309 BN_rand_range(priv_key, dh->p).
15311 BN_rand_range() is not necessary for Diffie-Hellman, and this
15312 specific range makes Diffie-Hellman unnecessarily inefficient if
15313 dh->length (recommended exponent length) is much smaller than the
15314 length of dh->p. We could use BN_rand_range() if the order of
15315 the subgroup was stored in the DH structure, but we only have
15320 BN_rand(priv_key, l, ...)
15322 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15329 RSA_eay_public_encrypt
15330 RSA_eay_private_decrypt
15331 RSA_eay_private_encrypt (signing)
15332 RSA_eay_public_decrypt (signature verification)
15334 (default implementations for RSA_public_encrypt,
15335 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15336 always reject numbers >= n.
15340 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15341 to synchronize access to 'locking_thread'. This is necessary on
15342 systems where access to 'locking_thread' (an 'unsigned long'
15343 variable) is not atomic.
15347 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15348 *before* setting the 'crypto_lock_rand' flag. The previous code had
15349 a race condition if 0 is a valid thread ID.
15351 *Travis Vitek <vitek@roguewave.com>*
15353 * Add support for shared libraries under Irix.
15355 *Albert Chin-A-Young <china@thewrittenword.com>*
15357 * Add configuration option to build on Linux on both big-endian and
15358 little-endian MIPS.
15360 *Ralf Baechle <ralf@uni-koblenz.de>*
15362 * Add the possibility to create shared libraries on HP-UX.
15366 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15368 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15369 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15370 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15371 PRNG state recovery was possible based on the output of
15372 one PRNG request appropriately sized to gain knowledge on
15373 'md' followed by enough consecutive 1-byte PRNG requests
15374 to traverse all of 'state'.
15376 1. When updating 'md_local' (the current thread's copy of 'md')
15377 during PRNG output generation, hash all of the previous
15378 'md_local' value, not just the half used for PRNG output.
15380 2. Make the number of bytes from 'state' included into the hash
15381 independent from the number of PRNG bytes requested.
15383 The first measure alone would be sufficient to avoid
15384 Markku-Juhani's attack. (Actually it had never occurred
15385 to me that the half of 'md_local' used for chaining was the
15386 half from which PRNG output bytes were taken -- I had always
15387 assumed that the secret half would be used.) The second
15388 measure makes sure that additional data from 'state' is never
15389 mixed into 'md_local' in small portions; this heuristically
15390 further strengthens the PRNG.
15394 * Fix crypto/bn/asm/mips3.s.
15398 * When only the key is given to "enc", the IV is undefined. Print out
15399 an error message in this case.
15403 * Handle special case when X509_NAME is empty in X509 printing routines.
15407 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15408 positive and less than q.
15412 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15413 used: it isn't thread safe and the add_lock_callback should handle
15416 *Paul Rose <Paul.Rose@bridge.com>*
15418 * Verify that incoming data obeys the block size in
15419 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15425 *Ulf Möller, Bodo Möller*
15427 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15428 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15429 when fixing the server behaviour for backwards-compatible 'client
15430 hello' messages. (Note that the attack is impractical against
15431 SSL 3.0 and TLS 1.0 anyway because length and version checking
15432 means that the probability of guessing a valid ciphertext is
15433 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15436 Before 0.9.5, the countermeasure (hide the error by generating a
15437 random 'decryption result') did not work properly because
15438 ERR_clear_error() was missing, meaning that SSL_get_error() would
15439 detect the supposedly ignored error.
15441 Both problems are now fixed.
15445 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15446 (previously it was 1024).
15450 * Fix for compatibility mode trust settings: ignore trust settings
15451 unless some valid trust or reject settings are present.
15455 * Fix for blowfish EVP: its a variable length cipher.
15459 * Fix various bugs related to DSA S/MIME verification. Handle missing
15460 parameters in DSA public key structures and return an error in the
15461 DSA routines if parameters are absent.
15465 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15466 in the current directory if neither $RANDFILE nor $HOME was set.
15467 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15468 caused some confusion to Windows users who haven't defined $HOME.
15469 Thus RAND_file_name() is changed again: e_os.h can define a
15470 DEFAULT_HOME, which will be used if $HOME is not set.
15471 For Windows, we use "C:"; on other platforms, we still require
15472 environment variables.
15474 * Move 'if (!initialized) RAND_poll()' into regions protected by
15475 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15476 having multiple threads call RAND_poll() concurrently.
15480 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15481 combination of a flag and a thread ID variable.
15482 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15483 flag), *other* threads can enter ssleay_add_bytes without obeying
15484 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15485 that they do not hold after the first thread unsets add_do_not_lock).
15489 * Change bctest again: '-x' expressions are not available in all
15490 versions of 'test'.
15494 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15496 * Fix a couple of memory leaks in PKCS7_dataDecode()
15498 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15500 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15501 the default extension for executables, if any. Also, make the perl
15502 scripts that use symlink() to test if it really exists and use "cp"
15503 if it doesn't. All this made OpenSSL compilable and installable in
15508 * Fix for asn1_GetSequence() for indefinite length constructed data.
15509 If SEQUENCE is length is indefinite just set c->slen to the total
15510 amount of data available.
15512 *Steve Henson, reported by shige@FreeBSD.org*
15514 *This change does not apply to 0.9.7.*
15516 * Change bctest to avoid here-documents inside command substitution
15517 (workaround for FreeBSD /bin/sh bug).
15518 For compatibility with Ultrix, avoid shell functions (introduced
15519 in the bctest version that searches along $PATH).
15523 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15524 with des_encrypt() defined on some operating systems, like Solaris
15529 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15530 On the Importance of Eliminating Errors in Cryptographic
15531 Computations, J. Cryptology 14 (2001) 2, 101-119,
15532 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15536 * MIPS assembler BIGNUM division bug fix.
15540 * Disabled incorrect Alpha assembler code.
15544 * Fix PKCS#7 decode routines so they correctly update the length
15545 after reading an EOC for the EXPLICIT tag.
15549 *This change does not apply to 0.9.7.*
15551 * Fix bug in PKCS#12 key generation routines. This was triggered
15552 if a 3DES key was generated with a 0 initial byte. Include
15553 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15554 (but broken) behaviour.
15558 * Enhance bctest to search for a working bc along $PATH and print
15561 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15563 * Fix memory leaks in err.c: free err_data string if necessary;
15564 don't write to the wrong index in ERR_set_error_data.
15568 * Implement ssl23_peek (analogous to ssl23_read), which previously
15573 * Replace rdtsc with `_emit` statements for VC++ version 5.
15575 *Jeremy Cooper <jeremy@baymoo.org>*
15577 * Make it possible to reuse SSLv2 sessions.
15581 * In copy_email() check for >= 0 as a return value for
15582 X509_NAME_get_index_by_NID() since 0 is a valid index.
15584 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15586 * Avoid coredump with unsupported or invalid public keys by checking if
15587 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15588 PKCS7_verify() fails with non detached data.
15592 * Don't use getenv in library functions when run as setuid/setgid.
15593 New function OPENSSL_issetugid().
15597 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15598 due to incorrect handling of multi-threading:
15600 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15602 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15604 3. Count how many times MemCheck_off() has been called so that
15605 nested use can be treated correctly. This also avoids
15606 inband-signalling in the previous code (which relied on the
15607 assumption that thread ID 0 is impossible).
15611 * Add "-rand" option also to s_client and s_server.
15615 * Fix CPU detection on Irix 6.x.
15616 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15617 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15619 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15624 *This change does not apply to 0.9.7.*
15626 * Use the cached encoding of an X509_NAME structure rather than
15627 copying it. This is apparently the reason for the libsafe "errors"
15628 but the code is actually correct.
15632 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15633 Bleichenbacher's DSA attack.
15634 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15635 to be set and top=0 forces the highest bit to be set; top=-1 is new
15636 and leaves the highest bit random.
15638 *Ulf Moeller, Bodo Moeller*
15640 * In the `NCONF_...`-based implementations for `CONF_...` queries
15641 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15642 a temporary CONF structure with the data component set to NULL
15643 (which gives segmentation faults in lh_retrieve).
15644 Instead, use NULL for the CONF pointer in CONF_get_string and
15645 CONF_get_number (which may use environment variables) and directly
15646 return NULL from CONF_get_section.
15650 * Fix potential buffer overrun for EBCDIC.
15654 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15655 keyUsage if basicConstraints absent for a CA.
15659 * Make SMIME_write_PKCS7() write mail header values with a format that
15660 is more generally accepted (no spaces before the semicolon), since
15661 some programs can't parse those values properly otherwise. Also make
15662 sure BIO's that break lines after each write do not create invalid
15667 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15668 macros previously used would not encode an empty SEQUENCE OF
15669 and break the signature.
15673 *This change does not apply to 0.9.7.*
15675 * Zero the premaster secret after deriving the master secret in
15680 * Add some EVP_add_digest_alias registrations (as found in
15681 OpenSSL_add_all_digests()) to SSL_library_init()
15682 aka OpenSSL_add_ssl_algorithms(). This provides improved
15683 compatibility with peers using X.509 certificates
15684 with unconventional AlgorithmIdentifier OIDs.
15688 * Fix for Irix with NO_ASM.
15690 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15692 * ./config script fixes.
15694 *Ulf Moeller, Richard Levitte*
15696 * Fix 'openssl passwd -1'.
15700 * Change PKCS12_key_gen_asc() so it can cope with non null
15701 terminated strings whose length is passed in the passlen
15702 parameter, for example from PEM callbacks. This was done
15703 by adding an extra length parameter to asc2uni().
15705 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15707 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15708 call failed, free the DSA structure.
15712 * Fix to uni2asc() to cope with zero length Unicode strings.
15713 These are present in some PKCS#12 files.
15717 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15718 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15719 when writing a 32767 byte record.
15721 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15723 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15724 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15726 (RSA objects have a reference count access to which is protected
15727 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15728 so they are meant to be shared between threads.)
15729 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15730 "Reddie, Steven" <Steven.Reddie@ca.com>*
15732 * Fix a deadlock in CRYPTO_mem_leaks().
15736 * Use better test patterns in bntest.
15740 * rand_win.c fix for Borland C.
15744 * BN_rshift bugfix for n == 0.
15748 * Add a 'bctest' script that checks for some known 'bc' bugs
15749 so that 'make test' does not abort just because 'bc' is broken.
15753 * Store verify_result within SSL_SESSION also for client side to
15754 avoid potential security hole. (Re-used sessions on the client side
15755 always resulted in verify_result==X509_V_OK, not using the original
15756 result of the server certificate verification.)
15760 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15761 SSL3_RT_APPLICATION_DATA, return 0.
15762 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15767 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15768 releases, have been re-implemented by renaming the previous
15769 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15770 and ssl3_read_internal, respectively, and adding 'peek' parameters
15771 to them. The new ssl[23]_{read,peek} functions are calls to
15772 ssl[23]_read_internal with the 'peek' flag set appropriately.
15773 A 'peek' parameter has also been added to ssl3_read_bytes, which
15774 does the actual work for ssl3_read_internal.
15778 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15779 the method-specific "init()" handler. Also clean up ex_data after
15780 calling the method-specific "finish()" handler. Previously, this was
15781 happening the other way round.
15785 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15786 The previous value, 12, was not always sufficient for BN_mod_exp().
15790 * Make sure that shared libraries get the internal name engine with
15791 the full version number and not just 0. This should mark the
15792 shared libraries as not backward compatible. Of course, this should
15793 be changed again when we can guarantee backward binary compatibility.
15797 * Fix typo in get_cert_by_subject() in by_dir.c
15799 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15801 * Rework the system to generate shared libraries:
15803 - Make note of the expected extension for the shared libraries and
15804 if there is a need for symbolic links from for example libcrypto.so.0
15805 to libcrypto.so.0.9.7. There is extended info in Configure for
15808 - Make as few rebuilds of the shared libraries as possible.
15810 - Still avoid linking the OpenSSL programs with the shared libraries.
15812 - When installing, install the shared libraries separately from the
15817 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15819 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15820 and not in SSL_clear because the latter is also used by the
15821 accept/connect functions; previously, the settings made by
15822 SSL_set_read_ahead would be lost during the handshake.
15824 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15826 * Correct util/mkdef.pl to be selective about disabled algorithms.
15827 Previously, it would create entries for disabled algorithms no
15832 * Added several new manual pages for SSL_* function.
15836 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15838 * In ssl23_get_client_hello, generate an error message when faced
15839 with an initial SSL 3.0/TLS record that is too small to contain the
15840 first two bytes of the ClientHello message, i.e. client_version.
15841 (Note that this is a pathologic case that probably has never happened
15842 in real life.) The previous approach was to use the version number
15843 from the record header as a substitute; but our protocol choice
15844 should not depend on that one because it is not authenticated
15845 by the Finished messages.
15849 * More robust randomness gathering functions for Windows.
15851 *Jeffrey Altman <jaltman@columbia.edu>*
15853 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15854 not set then we don't setup the error code for issuer check errors
15855 to avoid possibly overwriting other errors which the callback does
15856 handle. If an application does set the flag then we assume it knows
15857 what it is doing and can handle the new informational codes
15862 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15863 a general "ANY" type, as such it should be able to decode anything
15864 including tagged types. However it didn't check the class so it would
15865 wrongly interpret tagged types in the same way as their universal
15866 counterpart and unknown types were just rejected. Changed so that the
15867 tagged and unknown types are handled in the same way as a SEQUENCE:
15868 that is the encoding is stored intact. There is also a new type
15869 "V_ASN1_OTHER" which is used when the class is not universal, in this
15870 case we have no idea what the actual type is so we just lump them all
15875 * On VMS, stdout may very well lead to a file that is written to
15876 in a record-oriented fashion. That means that every write() will
15877 write a separate record, which will be read separately by the
15878 programs trying to read from it. This can be very confusing.
15880 The solution is to put a BIO filter in the way that will buffer
15881 text until a linefeed is reached, and then write everything a
15882 line at a time, so every record written will be an actual line,
15883 not chunks of lines and not (usually doesn't happen, but I've
15884 seen it once) several lines in one record. BIO_f_linebuffer() is
15887 Currently, it's a VMS-only method, because that's where it has
15888 been tested well enough.
15892 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15893 it can return incorrect results.
15894 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15895 but it was in 0.9.6-beta[12].)
15899 * Disable the check for content being present when verifying detached
15900 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15901 include zero length content when signing messages.
15905 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15906 BIO_ctrl (for BIO pairs).
15910 * Add DSO method for VMS.
15914 * Bug fix: Montgomery multiplication could produce results with the
15919 * Add RPM specification openssl.spec and modify it to build three
15920 packages. The default package contains applications, application
15921 documentation and run-time libraries. The devel package contains
15922 include files, static libraries and function documentation. The
15923 doc package contains the contents of the doc directory. The original
15924 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15928 * Add a large number of documentation files for many SSL routines.
15930 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15932 * Add a configuration entry for Sony News 4.
15934 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15936 * Don't set the two most significant bits to one when generating a
15937 random number < q in the DSA library.
15941 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15942 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15943 the underlying transport is blocking) if a handshake took place.
15944 (The default behaviour is needed by applications such as s_client
15945 and s_server that use select() to determine when to use SSL_read;
15946 but for applications that know in advance when to expect data, it
15947 just makes things more complicated.)
15951 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15956 * Add a few more EBCDIC conditionals that make `req` and `x509`
15957 work better on such systems.
15959 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15961 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15962 Update PKCS12_parse() so it copies the friendlyName and the
15963 keyid to the certificates aux info.
15967 * Fix bug in PKCS7_verify() which caused an infinite loop
15968 if there was more than one signature.
15970 *Sven Uszpelkat <su@celocom.de>*
15972 * Major change in util/mkdef.pl to include extra information
15973 about each symbol, as well as presenting variables as well
15974 as functions. This change means that there's n more need
15975 to rebuild the .num files when some algorithms are excluded.
15979 * Allow the verify time to be set by an application,
15980 rather than always using the current time.
15984 * Phase 2 verify code reorganisation. The certificate
15985 verify code now looks up an issuer certificate by a
15986 number of criteria: subject name, authority key id
15987 and key usage. It also verifies self signed certificates
15988 by the same criteria. The main comparison function is
15989 X509_check_issued() which performs these checks.
15991 Lot of changes were necessary in order to support this
15992 without completely rewriting the lookup code.
15994 Authority and subject key identifier are now cached.
15996 The LHASH 'certs' is X509_STORE has now been replaced
15997 by a STACK_OF(X509_OBJECT). This is mainly because an
15998 LHASH can't store or retrieve multiple objects with
15999 the same hash value.
16001 As a result various functions (which were all internal
16002 use only) have changed to handle the new X509_STORE
16003 structure. This will break anything that messed round
16004 with X509_STORE internally.
16006 The functions X509_STORE_add_cert() now checks for an
16007 exact match, rather than just subject name.
16009 The X509_STORE API doesn't directly support the retrieval
16010 of multiple certificates matching a given criteria, however
16011 this can be worked round by performing a lookup first
16012 (which will fill the cache with candidate certificates)
16013 and then examining the cache for matches. This is probably
16014 the best we can do without throwing out X509_LOOKUP
16015 entirely (maybe later...).
16017 The X509_VERIFY_CTX structure has been enhanced considerably.
16019 All certificate lookup operations now go via a get_issuer()
16020 callback. Although this currently uses an X509_STORE it
16021 can be replaced by custom lookups. This is a simple way
16022 to bypass the X509_STORE hackery necessary to make this
16023 work and makes it possible to use more efficient techniques
16024 in future. A very simple version which uses a simple
16025 STACK for its trusted certificate store is also provided
16026 using X509_STORE_CTX_trusted_stack().
16028 The verify_cb() and verify() callbacks now have equivalents
16029 in the X509_STORE_CTX structure.
16031 X509_STORE_CTX also has a 'flags' field which can be used
16032 to customise the verify behaviour.
16036 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
16037 excludes S/MIME capabilities.
16041 * When a certificate request is read in keep a copy of the
16042 original encoding of the signed data and use it when outputting
16043 again. Signatures then use the original encoding rather than
16044 a decoded, encoded version which may cause problems if the
16045 request is improperly encoded.
16049 * For consistency with other BIO_puts implementations, call
16050 buffer_write(b, ...) directly in buffer_puts instead of calling
16053 In BIO_puts, increment b->num_write as in BIO_write.
16055 *Peter.Sylvester@EdelWeb.fr*
16057 * Fix BN_mul_word for the case where the word is 0. (We have to use
16058 BN_zero, we may not return a BIGNUM with an array consisting of
16059 words set to zero.)
16063 * Avoid calling abort() from within the library when problems are
16064 detected, except if preprocessor symbols have been defined
16065 (such as REF_CHECK, BN_DEBUG etc.).
16069 * New openssl application 'rsautl'. This utility can be
16070 used for low-level RSA operations. DER public key
16071 BIO/fp routines also added.
16075 * New Configure entry and patches for compiling on QNX 4.
16077 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16079 * A demo state-machine implementation was sponsored by
16080 Nuron (<http://www.nuron.com/>) and is now available in
16081 demos/state_machine.
16085 * New options added to the 'dgst' utility for signature
16086 generation and verification.
16090 * Unrecognized PKCS#7 content types are now handled via a
16091 catch all ASN1_TYPE structure. This allows unsupported
16092 types to be stored as a "blob" and an application can
16093 encode and decode it manually.
16097 * Fix various signed/unsigned issues to make a_strex.c
16098 compile under VC++.
16100 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16102 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16103 length if passed a buffer. ASN1_INTEGER_to_BN failed
16104 if passed a NULL BN and its argument was negative.
16106 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16108 * Modification to PKCS#7 encoding routines to output definite
16109 length encoding. Since currently the whole structures are in
16110 memory there's not real point in using indefinite length
16111 constructed encoding. However if OpenSSL is compiled with
16112 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16116 * Added BIO_vprintf() and BIO_vsnprintf().
16120 * Added more prefixes to parse for in the strings written
16121 through a logging bio, to cover all the levels that are available
16122 through syslog. The prefixes are now:
16124 PANIC, EMERG, EMR => LOG_EMERG
16125 ALERT, ALR => LOG_ALERT
16126 CRIT, CRI => LOG_CRIT
16127 ERROR, ERR => LOG_ERR
16128 WARNING, WARN, WAR => LOG_WARNING
16129 NOTICE, NOTE, NOT => LOG_NOTICE
16130 INFO, INF => LOG_INFO
16131 DEBUG, DBG => LOG_DEBUG
16133 and as before, if none of those prefixes are present at the
16134 beginning of the string, LOG_ERR is chosen.
16136 On Win32, the `LOG_*` levels are mapped according to this:
16138 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16139 LOG_WARNING => EVENTLOG_WARNING_TYPE
16140 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16144 * Made it possible to reconfigure with just the configuration
16145 argument "reconf" or "reconfigure". The command line arguments
16146 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16147 and are retrieved from there when reconfiguring.
16153 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16155 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16159 * The obj_dat.pl script was messing up the sorting of object
16160 names. The reason was that it compared the quoted version
16161 of strings as a result "OCSP" > "OCSP Signing" because
16162 " > SPACE. Changed script to store unquoted versions of
16163 names and add quotes on output. It was also omitting some
16164 names from the lookup table if they were given a default
16165 value (that is if SN is missing it is given the same
16166 value as LN and vice versa), these are now added on the
16167 grounds that if an object has a name we should be able to
16168 look it up. Finally added warning output when duplicate
16169 short or long names are found.
16173 * Changes needed for Tandem NSK.
16175 *Scott Uroff <scott@xypro.com>*
16177 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16178 RSA_padding_check_SSLv23(), special padding was never detected
16179 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16180 version rollback attacks was not effective.
16182 In s23_clnt.c, don't use special rollback-attack detection padding
16183 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16184 client; similarly, in s23_srvr.c, don't do the rollback check if
16185 SSL 2.0 is the only protocol enabled in the server.
16189 * Make it possible to get hexdumps of unprintable data with 'openssl
16190 asn1parse'. By implication, the functions ASN1_parse_dump() and
16191 BIO_dump_indent() are added.
16195 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16196 these print out strings and name structures based on various
16197 flags including RFC2253 support and proper handling of
16198 multibyte characters. Added options to the 'x509' utility
16199 to allow the various flags to be set.
16203 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16204 Also change the functions X509_cmp_current_time() and
16205 X509_gmtime_adj() work with an ASN1_TIME structure,
16206 this will enable certificates using GeneralizedTime in validity
16207 dates to be checked.
16211 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16212 negative public key encodings) on by default,
16213 NO_NEG_PUBKEY_BUG can be set to disable it.
16217 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16218 content octets. An i2c_ASN1_OBJECT is unnecessary because
16219 the encoding can be trivially obtained from the structure.
16223 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16224 not read locks (`CRYPTO_r_[un]lock`).
16228 * A first attempt at creating official support for shared
16229 libraries through configuration. I've kept it so the
16230 default is static libraries only, and the OpenSSL programs
16231 are always statically linked for now, but there are
16232 preparations for dynamic linking in place.
16233 This has been tested on Linux and Tru64.
16237 * Randomness polling function for Win9x, as described in:
16238 Peter Gutmann, Software Generation of Practically Strong
16243 * Fix so PRNG is seeded in req if using an already existing
16248 * New options to smime application. -inform and -outform
16249 allow alternative formats for the S/MIME message including
16250 PEM and DER. The -content option allows the content to be
16251 specified separately. This should allow things like Netscape
16252 form signing output easier to verify.
16256 * Fix the ASN1 encoding of tags using the 'long form'.
16260 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16261 STRING types. These convert content octets to and from the
16262 underlying type. The actual tag and length octets are
16263 already assumed to have been read in and checked. These
16264 are needed because all other string types have virtually
16265 identical handling apart from the tag. By having versions
16266 of the ASN1 functions that just operate on content octets
16267 IMPLICIT tagging can be handled properly. It also allows
16268 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16269 and ASN1_INTEGER are identical apart from the tag.
16273 * Change the handling of OID objects as follows:
16275 - New object identifiers are inserted in objects.txt, following
16276 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16277 - objects.pl is used to process obj_mac.num and create a new
16279 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16282 This is currently kind of a hack, and the perl code in objects.pl
16283 isn't very elegant, but it works as I intended. The simplest way
16284 to check that it worked correctly is to look in obj_dat.h and
16285 check the array nid_objs and make sure the objects haven't moved
16286 around (this is important!). Additions are OK, as well as
16287 consistent name changes.
16291 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16295 * Addition of the command line parameter '-rand file' to 'openssl req'.
16296 The given file adds to whatever has already been seeded into the
16297 random pool through the RANDFILE configuration file option or
16298 environment variable, or the default random state file.
16302 * mkstack.pl now sorts each macro group into lexical order.
16303 Previously the output order depended on the order the files
16304 appeared in the directory, resulting in needless rewriting
16309 * Patches to make OpenSSL compile under Win32 again. Mostly
16310 work arounds for the VC++ problem that it treats func() as
16311 func(void). Also stripped out the parts of mkdef.pl that
16312 added extra typesafe functions: these no longer exist.
16316 * Reorganisation of the stack code. The macros are now all
16317 collected in safestack.h . Each macro is defined in terms of
16318 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16319 DEBUG_SAFESTACK is now handled in terms of function casts,
16320 this has the advantage of retaining type safety without the
16321 use of additional functions. If DEBUG_SAFESTACK is not defined
16322 then the non typesafe macros are used instead. Also modified the
16323 mkstack.pl script to handle the new form. Needs testing to see
16324 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16325 the default if no major problems. Similar behaviour for ASN1_SET_OF
16326 and PKCS12_STACK_OF.
16330 * When some versions of IIS use the 'NET' form of private key the
16331 key derivation algorithm is different. Normally MD5(password) is
16332 used as a 128 bit RC4 key. In the modified case
16333 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16334 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16335 as the old Netscape_RSA functions except they have an additional
16336 'sgckey' parameter which uses the modified algorithm. Also added
16337 an -sgckey command line option to the rsa utility. Thanks to
16338 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16339 algorithm to openssl-dev.
16343 * The evp_local.h macros were using 'c.##kname' which resulted in
16344 invalid expansion on some systems (SCO 5.0.5 for example).
16345 Corrected to 'c.kname'.
16347 *Phillip Porch <root@theporch.com>*
16349 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16350 a STACK of email addresses from a certificate or request, these look
16351 in the subject name and the subject alternative name extensions and
16352 omit any duplicate addresses.
16356 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16357 This makes DSA verification about 2 % faster.
16361 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16362 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16363 plus overhead for 1024 bit moduli).
16364 This makes exponentiations about 0.5 % faster for 1024 bit
16365 exponents (as measured by "openssl speed rsa2048").
16369 * Rename memory handling macros to avoid conflicts with other
16371 Malloc => OPENSSL_malloc
16372 Malloc_locked => OPENSSL_malloc_locked
16373 Realloc => OPENSSL_realloc
16374 Free => OPENSSL_free
16378 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16379 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16383 * CygWin32 support.
16385 *John Jarvie <jjarvie@newsguy.com>*
16387 * The type-safe stack code has been rejigged. It is now only compiled
16388 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16389 by default all type-specific stack functions are "#define"d back to
16390 standard stack functions. This results in more streamlined output
16391 but retains the type-safety checking possibilities of the original
16396 * The STACK code has been cleaned up, and certain type declarations
16397 that didn't make a lot of sense have been brought in line. This has
16398 also involved a cleanup of sorts in safestack.h to more correctly
16399 map type-safe stack functions onto their plain stack counterparts.
16400 This work has also resulted in a variety of "const"ifications of
16401 lots of the code, especially `_cmp` operations which should normally
16402 be prototyped with "const" parameters anyway.
16406 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16407 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16408 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16409 where all of 'md' is used each time the PRNG is used, but 'state'
16410 is used only indexed by a cyclic counter. As entropy may not be
16411 well distributed from the beginning, 'md' is important as a
16412 chaining variable. However, the output function chains only half
16413 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16414 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16415 in all of 'state' being rewritten, with the new values depending
16416 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16420 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16421 the handshake is continued after ssl_verify_cert_chain();
16422 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16423 can lead to 'unexplainable' connection aborts later.
16425 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16427 * Major EVP API cipher revision.
16428 Add hooks for extra EVP features. This allows various cipher
16429 parameters to be set in the EVP interface. Support added for variable
16430 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16431 setting of RC2 and RC5 parameters.
16433 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16436 Remove lots of duplicated code from the EVP library. For example *every*
16437 cipher init() function handles the 'iv' in the same way according to the
16438 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16439 for CFB and OFB modes they zero ctx->num.
16441 New functionality allows removal of S/MIME code RC2 hack.
16443 Most of the routines have the same form and so can be declared in terms
16446 By shifting this to the top level EVP_CipherInit() it can be removed from
16447 all individual ciphers. If the cipher wants to handle IVs or keys
16448 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16451 Change lots of functions like EVP_EncryptUpdate() to now return a
16452 value: although software versions of the algorithms cannot fail
16453 any installed hardware versions can.
16457 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16458 this option is set, tolerate broken clients that send the negotiated
16459 protocol version number instead of the requested protocol version
16464 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16465 i.e. non-zero for export ciphersuites, zero otherwise.
16466 Previous versions had this flag inverted, inconsistent with
16467 rsa_tmp_cb (..._TMP_RSA_CB).
16469 *Bodo Moeller; problem reported by Amit Chopra*
16471 * Add missing DSA library text string. Work around for some IIS
16472 key files with invalid SEQUENCE encoding.
16476 * Add a document (doc/standards.txt) that list all kinds of standards
16477 and so on that are implemented in OpenSSL.
16481 * Enhance c_rehash script. Old version would mishandle certificates
16482 with the same subject name hash and wouldn't handle CRLs at all.
16483 Added -fingerprint option to crl utility, to support new c_rehash
16488 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16492 * Fix for SSL server purpose checking. Server checking was
16493 rejecting certificates which had extended key usage present
16494 but no ssl client purpose.
16496 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16498 * Make PKCS#12 code work with no password. The PKCS#12 spec
16499 is a little unclear about how a blank password is handled.
16500 Since the password in encoded as a BMPString with terminating
16501 double NULL a zero length password would end up as just the
16502 double NULL. However no password at all is different and is
16503 handled differently in the PKCS#12 key generation code. NS
16504 treats a blank password as zero length. MSIE treats it as no
16505 password on export: but it will try both on import. We now do
16506 the same: PKCS12_parse() tries zero length and no password if
16507 the password is set to "" or NULL (NULL is now a valid password:
16508 it wasn't before) as does the pkcs12 application.
16512 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16513 perror when PEM_read_bio_X509_REQ fails, the error message must
16514 be obtained from the error queue.
16518 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16519 it in ERR_remove_state if appropriate, and change ERR_get_state
16520 accordingly to avoid race conditions (this is necessary because
16521 thread_hash is no longer constant once set).
16525 * Bugfix for linux-elf makefile.one.
16529 * RSA_get_default_method() will now cause a default
16530 RSA_METHOD to be chosen if one doesn't exist already.
16531 Previously this was only set during a call to RSA_new()
16532 or RSA_new_method(NULL) meaning it was possible for
16533 RSA_get_default_method() to return NULL.
16537 * Added native name translation to the existing DSO code
16538 that will convert (if the flag to do so is set) filenames
16539 that are sufficiently small and have no path information
16540 into a canonical native form. Eg. "blah" converted to
16541 "libblah.so" or "blah.dll" etc.
16545 * New function ERR_error_string_n(e, buf, len) which is like
16546 ERR_error_string(e, buf), but writes at most 'len' bytes
16547 including the 0 terminator. For ERR_error_string_n, 'buf'
16550 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16552 * CONF library reworked to become more general. A new CONF
16553 configuration file reader "class" is implemented as well as a
16554 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16555 old `CONF_*` functions are still there, but are reimplemented to
16556 work in terms of the new functions. Also, a set of functions
16557 to handle the internal storage of the configuration data is
16558 provided to make it easier to write new configuration file
16559 reader "classes" (I can definitely see something reading a
16560 configuration file in XML format, for example), called `_CONF_*`,
16561 or "the configuration storage API"...
16563 The new configuration file reading functions are:
16565 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16566 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16568 NCONF_default, NCONF_WIN32
16570 NCONF_dump_fp, NCONF_dump_bio
16572 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16573 NCONF_new creates a new CONF object. This works in the same way
16574 as other interfaces in OpenSSL, like the BIO interface.
16575 `NCONF_dump_*` dump the internal storage of the configuration file,
16576 which is useful for debugging. All other functions take the same
16577 arguments as the old `CONF_*` functions with the exception of the
16578 first that must be a `CONF *` instead of a `LHASH *`.
16580 To make it easier to use the new classes with the old `CONF_*` functions,
16581 the function CONF_set_default_method is provided.
16585 * Add '-tls1' option to 'openssl ciphers', which was already
16586 mentioned in the documentation but had not been implemented.
16587 (This option is not yet really useful because even the additional
16588 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16592 * Initial DSO code added into libcrypto for letting OpenSSL (and
16593 OpenSSL-based applications) load shared libraries and bind to
16594 them in a portable way.
16596 *Geoff Thorpe, with contributions from Richard Levitte*
16598 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16600 * Make sure _lrotl and _lrotr are only used with MSVC.
16602 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16603 (the default implementation of RAND_status).
16605 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16606 to '-clrext' (= clear extensions), as intended and documented.
16607 *Bodo Moeller; inconsistency pointed out by Michael Attili
16608 <attili@amaxo.com>*
16610 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16611 was larger than the MD block size.
16613 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16615 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16616 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16617 using the passed key: if the passed key was a private key the result
16618 of X509_print(), for example, would be to print out all the private key
16623 * des_quad_cksum() byte order bug fix.
16624 *Ulf Möller, using the problem description in krb4-0.9.7, where
16625 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16627 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16630 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16632 * For easily testing in shell scripts whether some command
16633 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16634 returns with exit code 0 iff no command of the given name is available.
16635 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16636 the output goes to stdout and nothing is printed to stderr.
16637 Additional arguments are always ignored.
16639 Since for each cipher there is a command of the same name,
16640 the 'no-cipher' compilation switches can be tested this way.
16642 ('openssl no-XXX' is not able to detect pseudo-commands such
16643 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16647 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16651 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16652 is set; it will be thrown away anyway because each handshake creates
16654 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16655 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16656 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16657 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16661 * New s_client option -ign_eof: EOF at stdin is ignored, and
16662 'Q' and 'R' lose their special meanings (quit/renegotiate).
16663 This is part of what -quiet does; unlike -quiet, -ign_eof
16664 does not suppress any output.
16668 * Add compatibility options to the purpose and trust code. The
16669 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16670 accepts a certificate or CA, this was the previous behaviour,
16671 with all the associated security issues.
16673 X509_TRUST_COMPAT is the old trust behaviour: only and
16674 automatically trust self signed roots in certificate store. A
16675 new trust setting X509_TRUST_DEFAULT is used to specify that
16676 a purpose has no associated trust setting and it should instead
16677 use the value in the default purpose.
16681 * Fix the PKCS#8 DSA private key code so it decodes keys again
16682 and fix a memory leak.
16686 * In util/mkerr.pl (which implements 'make errors'), preserve
16687 reason strings from the previous version of the .c file, as
16688 the default to have only downcase letters (and digits) in
16689 automatically generated reasons codes is not always appropriate.
16693 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16694 using strerror. Previously, ERR_reason_error_string() returned
16695 library names as reason strings for SYSerr; but SYSerr is a special
16696 case where small numbers are errno values, not library numbers.
16700 * Add '-dsaparam' option to 'openssl dhparam' application. This
16701 converts DSA parameters into DH parameters. (When creating parameters,
16702 DSA_generate_parameters is used.)
16706 * Include 'length' (recommended exponent length) in C code generated
16707 by 'openssl dhparam -C'.
16711 * The second argument to set_label in perlasm was already being used
16712 so couldn't be used as a "file scope" flag. Moved to third argument
16717 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16718 instead of RAND_bytes for encryption IVs and salts.
16722 * Include RAND_status() into RAND_METHOD instead of implementing
16723 it only for md_rand.c Otherwise replacing the PRNG by calling
16724 RAND_set_rand_method would be impossible.
16728 * Don't let DSA_generate_key() enter an infinite loop if the random
16729 number generation fails.
16733 * New 'rand' application for creating pseudo-random output.
16737 * Added configuration support for Linux/IA64
16739 *Rolf Haberrecker <rolf@suse.de>*
16741 * Assembler module support for Mingw32.
16745 * Shared library support for HPUX (in shlib/).
16747 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16749 * Shared library support for Solaris gcc.
16751 *Lutz Behnke <behnke@trustcenter.de>*
16753 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16755 * PKCS7_encrypt() was adding text MIME headers twice because they
16756 were added manually and by SMIME_crlf_copy().
16760 * In bntest.c don't call BN_rand with zero bits argument.
16762 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16764 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16765 case was implemented. This caused BN_div_recp() to fail occasionally.
16769 * Add an optional second argument to the set_label() in the perl
16770 assembly language builder. If this argument exists and is set
16771 to 1 it signals that the assembler should use a symbol whose
16772 scope is the entire file, not just the current function. This
16773 is needed with MASM which uses the format label:: for this scope.
16775 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16777 * Change the ASN1 types so they are typedefs by default. Before
16778 almost all types were #define'd to ASN1_STRING which was causing
16779 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16784 * Change names of new functions to the new get1/get0 naming
16785 convention: After 'get1', the caller owns a reference count
16786 and has to call `..._free`; 'get0' returns a pointer to some
16787 data structure without incrementing reference counters.
16788 (Some of the existing 'get' functions increment a reference
16789 counter, some don't.)
16790 Similarly, 'set1' and 'add1' functions increase reference
16791 counters or duplicate objects.
16795 * Allow for the possibility of temp RSA key generation failure:
16796 the code used to assume it always worked and crashed on failure.
16800 * Fix potential buffer overrun problem in BIO_printf().
16801 *Ulf Möller, using public domain code by Patrick Powell; problem
16802 pointed out by David Sacerdote <das33@cornell.edu>*
16804 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16805 RAND_egd() and RAND_status(). In the command line application,
16806 the EGD socket can be specified like a seed file using RANDFILE
16811 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16812 Some CAs (e.g. Verisign) distribute certificates in this form.
16816 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16817 list to exclude them. This means that no special compilation option
16818 is needed to use anonymous DH: it just needs to be included in the
16823 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16824 EVP_MD_type. The old functionality is available in a new macro called
16825 EVP_MD_md(). Change code that uses it and update docs.
16829 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16830 where the `void *` argument is replaced by a function pointer argument.
16831 Previously `void *` was abused to point to functions, which works on
16832 many platforms, but is not correct. As these functions are usually
16833 called by macros defined in OpenSSL header files, most source code
16834 should work without changes.
16838 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16839 sections with information on -D... compiler switches used for
16840 compiling the library so that applications can see them. To enable
16841 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16842 must be defined. E.g.,
16843 #define OPENSSL_ALGORITHM_DEFINES
16844 #include <openssl/opensslconf.h>
16845 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16847 *Richard Levitte, Ulf and Bodo Möller*
16849 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16854 * Change the 'other' type in certificate aux info to a STACK_OF
16855 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16856 the required ASN1 format: arbitrary types determined by an OID.
16860 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16861 argument to 'req'. This is not because the function is newer or
16862 better than others it just uses the work 'NEW' in the certificate
16863 request header lines. Some software needs this.
16867 * Reorganise password command line arguments: now passwords can be
16868 obtained from various sources. Delete the PEM_cb function and make
16869 it the default behaviour: i.e. if the callback is NULL and the
16870 usrdata argument is not NULL interpret it as a null terminated pass
16871 phrase. If usrdata and the callback are NULL then the pass phrase
16872 is prompted for as usual.
16876 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16877 the support is automatically enabled. The resulting binaries will
16878 autodetect the card and use it if present.
16880 *Ben Laurie and Compaq Inc.*
16882 * Work around for Netscape hang bug. This sends certificate request
16883 and server done in one record. Since this is perfectly legal in the
16884 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16885 the bugs/SSLv3 entry for more info.
16889 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16893 * Add -rand argument to smime and pkcs12 applications and read/write
16898 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16902 * Add command line password options to the remaining applications.
16906 * Bug fix for BN_div_recp() for numerators with an even number of
16911 * More tests in bntest.c, and changed test_bn output.
16915 * ./config recognizes MacOS X now.
16919 * Bug fix for BN_div() when the first words of num and divisor are
16920 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16924 * Add support for various broken PKCS#8 formats, and command line
16925 options to produce them.
16929 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16930 get temporary BIGNUMs from a BN_CTX.
16934 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16939 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16940 include a #define from the old name to the new. The original intent
16941 was that statically linked binaries could for example just call
16942 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16943 link with digests. This never worked because SSLeay_add_all_digests()
16944 and SSLeay_add_all_ciphers() were in the same source file so calling
16945 one would link with the other. They are now in separate source files.
16949 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16953 * Use a less unusual form of the Miller-Rabin primality test (it used
16954 a binary algorithm for exponentiation integrated into the Miller-Rabin
16955 loop, our standard modexp algorithms are faster).
16959 * Support for the EBCDIC character set completed.
16961 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16963 * Source code cleanups: use const where appropriate, eliminate casts,
16964 use `void *` instead of `char *` in lhash.
16968 * Bugfix: ssl3_send_server_key_exchange was not restartable
16969 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16970 this the server could overwrite ephemeral keys that the client
16975 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16976 using 50 iterations of the Rabin-Miller test.
16978 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16979 iterations of the Rabin-Miller test as required by the appendix
16980 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16981 As BN_is_prime_fasttest includes trial division, DSA parameter
16982 generation becomes much faster.
16984 This implies a change for the callback functions in DSA_is_prime
16985 and DSA_generate_parameters: The callback function is called once
16986 for each positive witness in the Rabin-Miller test, not just
16987 occasionally in the inner loop; and the parameters to the
16988 callback function now provide an iteration count for the outer
16989 loop rather than for the current invocation of the inner loop.
16990 DSA_generate_parameters additionally can call the callback
16991 function with an 'iteration count' of -1, meaning that a
16992 candidate has passed the trial division test (when q is generated
16993 from an application-provided seed, trial division is skipped).
16997 * New function BN_is_prime_fasttest that optionally does trial
16998 division before starting the Rabin-Miller test and has
16999 an additional BN_CTX * argument (whereas BN_is_prime always
17000 has to allocate at least one BN_CTX).
17001 'callback(1, -1, cb_arg)' is called when a number has passed the
17002 trial division stage.
17006 * Fix for bug in CRL encoding. The validity dates weren't being handled
17011 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
17015 * New function BN_pseudo_rand().
17019 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
17020 bignum version of BN_from_montgomery() with the working code from
17021 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
17026 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
17027 made it impossible to use the same SSL_SESSION data structure in
17028 SSL2 clients in multiple threads.
17032 * The return value of RAND_load_file() no longer counts bytes obtained
17033 by stat(). RAND_load_file(..., -1) is new and uses the complete file
17034 to seed the PRNG (previously an explicit byte count was required).
17036 *Ulf Möller, Bodo Möller*
17038 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
17039 used `char *` instead of `void *` and had casts all over the place.
17043 * Make BN_generate_prime() return NULL on error if ret!=NULL.
17047 * Retain source code compatibility for BN_prime_checks macro:
17048 BN_is_prime(..., BN_prime_checks, ...) now uses
17049 BN_prime_checks_for_size to determine the appropriate number of
17050 Rabin-Miller iterations.
17054 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
17055 DH_CHECK_P_NOT_SAFE_PRIME.
17056 (Check if this is true? OpenPGP calls them "strong".)
17060 * Merge the functionality of "dh" and "gendh" programs into a new program
17061 "dhparam". The old programs are retained for now but will handle DH keys
17062 (instead of parameters) in future.
17066 * Make the ciphers, s_server and s_client programs check the return values
17067 when a new cipher list is set.
17071 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17072 ciphers. Before when the 56bit ciphers were enabled the sorting was
17075 The syntax for the cipher sorting has been extended to support sorting by
17076 cipher-strength (using the strength_bits hard coded in the tables).
17077 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17079 Fix a bug in the cipher-command parser: when supplying a cipher command
17080 string with an "undefined" symbol (neither command nor alphanumeric
17081 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17082 an error is flagged.
17084 Due to the strength-sorting extension, the code of the
17085 ssl_create_cipher_list() function was completely rearranged. I hope that
17086 the readability was also increased :-)
17088 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17090 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17091 for the first serial number and places 2 in the serial number file. This
17092 avoids problems when the root CA is created with serial number zero and
17093 the first user certificate has the same issuer name and serial number
17098 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17099 the new code. Add documentation for this stuff.
17103 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17104 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17105 structures and behave in an analogous way to the X509v3 functions:
17106 they shouldn't be called directly but wrapper functions should be used
17109 So we also now have some wrapper functions that call the X509at functions
17110 when passed certificate requests. (TO DO: similar things can be done with
17111 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17112 things. Some of these need some d2i or i2d and print functionality
17113 because they handle more complex structures.)
17117 * Add missing #ifndefs that caused missing symbols when building libssl
17118 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17119 NO_RSA in `ssl/s2*.c`.
17121 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17123 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17124 has a return value which indicates the quality of the random data
17125 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17126 error queue. New function RAND_pseudo_bytes() generates output that is
17127 guaranteed to be unique but not unpredictable. RAND_add is like
17128 RAND_seed, but takes an extra argument for an entropy estimate
17129 (RAND_seed always assumes full entropy).
17133 * Do more iterations of Rabin-Miller probable prime test (specifically,
17134 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17135 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17136 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17137 false-positive rate of at most 2^-80 for random input.
17141 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17145 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17146 in the 0.9.5 release), this returns the chain
17147 from an X509_CTX structure with a dup of the stack and all
17148 the X509 reference counts upped: so the stack will exist
17149 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17152 Also make SSL_SESSION_print() print out the verify return
17157 * Add manpage for the pkcs12 command. Also change the default
17158 behaviour so MAC iteration counts are used unless the new
17159 -nomaciter option is used. This improves file security and
17160 only older versions of MSIE (4.0 for example) need it.
17164 * Honor the no-xxx Configure options when creating .DEF files.
17168 * Add PKCS#10 attributes to field table: challengePassword,
17169 unstructuredName and unstructuredAddress. These are taken from
17170 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17171 international characters are used.
17173 More changes to X509_ATTRIBUTE code: allow the setting of types
17174 based on strings. Remove the 'loc' parameter when adding
17175 attributes because these will be a SET OF encoding which is sorted
17180 * Initial changes to the 'req' utility to allow request generation
17181 automation. This will allow an application to just generate a template
17182 file containing all the field values and have req construct the
17185 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17186 used all over the place including certificate requests and PKCS#7
17187 structures. They are currently handled manually where necessary with
17188 some primitive wrappers for PKCS#7. The new functions behave in a
17189 manner analogous to the X509 extension functions: they allow
17190 attributes to be looked up by NID and added.
17192 Later something similar to the X509V3 code would be desirable to
17193 automatically handle the encoding, decoding and printing of the
17194 more complex types. The string types like challengePassword can
17195 be handled by the string table functions.
17197 Also modified the multi byte string table handling. Now there is
17198 a 'global mask' which masks out certain types. The table itself
17199 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17200 is useful when for example there is only one permissible type
17201 (as in countryName) and using the mask might result in no valid
17206 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17207 SSL_get_peer_finished to allow applications to obtain the latest
17208 Finished messages sent to the peer or expected from the peer,
17209 respectively. (SSL_get_peer_finished is usually the Finished message
17210 actually received from the peer, otherwise the protocol will be aborted.)
17212 As the Finished message are message digests of the complete handshake
17213 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17214 be used for external authentication procedures when the authentication
17215 provided by SSL/TLS is not desired or is not enough.
17219 * Enhanced support for Alpha Linux is added. Now ./config checks if
17220 the host supports BWX extension and if Compaq C is present on the
17221 $PATH. Just exploiting of the BWX extension results in 20-30%
17222 performance kick for some algorithms, e.g. DES and RC4 to mention
17223 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17228 * Add support for MS "fast SGC". This is arguably a violation of the
17229 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17230 weak crypto and after checking the certificate is SGC a second one
17231 with strong crypto. MS SGC stops the first handshake after receiving
17232 the server certificate message and sends a second client hello. Since
17233 a server will typically do all the time consuming operations before
17234 expecting any further messages from the client (server key exchange
17235 is the most expensive) there is little difference between the two.
17237 To get OpenSSL to support MS SGC we have to permit a second client
17238 hello message after we have sent server done. In addition we have to
17239 reset the MAC if we do get this second client hello.
17243 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17244 if a DER encoded private key is RSA or DSA traditional format. Changed
17245 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17246 format DER encoded private key. Newer code should use PKCS#8 format which
17247 has the key type encoded in the ASN1 structure. Added DER private key
17248 support to pkcs8 application.
17252 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17253 ciphersuites has been selected (as required by the SSL 3/TLS 1
17254 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17255 is set, we interpret this as a request to violate the specification
17256 (the worst that can happen is a handshake failure, and 'correct'
17257 behaviour would result in a handshake failure anyway).
17261 * In SSL_CTX_add_session, take into account that there might be multiple
17262 SSL_SESSION structures with the same session ID (e.g. when two threads
17263 concurrently obtain them from an external cache).
17264 The internal cache can handle only one SSL_SESSION with a given ID,
17265 so if there's a conflict, we now throw out the old one to achieve
17270 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17271 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17272 some routines that use cipher OIDs: some ciphers do not have OIDs
17273 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17278 * Simplify the trust setting structure and code. Now we just have
17279 two sequences of OIDs for trusted and rejected settings. These will
17280 typically have values the same as the extended key usage extension
17281 and any application specific purposes.
17283 The trust checking code now has a default behaviour: it will just
17284 check for an object with the same NID as the passed id. Functions can
17285 be provided to override either the default behaviour or the behaviour
17286 for a given id. SSL client, server and email already have functions
17287 in place for compatibility: they check the NID and also return "trusted"
17288 if the certificate is self signed.
17292 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17293 traditional format into an EVP_PKEY structure.
17297 * Add a password callback function PEM_cb() which either prompts for
17298 a password if usr_data is NULL or otherwise assumes it is a null
17299 terminated password. Allow passwords to be passed on command line
17300 environment or config files in a few more utilities.
17304 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17305 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17306 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17307 Update documentation.
17311 * Support for ASN1 "NULL" type. This could be handled before by using
17312 ASN1_TYPE but there wasn't any function that would try to read a NULL
17313 and produce an error if it couldn't. For compatibility we also have
17314 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17315 don't allocate anything because they don't need to.
17319 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17322 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17324 * Rebuild of the memory allocation routines used by OpenSSL code and
17325 possibly others as well. The purpose is to make an interface that
17326 provide hooks so anyone can build a separate set of allocation and
17327 deallocation routines to be used by OpenSSL, for example memory
17328 pool implementations, or something else, which was previously hard
17329 since Malloc(), Realloc() and Free() were defined as macros having
17330 the values malloc, realloc and free, respectively (except for Win32
17331 compilations). The same is provided for memory debugging code.
17332 OpenSSL already comes with functionality to find memory leaks, but
17333 this gives people a chance to debug other memory problems.
17335 With these changes, a new set of functions and macros have appeared:
17337 CRYPTO_set_mem_debug_functions() [F]
17338 CRYPTO_get_mem_debug_functions() [F]
17339 CRYPTO_dbg_set_options() [F]
17340 CRYPTO_dbg_get_options() [F]
17341 CRYPTO_malloc_debug_init() [M]
17343 The memory debug functions are NULL by default, unless the library
17344 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17345 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17346 gives the standard debugging functions that come with OpenSSL) or
17347 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17348 provided by the library user) must be used. When the standard
17349 debugging functions are used, CRYPTO_dbg_set_options can be used to
17350 request additional information:
17351 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17352 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17354 Also, things like CRYPTO_set_mem_functions will always give the
17355 expected result (the new set of functions is used for allocation
17356 and deallocation) at all times, regardless of platform and compiler
17359 To finish it up, some functions that were never use in any other
17360 way than through macros have a new API and new semantic:
17362 CRYPTO_dbg_malloc()
17363 CRYPTO_dbg_realloc()
17366 All macros of value have retained their old syntax.
17368 *Richard Levitte and Bodo Moeller*
17370 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17371 ordering of SMIMECapabilities wasn't in "strength order" and there
17372 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17377 * Some ASN1 types with illegal zero length encoding (INTEGER,
17378 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17380 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17382 * Merge in my S/MIME library for OpenSSL. This provides a simple
17383 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17384 functionality to handle multipart/signed properly) and a utility
17385 called 'smime' to call all this stuff. This is based on code I
17386 originally wrote for Celo who have kindly allowed it to be
17387 included in OpenSSL.
17391 * Add variants des_set_key_checked and des_set_key_unchecked of
17392 des_set_key (aka des_key_sched). Global variable des_check_key
17393 decides which of these is called by des_set_key; this way
17394 des_check_key behaves as it always did, but applications and
17395 the library itself, which was buggy for des_check_key == 1,
17396 have a cleaner way to pick the version they need.
17400 * New function PKCS12_newpass() which changes the password of a
17405 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17406 dynamic mix. In both cases the ids can be used as an index into the
17407 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17408 functions so they accept a list of the field values and the
17409 application doesn't need to directly manipulate the X509_TRUST
17414 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17419 * Modify the way the V3 extension code looks up extensions. This now
17420 works in a similar way to the object code: we have some "standard"
17421 extensions in a static table which is searched with OBJ_bsearch()
17422 and the application can add dynamic ones if needed. The file
17423 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17424 updated whenever a new extension is added to the core code and kept
17425 in ext_nid order. There is a simple program 'tabtest.c' which checks
17426 this. New extensions are not added too often so this file can readily
17427 be maintained manually.
17429 There are two big advantages in doing things this way. The extensions
17430 can be looked up immediately and no longer need to be "added" using
17431 X509V3_add_standard_extensions(): this function now does nothing.
17432 Side note: I get *lots* of email saying the extension code doesn't
17433 work because people forget to call this function.
17434 Also no dynamic allocation is done unless new extensions are added:
17435 so if we don't add custom extensions there is no need to call
17436 X509V3_EXT_cleanup().
17440 * Modify enc utility's salting as follows: make salting the default. Add a
17441 magic header, so unsalted files fail gracefully instead of just decrypting
17442 to garbage. This is because not salting is a big security hole, so people
17443 should be discouraged from doing it.
17447 * Fixes and enhancements to the 'x509' utility. It allowed a message
17448 digest to be passed on the command line but it only used this
17449 parameter when signing a certificate. Modified so all relevant
17450 operations are affected by the digest parameter including the
17451 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17452 DSA key was used because it didn't fix the digest.
17456 * Initial certificate chain verify code. Currently tests the untrusted
17457 certificates for consistency with the verify purpose (which is set
17458 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17460 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17461 this is because it will reject chains with invalid extensions whereas
17462 every previous version of OpenSSL and SSLeay made no checks at all.
17464 Trust code: checks the root CA for the relevant trust settings. Trust
17465 settings have an initial value consistent with the verify purpose: e.g.
17466 if the verify purpose is for SSL client use it expects the CA to be
17467 trusted for SSL client use. However the default value can be changed to
17468 permit custom trust settings: one example of this would be to only trust
17469 certificates from a specific "secure" set of CAs.
17471 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17472 which should be used for version portability: especially since the
17473 verify structure is likely to change more often now.
17475 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17476 to set them. If not set then assume SSL clients will verify SSL servers
17479 Two new options to the verify program: -untrusted allows a set of
17480 untrusted certificates to be passed in and -purpose which sets the
17481 intended purpose of the certificate. If a purpose is set then the
17482 new chain verify code is used to check extension consistency.
17486 * Support for the authority information access extension.
17490 * Modify RSA and DSA PEM read routines to transparently handle
17491 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17492 public keys in a format compatible with certificate
17493 SubjectPublicKeyInfo structures. Unfortunately there were already
17494 functions called *_PublicKey_* which used various odd formats so
17495 these are retained for compatibility: however the DSA variants were
17496 never in a public release so they have been deleted. Changed dsa/rsa
17497 utilities to handle the new format: note no releases ever handled public
17498 keys so we should be OK.
17500 The primary motivation for this change is to avoid the same fiasco
17501 that dogs private keys: there are several incompatible private key
17502 formats some of which are standard and some OpenSSL specific and
17503 require various evil hacks to allow partial transparent handling and
17504 even then it doesn't work with DER formats. Given the option anything
17505 other than PKCS#8 should be dumped: but the other formats have to
17506 stay in the name of compatibility.
17508 With public keys and the benefit of hindsight one standard format
17509 is used which works with EVP_PKEY, RSA or DSA structures: though
17510 it clearly returns an error if you try to read the wrong kind of key.
17512 Added a -pubkey option to the 'x509' utility to output the public key.
17513 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17514 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17515 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17516 that do the same as the `EVP_PKEY_assign_*()` except they up the
17517 reference count of the added key (they don't "swallow" the
17522 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17523 CRLs would fail if the file contained no certificates or no CRLs:
17524 added a new function to read in both types and return the number
17525 read: this means that if none are read it will be an error. The
17526 DER versions of the certificate and CRL reader would always fail
17527 because it isn't possible to mix certificates and CRLs in DER format
17528 without choking one or the other routine. Changed this to just read
17529 a certificate: this is the best we can do. Also modified the code
17530 in `apps/verify.c` to take notice of return codes: it was previously
17531 attempting to read in certificates from NULL pointers and ignoring
17532 any errors: this is one reason why the cert and CRL reader seemed
17533 to work. It doesn't check return codes from the default certificate
17534 routines: these may well fail if the certificates aren't installed.
17538 * Code to support otherName option in GeneralName.
17542 * First update to verify code. Change the verify utility
17543 so it warns if it is passed a self signed certificate:
17544 for consistency with the normal behaviour. X509_verify
17545 has been modified to it will now verify a self signed
17546 certificate if *exactly* the same certificate appears
17547 in the store: it was previously impossible to trust a
17548 single self signed certificate. This means that:
17549 openssl verify ss.pem
17550 now gives a warning about a self signed certificate but
17551 openssl verify -CAfile ss.pem ss.pem
17556 * For servers, store verify_result in SSL_SESSION data structure
17557 (and add it to external session representation).
17558 This is needed when client certificate verifications fails,
17559 but an application-provided verification callback (set by
17560 SSL_CTX_set_cert_verify_callback) allows accepting the session
17561 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17562 but returns 1): When the session is reused, we have to set
17563 ssl->verify_result to the appropriate error code to avoid
17566 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17568 * Fix a bug in the new PKCS#7 code: it didn't consider the
17569 case in PKCS7_dataInit() where the signed PKCS7 structure
17570 didn't contain any existing data because it was being created.
17572 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17574 * Add a salt to the key derivation routines in enc.c. This
17575 forms the first 8 bytes of the encrypted file. Also add a
17576 -S option to allow a salt to be input on the command line.
17580 * New function X509_cmp(). Oddly enough there wasn't a function
17581 to compare two certificates. We do this by working out the SHA1
17582 hash and comparing that. X509_cmp() will be needed by the trust
17587 * SSL_get1_session() is like SSL_get_session(), but increments
17588 the reference count in the SSL_SESSION returned.
17590 *Geoff Thorpe <geoff@eu.c2.net>*
17592 * Fix for 'req': it was adding a null to request attributes.
17593 Also change the X509_LOOKUP and X509_INFO code to handle
17594 certificate auxiliary information.
17598 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17603 * Add the possibility to add extra information to the memory leak
17604 detecting output, to form tracebacks, showing from where each
17605 allocation was originated: CRYPTO_push_info("constant string") adds
17606 the string plus current file name and line number to a per-thread
17607 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17608 is like calling CYRPTO_pop_info() until the stack is empty.
17609 Also updated memory leak detection code to be multi-thread-safe.
17613 * Add options -text and -noout to pkcs7 utility and delete the
17614 encryption options which never did anything. Update docs.
17618 * Add options to some of the utilities to allow the pass phrase
17619 to be included on either the command line (not recommended on
17620 OSes like Unix) or read from the environment. Update the
17621 manpages and fix a few bugs.
17625 * Add a few manpages for some of the openssl commands.
17629 * Fix the -revoke option in ca. It was freeing up memory twice,
17630 leaking and not finding already revoked certificates.
17634 * Extensive changes to support certificate auxiliary information.
17635 This involves the use of X509_CERT_AUX structure and X509_AUX
17636 functions. An X509_AUX function such as PEM_read_X509_AUX()
17637 can still read in a certificate file in the usual way but it
17638 will also read in any additional "auxiliary information". By
17639 doing things this way a fair degree of compatibility can be
17640 retained: existing certificates can have this information added
17641 using the new 'x509' options.
17643 Current auxiliary information includes an "alias" and some trust
17644 settings. The trust settings will ultimately be used in enhanced
17645 certificate chain verification routines: currently a certificate
17646 can only be trusted if it is self signed and then it is trusted
17651 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17652 The problem was that one of the replacement routines had not been working
17653 since SSLeay releases. For now the offending routine has been replaced
17654 with non-optimised assembler. Even so, this now gives around 95%
17655 performance improvement for 1024 bit RSA signs.
17659 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17660 handling. Most clients have the effective key size in bits equal to
17661 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17662 A few however don't do this and instead use the size of the decrypted key
17663 to determine the RC2 key length and the AlgorithmIdentifier to determine
17664 the effective key length. In this case the effective key length can still
17665 be 40 bits but the key length can be 168 bits for example. This is fixed
17666 by manually forcing an RC2 key into the EVP_PKEY structure because the
17667 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17668 the key length and effective key length are equal.
17672 * Add a bunch of functions that should simplify the creation of
17673 X509_NAME structures. Now you should be able to do:
17674 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17675 and have it automatically work out the correct field type and fill in
17676 the structures. The more adventurous can try:
17677 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17678 and it will (hopefully) work out the correct multibyte encoding.
17682 * Change the 'req' utility to use the new field handling and multibyte
17683 copy routines. Before the DN field creation was handled in an ad hoc
17684 way in req, ca, and x509 which was rather broken and didn't support
17685 BMPStrings or UTF8Strings. Since some software doesn't implement
17686 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17687 using the dirstring_type option. See the new comment in the default
17688 openssl.cnf for more info.
17692 * Make crypto/rand/md_rand.c more robust:
17693 - Assure unique random numbers after fork().
17694 - Make sure that concurrent threads access the global counter and
17695 md serializably so that we never lose entropy in them
17696 or use exactly the same state in multiple threads.
17697 Access to the large state is not always serializable because
17698 the additional locking could be a performance killer, and
17699 md should be large enough anyway.
17703 * New file `apps/app_rand.c` with commonly needed functionality
17704 for handling the random seed file.
17706 Use the random seed file in some applications that previously did not:
17708 dsaparam -genkey (which also ignored its '-rand' option),
17711 x509 (when signing).
17712 Except on systems with /dev/urandom, it is crucial to have a random
17713 seed file at least for key creation, DSA signing, and for DH exchanges;
17714 for RSA signatures we could do without one.
17716 gendh and gendsa (unlike genrsa) used to read only the first byte
17717 of each file listed in the '-rand' option. The function as previously
17718 found in genrsa is now in app_rand.c and is used by all programs
17719 that support '-rand'.
17723 * In RAND_write_file, use mode 0600 for creating files;
17724 don't just chmod when it may be too late.
17728 * Report an error from X509_STORE_load_locations
17729 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17733 * New function ASN1_mbstring_copy() this copies a string in either
17734 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17735 into an ASN1_STRING type. A mask of permissible types is passed
17736 and it chooses the "minimal" type to use or an error if not type
17741 * Add function equivalents to the various macros in asn1.h. The old
17742 macros are retained with an `M_` prefix. Code inside the library can
17743 use the `M_` macros. External code (including the openssl utility)
17744 should *NOT* in order to be "shared library friendly".
17748 * Add various functions that can check a certificate's extensions
17749 to see if it usable for various purposes such as SSL client,
17750 server or S/MIME and CAs of these types. This is currently
17751 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17752 verification. Also added a -purpose flag to x509 utility to
17753 print out all the purposes.
17757 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17762 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17763 for, obtain and decode and extension and obtain its critical flag.
17764 This allows all the necessary extension code to be handled in a
17765 single function call.
17769 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17770 platforms. See crypto/rc4/rc4_enc.c for further details.
17774 * New -noout option to asn1parse. This causes no output to be produced
17775 its main use is when combined with -strparse and -out to extract data
17776 from a file (which may not be in ASN.1 format).
17780 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17781 when producing the local key id.
17783 *Richard Levitte <levitte@stacken.kth.se>*
17785 * New option -dhparam in s_server. This allows a DH parameter file to be
17786 stated explicitly. If it is not stated then it tries the first server
17787 certificate file. The previous behaviour hard coded the filename
17792 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17793 a public key to be input or output. For example:
17794 openssl rsa -in key.pem -pubout -out pubkey.pem
17795 Also added necessary DSA public key functions to handle this.
17799 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17800 in the message. This was handled by allowing
17801 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17803 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17805 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17806 to the end of the strings whereas this didn't. This would cause problems
17807 if strings read with d2i_ASN1_bytes() were later modified.
17809 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17811 * Fix for base64 decode bug. When a base64 bio reads only one line of
17812 data and it contains EOF it will end up returning an error. This is
17813 caused by input 46 bytes long. The cause is due to the way base64
17814 BIOs find the start of base64 encoded data. They do this by trying a
17815 trial decode on each line until they find one that works. When they
17816 do a flag is set and it starts again knowing it can pass all the
17817 data directly through the decoder. Unfortunately it doesn't reset
17818 the context it uses. This means that if EOF is reached an attempt
17819 is made to pass two EOFs through the context and this causes the
17820 resulting error. This can also cause other problems as well. As is
17821 usual with these problems it takes *ages* to find and the fix is
17822 trivial: move one line.
17824 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17826 * Ugly workaround to get s_client and s_server working under Windows. The
17827 old code wouldn't work because it needed to select() on sockets and the
17828 tty (for keypresses and to see if data could be written). Win32 only
17829 supports select() on sockets so we select() with a 1s timeout on the
17830 sockets and then see if any characters are waiting to be read, if none
17831 are present then we retry, we also assume we can always write data to
17832 the tty. This isn't nice because the code then blocks until we've
17833 received a complete line of data and it is effectively polling the
17834 keyboard at 1s intervals: however it's quite a bit better than not
17835 working at all :-) A dedicated Windows application might handle this
17836 with an event loop for example.
17840 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17841 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17842 will be called when RSA_sign() and RSA_verify() are used. This is useful
17843 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17844 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17845 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17846 This necessitated the support of an extra signature type NID_md5_sha1
17847 for SSL signatures and modifications to the SSL library to use it instead
17848 of calling RSA_public_decrypt() and RSA_private_encrypt().
17852 * Add new -verify -CAfile and -CApath options to the crl program, these
17853 will lookup a CRL issuers certificate and verify the signature in a
17854 similar way to the verify program. Tidy up the crl program so it
17855 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17856 less strict. It will now permit CRL extensions even if it is not
17857 a V2 CRL: this will allow it to tolerate some broken CRLs.
17861 * Initialize all non-automatic variables each time one of the openssl
17862 sub-programs is started (this is necessary as they may be started
17863 multiple times from the "OpenSSL>" prompt).
17865 *Lennart Bang, Bodo Moeller*
17867 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17868 removing all other RSA functionality (this is what NO_RSA does). This
17869 is so (for example) those in the US can disable those operations covered
17870 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17875 * Non-copying interface to BIO pairs.
17876 (still largely untested)
17880 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17881 ASCII string. This was handled independently in various places before.
17885 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17886 UTF8 strings a character at a time.
17890 * Use client_version from client hello to select the protocol
17891 (s23_srvr.c) and for RSA client key exchange verification
17892 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17896 * Add various utility functions to handle SPKACs, these were previously
17897 handled by poking round in the structure internals. Added new function
17898 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17899 print, verify and generate SPKACs. Based on an original idea from
17900 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17904 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17908 * Allow the config file extension section to be overwritten on the
17909 command line. Based on an original idea from Massimiliano Pala
17910 <madwolf@comune.modena.it>. The new option is called -extensions
17911 and can be applied to ca, req and x509. Also -reqexts to override
17912 the request extensions in req and -crlexts to override the crl extensions
17917 * Add new feature to the SPKAC handling in ca. Now you can include
17918 the same field multiple times by preceding it by "XXXX." for example:
17921 this is the same syntax as used in the req config file.
17925 * Allow certificate extensions to be added to certificate requests. These
17926 are specified in a 'req_extensions' option of the req section of the
17927 config file. They can be printed out with the -text option to req but
17928 are otherwise ignored at present.
17932 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17933 data read consists of only the final block it would not decrypted because
17934 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17935 A misplaced 'break' also meant the decrypted final block might not be
17936 copied until the next read.
17940 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17941 a few extra parameters to the DH structure: these will be useful if
17942 for example we want the value of 'q' or implement X9.42 DH.
17946 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17947 provides hooks that allow the default DSA functions or functions on a
17948 "per key" basis to be replaced. This allows hardware acceleration and
17949 hardware key storage to be handled without major modification to the
17950 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17951 associated functions.
17955 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17956 as "read only": it can't be written to and the buffer it points to will
17957 not be freed. Reading from a read only BIO is much more efficient than
17958 a normal memory BIO. This was added because there are several times when
17959 an area of memory needs to be read from a BIO. The previous method was
17960 to create a memory BIO and write the data to it, this results in two
17961 copies of the data and an O(n^2) reading algorithm. There is a new
17962 function BIO_new_mem_buf() which creates a read only memory BIO from
17963 an area of memory. Also modified the PKCS#7 routines to use read only
17968 * Bugfix: ssl23_get_client_hello did not work properly when called in
17969 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17970 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17971 but a retry condition occurred while trying to read the rest.
17975 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17976 NID_pkcs7_encrypted by default: this was wrong since this should almost
17977 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17978 the encrypted data type: this is a more sensible place to put it and it
17979 allows the PKCS#12 code to be tidied up that duplicated this
17984 * Changed obj_dat.pl script so it takes its input and output files on
17985 the command line. This should avoid shell escape redirection problems
17990 * Initial support for certificate extension requests, these are included
17991 in things like Xenroll certificate requests. Included functions to allow
17992 extensions to be obtained and added.
17996 * -crlf option to s_client and s_server for sending newlines as
17997 CRLF (as required by many protocols).
18001 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
18003 * Install libRSAglue.a when OpenSSL is built with RSAref.
18005 *Ralf S. Engelschall*
18007 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
18009 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
18011 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
18016 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
18017 DH parameters/keys (q is lost during that conversion, but the resulting
18018 DH parameters contain its length).
18020 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
18021 much faster than DH_generate_parameters (which creates parameters
18022 where `p = 2*q + 1`), and also the smaller q makes DH computations
18023 much more efficient (160-bit exponentiation instead of 1024-bit
18024 exponentiation); so this provides a convenient way to support DHE
18025 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
18026 utter importance to use
18027 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18029 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18030 when such DH parameters are used, because otherwise small subgroup
18031 attacks may become possible!
18035 * Avoid memory leak in i2d_DHparams.
18039 * Allow the -k option to be used more than once in the enc program:
18040 this allows the same encrypted message to be read by multiple recipients.
18044 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
18045 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
18046 it will always use the numerical form of the OID, even if it has a short
18051 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
18052 method only got called if p,q,dmp1,dmq1,iqmp components were present,
18053 otherwise bn_mod_exp was called. In the case of hardware keys for example
18054 no private key components need be present and it might store extra data
18055 in the RSA structure, which cannot be accessed from bn_mod_exp.
18056 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
18057 private key operations.
18061 * Added support for SPARC Linux.
18065 * pem_password_cb function type incompatibly changed from
18066 typedef int pem_password_cb(char *buf, int size, int rwflag);
18068 ....(char *buf, int size, int rwflag, void *userdata);
18069 so that applications can pass data to their callbacks:
18070 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18071 additional void * argument, which is just handed through whenever
18072 the password callback is called.
18074 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18076 New function SSL_CTX_set_default_passwd_cb_userdata.
18078 Compatibility note: As many C implementations push function arguments
18079 onto the stack in reverse order, the new library version is likely to
18080 interoperate with programs that have been compiled with the old
18081 pem_password_cb definition (PEM_whatever takes some data that
18082 happens to be on the stack as its last argument, and the callback
18083 just ignores this garbage); but there is no guarantee whatsoever that
18086 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18087 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18088 problems not only on Windows, but also on some Unix platforms.
18089 To avoid problematic command lines, these definitions are now in an
18090 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18091 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18095 * MIPS III/IV assembler module is reimplemented.
18099 * More DES library cleanups: remove references to srand/rand and
18100 delete an unused file.
18104 * Add support for the free Netwide assembler (NASM) under Win32,
18105 since not many people have MASM (ml) and it can be hard to obtain.
18106 This is currently experimental but it seems to work OK and pass all
18107 the tests. Check out INSTALL.W32 for info.
18111 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18112 without temporary keys kept an extra copy of the server key,
18113 and connections with temporary keys did not free everything in case
18118 * New function RSA_check_key and new openssl rsa option -check
18119 for verifying the consistency of RSA keys.
18121 *Ulf Moeller, Bodo Moeller*
18123 * Various changes to make Win32 compile work:
18124 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18125 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18126 comparison" warnings.
18127 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18131 * Add a debugging option to PKCS#5 v2 key generation function: when
18132 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18133 derived keys are printed to stderr.
18137 * Copy the flags in ASN1_STRING_dup().
18139 *Roman E. Pavlov <pre@mo.msk.ru>*
18141 * The x509 application mishandled signing requests containing DSA
18142 keys when the signing key was also DSA and the parameters didn't match.
18144 It was supposed to omit the parameters when they matched the signing key:
18145 the verifying software was then supposed to automatically use the CA's
18146 parameters if they were absent from the end user certificate.
18148 Omitting parameters is no longer recommended. The test was also
18149 the wrong way round! This was probably due to unusual behaviour in
18150 EVP_cmp_parameters() which returns 1 if the parameters match.
18151 This meant that parameters were omitted when they *didn't* match and
18152 the certificate was useless. Certificates signed with 'ca' didn't have
18155 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18157 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18158 The interface is as follows:
18159 Applications can use
18160 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18161 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18162 "off" is now the default.
18163 The library internally uses
18164 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18165 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18166 to disable memory-checking temporarily.
18168 Some inconsistent states that previously were possible (and were
18169 even the default) are now avoided.
18171 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18172 with each memory chunk allocated; this is occasionally more helpful
18173 than just having a counter.
18175 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18177 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18182 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18183 which largely parallels "options", but is for changing API behaviour,
18184 whereas "options" are about protocol behaviour.
18185 Initial "mode" flags are:
18187 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18188 a single record has been written.
18189 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18190 retries use the same buffer location.
18191 (But all of the contents must be
18196 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18199 * Fix problems with no-hmac etc.
18201 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18203 * New functions RSA_get_default_method(), RSA_set_method() and
18204 RSA_get_method(). These allows replacement of RSA_METHODs without having
18205 to mess around with the internals of an RSA structure.
18209 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18210 Also really enable memory leak checks in openssl.c and in some
18213 *Chad C. Mulligan, Bodo Moeller*
18215 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18216 up the length of negative integers. This has now been simplified to just
18217 store the length when it is first determined and use it later, rather
18218 than trying to keep track of where data is copied and updating it to
18220 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18222 * Add a new function PKCS7_signatureVerify. This allows the verification
18223 of a PKCS#7 signature but with the signing certificate passed to the
18224 function itself. This contrasts with PKCS7_dataVerify which assumes the
18225 certificate is present in the PKCS#7 structure. This isn't always the
18226 case: certificates can be omitted from a PKCS#7 structure and be
18227 distributed by "out of band" means (such as a certificate database).
18231 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18232 function prototypes in pem.h, also change util/mkdef.pl to add the
18233 necessary function names.
18237 * mk1mf.pl (used by Windows builds) did not properly read the
18238 options set by Configure in the top level Makefile, and Configure
18239 was not even able to write more than one option correctly.
18240 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18244 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18245 file to be loaded from a BIO or FILE pointer. The BIO version will
18246 for example allow memory BIOs to contain config info.
18250 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18251 Whoever hopes to achieve shared-library compatibility across versions
18252 must use this, not the compile-time macro.
18253 (Exercise 0.9.4: Which is the minimum library version required by
18255 Note: All this applies only to multi-threaded programs, others don't
18260 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18261 through a BIO pair triggered the default case, i.e.
18262 SSLerr(...,SSL_R_UNKNOWN_STATE).
18266 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18267 can use the SSL library even if none of the specific BIOs is
18272 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18273 for the encoded length.
18275 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18277 * Add initial documentation of the X509V3 functions.
18281 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18282 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18283 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18284 secure PKCS#8 private key format with a high iteration count.
18288 * Fix determination of Perl interpreter: A perl or perl5
18289 *directory* in $PATH was also accepted as the interpreter.
18291 *Ralf S. Engelschall*
18293 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18294 wrong with it but it was very old and did things like calling
18295 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18296 unusual formatting.
18300 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18301 to use the new extension code.
18305 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18306 with macros. This should make it easier to change their form, add extra
18307 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18312 * Add to configuration table a new entry that can specify an alternative
18313 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18314 according to Mark Crispin <MRC@Panda.COM>.
18318 * DES CBC did not update the IV. Weird.
18322 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18323 Changing the behaviour of the former might break existing programs --
18324 where IV updating is needed, des_ncbc_encrypt can be used.
18327 * When bntest is run from "make test" it drives bc to check its
18328 calculations, as well as internally checking them. If an internal check
18329 fails, it needs to cause bc to give a non-zero result or make test carries
18330 on without noticing the failure. Fixed.
18334 * DES library cleanups.
18338 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18339 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18340 ciphers. NOTE: although the key derivation function has been verified
18341 against some published test vectors it has not been extensively tested
18342 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18347 * Instead of "mkdir -p", which is not fully portable, use new
18348 Perl script "util/mkdir-p.pl".
18352 * Rewrite the way password based encryption (PBE) is handled. It used to
18353 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18354 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18355 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18356 the 'parameter' field of the AlgorithmIdentifier is passed to the
18357 underlying key generation function so it must do its own ASN1 parsing.
18358 This has also changed the EVP_PBE_CipherInit() function which now has a
18359 'parameter' argument instead of literal salt and iteration count values
18360 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18364 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18365 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18366 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18367 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18368 value was just used as a "magic string" and not used directly its
18369 value doesn't matter.
18373 * Introduce some semblance of const correctness to BN. Shame C doesn't
18378 * "linux-sparc64" configuration (ultrapenguin).
18380 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18381 "linux-sparc" configuration.
18383 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18385 * config now generates no-xxx options for missing ciphers.
18389 * Support the EBCDIC character set (work in progress).
18390 File ebcdic.c not yet included because it has a different license.
18392 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18394 * Support BS2000/OSD-POSIX.
18396 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18398 * Make callbacks for key generation use `void *` instead of `char *`.
18402 * Make S/MIME samples compile (not yet tested).
18406 * Additional typesafe stacks.
18410 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18414 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18416 * New configuration variant "sco5-gcc".
18418 * Updated some demos.
18420 *Sean O Riordain, Wade Scholine*
18422 * Add missing BIO_free at exit of pkcs12 application.
18426 * Fix memory leak in conf.c.
18430 * Updates for Win32 to assembler version of MD5.
18434 * Set #! path to perl in `apps/der_chop` to where we found it
18435 instead of using a fixed path.
18439 * SHA library changes for irix64-mips4-cc.
18443 * Improvements for VMS support.
18447 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18449 * Bignum library bug fix. IRIX 6 passes "make test" now!
18450 This also avoids the problems with SC4.2 and unpatched SC5.
18452 *Andy Polyakov <appro@fy.chalmers.se>*
18454 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18455 These are required because of the typesafe stack would otherwise break
18456 existing code. If old code used a structure member which used to be STACK
18457 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18458 sk_num or sk_value it would produce an error because the num, data members
18459 are not present in STACK_OF. Now it just produces a warning. sk_set
18460 replaces the old method of assigning a value to sk_value
18461 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18462 that does this will no longer work (and should use sk_set instead) but
18463 this could be regarded as a "questionable" behaviour anyway.
18467 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18468 correctly handle encrypted S/MIME data.
18472 * Change type of various DES function arguments from des_cblock
18473 (which means, in function argument declarations, pointer to char)
18474 to des_cblock * (meaning pointer to array with 8 char elements),
18475 which allows the compiler to do more typechecking; it was like
18476 that back in SSLeay, but with lots of ugly casts.
18478 Introduce new type const_des_cblock.
18482 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18483 problems: find RecipientInfo structure that matches recipient certificate
18484 and initialise the ASN1 structures properly based on passed cipher.
18488 * Belatedly make the BN tests actually check the results.
18492 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18493 to and from BNs: it was completely broken. New compilation option
18494 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18495 key elements as negative integers.
18499 * Reorganize and speed up MD5.
18501 *Andy Polyakov <appro@fy.chalmers.se>*
18505 *Richard Levitte <richard@levitte.org>*
18507 * New option -out to asn1parse to allow the parsed structure to be
18508 output to a file. This is most useful when combined with the -strparse
18509 option to examine the output of things like OCTET STRINGS.
18513 * Make SSL library a little more fool-proof by not requiring any longer
18514 that `SSL_set_{accept,connect}_state` be called before
18515 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18516 in many applications because usually everything *appeared* to work as
18517 intended anyway -- now it really works as intended).
18521 * Move openssl.cnf out of lib/.
18525 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18526 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18527 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18529 *Ralf S. Engelschall*
18531 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18532 handle PKCS#7 enveloped data properly.
18534 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18536 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18537 copying pointers. The cert_st handling is changed by this in
18538 various ways (and thus what used to be known as ctx->default_cert
18539 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18540 any longer when s->cert does not give us what we need).
18541 ssl_cert_instantiate becomes obsolete by this change.
18542 As soon as we've got the new code right (possibly it already is?),
18543 we have solved a couple of bugs of the earlier code where s->cert
18544 was used as if it could not have been shared with other SSL structures.
18546 Note that using the SSL API in certain dirty ways now will result
18547 in different behaviour than observed with earlier library versions:
18548 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18549 does not influence s as it used to.
18551 In order to clean up things more thoroughly, inside SSL_SESSION
18552 we don't use CERT any longer, but a new structure SESS_CERT
18553 that holds per-session data (if available); currently, this is
18554 the peer's certificate chain and, for clients, the server's certificate
18555 and temporary key. CERT holds only those values that can have
18556 meaningful defaults in an SSL_CTX.
18560 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18561 from the internal representation. Various PKCS#7 fixes: remove some
18562 evil casts and set the enc_dig_alg field properly based on the signing
18567 * Allow PKCS#12 password to be set from the command line or the
18568 environment. Let 'ca' get its config file name from the environment
18569 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18574 * Allow certificate policies extension to use an IA5STRING for the
18575 organization field. This is contrary to the PKIX definition but
18576 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18581 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18582 without disallowing inline assembler and the like for non-pedantic builds.
18586 * Support Borland C++ builder.
18588 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18594 * SHA-1 cleanups and performance enhancements.
18596 *Andy Polyakov <appro@fy.chalmers.se>*
18598 * Sparc v8plus assembler for the bignum library.
18600 *Andy Polyakov <appro@fy.chalmers.se>*
18602 * Accept any -xxx and +xxx compiler options in Configure.
18606 * Update HPUX configuration.
18610 * Add missing `sk_<type>_unshift()` function to safestack.h
18612 *Ralf S. Engelschall*
18614 * New function SSL_CTX_use_certificate_chain_file that sets the
18615 "extra_cert"s in addition to the certificate. (This makes sense
18616 only for "PEM" format files, as chains as a whole are not
18621 * Support verify_depth from the SSL API.
18622 x509_vfy.c had what can be considered an off-by-one-error:
18623 Its depth (which was not part of the external interface)
18624 was actually counting the number of certificates in a chain;
18625 now it really counts the depth.
18629 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18630 instead of X509err, which often resulted in confusing error
18631 messages since the error codes are not globally unique
18632 (e.g. an alleged error in ssl3_accept when a certificate
18633 didn't match the private key).
18635 * New function SSL_CTX_set_session_id_context that allows to set a default
18636 value (so that you don't need SSL_set_session_id_context for each
18637 connection using the SSL_CTX).
18641 * OAEP decoding bug fix.
18645 * Support INSTALL_PREFIX for package builders, as proposed by
18650 * New Configure options "threads" and "no-threads". For systems
18651 where the proper compiler options are known (currently Solaris
18652 and Linux), "threads" is the default.
18656 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18660 * Install various scripts to $(OPENSSLDIR)/misc, not to
18661 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18662 such as /usr/local/bin.
18666 * "make linux-shared" to build shared libraries.
18668 *Niels Poppe <niels@netbox.org>*
18670 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18674 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18675 extension adding in x509 utility.
18679 * Remove NOPROTO sections and error code comments.
18683 * Partial rewrite of the DEF file generator to now parse the ANSI
18688 * New Configure options --prefix=DIR and --openssldir=DIR.
18692 * Complete rewrite of the error code script(s). It is all now handled
18693 by one script at the top level which handles error code gathering,
18694 header rewriting and C source file generation. It should be much better
18695 than the old method: it now uses a modified version of Ulf's parser to
18696 read the ANSI prototypes in all header files (thus the old K&R definitions
18697 aren't needed for error creation any more) and do a better job of
18698 translating function codes into names. The old 'ASN1 error code embedded
18699 in a comment' is no longer necessary and it doesn't use .err files which
18700 have now been deleted. Also the error code call doesn't have to appear all
18701 on one line (which resulted in some large lines...).
18705 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18709 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18710 0 (which usually indicates a closed connection), but continue reading.
18714 * Fix some race conditions.
18718 * Add support for CRL distribution points extension. Add Certificate
18719 Policies and CRL distribution points documentation.
18723 * Move the autogenerated header file parts to crypto/opensslconf.h.
18727 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18728 8 of keying material. Merlin has also confirmed interop with this fix
18729 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18731 *Merlin Hughes <merlin@baltimore.ie>*
18733 * Fix lots of warnings.
18735 *Richard Levitte <levitte@stacken.kth.se>*
18737 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18738 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18740 *Richard Levitte <levitte@stacken.kth.se>*
18742 * Fix problems with sizeof(long) == 8.
18744 *Andy Polyakov <appro@fy.chalmers.se>*
18746 * Change functions to ANSI C.
18750 * Fix typos in error codes.
18752 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18754 * Remove defunct assembler files from Configure.
18758 * SPARC v8 assembler BIGNUM implementation.
18760 *Andy Polyakov <appro@fy.chalmers.se>*
18762 * Support for Certificate Policies extension: both print and set.
18763 Various additions to support the r2i method this uses.
18767 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18768 return a const string when you are expecting an allocated buffer.
18772 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18773 types DirectoryString and DisplayText.
18777 * Add code to allow r2i extensions to access the configuration database,
18778 add an LHASH database driver and add several ctx helper functions.
18782 * Fix an evil bug in bn_expand2() which caused various BN functions to
18783 fail when they extended the size of a BIGNUM.
18787 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18788 support typesafe stack.
18792 * Fix typo in SSL_[gs]et_options().
18794 *Nils Frostberg <nils@medcom.se>*
18796 * Delete various functions and files that belonged to the (now obsolete)
18797 old X509V3 handling code.
18801 * New Configure option "rsaref".
18805 * Don't auto-generate pem.h.
18809 * Introduce type-safe ASN.1 SETs.
18813 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18815 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18817 * Introduce type-safe STACKs. This will almost certainly break lots of code
18818 that links with OpenSSL (well at least cause lots of warnings), but fear
18819 not: the conversion is trivial, and it eliminates loads of evil casts. A
18820 few STACKed things have been converted already. Feel free to convert more.
18821 In the fullness of time, I'll do away with the STACK type altogether.
18825 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18826 specified in `<certfile>` by updating the entry in the index.txt file.
18827 This way one no longer has to edit the index.txt file manually for
18828 revoking a certificate. The -revoke option does the gory details now.
18830 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18832 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18833 `-text` option at all and this way the `-noout -text` combination was
18834 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18836 *Ralf S. Engelschall*
18838 * Make sure a corresponding plain text error message exists for the
18839 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18840 verify callback function determined that a certificate was revoked.
18842 *Ralf S. Engelschall*
18844 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18845 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18846 all available ciphers including rc5, which was forgotten until now.
18847 In order to let the testing shell script know which algorithms
18848 are available, a new (up to now undocumented) command
18849 `openssl list-cipher-commands` is used.
18853 * Bugfix: s_client occasionally would sleep in select() when
18854 it should have checked SSL_pending() first.
18858 * New functions DSA_do_sign and DSA_do_verify to provide access to
18859 the raw DSA values prior to ASN.1 encoding.
18863 * Tweaks to Configure
18865 *Niels Poppe <niels@netbox.org>*
18867 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18872 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18876 * New config option to avoid instructions that are illegal on the 80386.
18877 The default code is faster, but requires at least a 486.
18881 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18882 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18883 same as SSL2_VERSION anyway.
18887 * New "-showcerts" option for s_client.
18891 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18892 application. Various cleanups and fixes.
18896 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18897 modify error routines to work internally. Add error codes and PBE init
18898 to library startup routines.
18902 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18903 packing functions to asn1 and evp. Changed function names and error
18904 codes along the way.
18908 * PKCS12 integration: and so it begins... First of several patches to
18909 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18910 objects to objects.h
18914 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18915 and display support for Thawte strong extranet extension.
18919 * Add LinuxPPC support.
18921 *Jeff Dubrule <igor@pobox.org>*
18923 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18924 bn_div_words in alpha.s.
18926 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18928 * Make sure the RSA OAEP test is skipped under -DRSAref because
18929 OAEP isn't supported when OpenSSL is built with RSAref.
18931 *Ulf Moeller <ulf@fitug.de>*
18933 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18934 so they no longer are missing under -DNOPROTO.
18936 *Soren S. Jorvang <soren@t.dk>*
18938 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18940 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18941 doesn't work when the session is reused. Coming soon!
18945 * Fix a security hole, that allows sessions to be reused in the wrong
18946 context thus bypassing client cert protection! All software that uses
18947 client certs and session caches in multiple contexts NEEDS PATCHING to
18948 allow session reuse! A fuller solution is in the works.
18950 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18952 * Some more source tree cleanups (removed obsolete files
18953 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18954 permission on "config" script to be executable) and a fix for the INSTALL
18957 *Ulf Moeller <ulf@fitug.de>*
18959 * Remove some legacy and erroneous uses of malloc, free instead of
18962 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18964 * Make rsa_oaep_test return non-zero on error.
18966 *Ulf Moeller <ulf@fitug.de>*
18968 * Add support for native Solaris shared libraries. Configure
18969 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18970 if someone would make that last step automatic.
18972 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18974 * ctx_size was not built with the right compiler during "make links". Fixed.
18978 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18979 except NULL ciphers". This means the default cipher list will no longer
18980 enable NULL ciphers. They need to be specifically enabled e.g. with
18981 the string "DEFAULT:eNULL".
18985 * Fix to RSA private encryption routines: if p < q then it would
18986 occasionally produce an invalid result. This will only happen with
18987 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18991 * Be less restrictive and allow also `perl util/perlpath.pl
18992 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18993 because this way one can also use an interpreter named `perl5` (which is
18994 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18995 installed as `perl`).
18997 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18999 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
19001 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19003 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
19004 advapi32.lib to Win32 build and change the pem test comparison
19005 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
19006 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
19007 and crypto/des/ede_cbcm_enc.c.
19011 * DES quad checksum was broken on big-endian architectures. Fixed.
19015 * Comment out two functions in bio.h that aren't implemented. Fix up the
19016 Win32 test batch file so it (might) work again. The Win32 test batch file
19017 is horrible: I feel ill....
19021 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
19022 in e_os.h. Audit of header files to check ANSI and non ANSI
19023 sections: 10 functions were absent from non ANSI section and not exported
19024 from Windows DLLs. Fixed up libeay.num for new functions.
19028 * Make `openssl version` output lines consistent.
19030 *Ralf S. Engelschall*
19032 * Fix Win32 symbol export lists for BIO functions: Added
19033 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
19034 to ms/libeay{16,32}.def.
19036 *Ralf S. Engelschall*
19038 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
19039 fine under Unix and passes some trivial tests I've now added. But the
19040 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
19041 added to make sure no one expects that this stuff really works in the
19042 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
19043 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
19046 *Ralf S. Engelschall*
19048 * Fix the generation of two part addresses in perl.
19050 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
19052 * Add config entry for Linux on MIPS.
19054 *John Tobey <jtobey@channel1.com>*
19056 * Make links whenever Configure is run, unless we are on Windoze.
19060 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
19061 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19066 * Add a useful kludge to allow package maintainers to specify compiler and
19067 other platforms details on the command line without having to patch the
19068 Configure script every time: One now can use
19069 `perl Configure <id>:<details>`,
19070 i.e. platform ids are allowed to have details appended
19071 to them (separated by colons). This is treated as there would be a static
19072 pre-configured entry in Configure's %table under key `<id>` with value
19073 `<details>` and `perl Configure <id>` is called. So, when you want to
19074 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19075 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19076 now, which overrides the FreeBSD-elf entry on-the-fly.
19078 *Ralf S. Engelschall*
19080 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19084 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19085 on the `perl Configure ...` command line. This way one can compile
19086 OpenSSL libraries with Position Independent Code (PIC) which is needed
19087 for linking it into DSOs.
19089 *Ralf S. Engelschall*
19091 * Remarkably, export ciphers were totally broken and no-one had noticed!
19096 * Cleaned up the LICENSE document: The official contact for any license
19097 questions now is the OpenSSL core team under openssl-core@openssl.org.
19098 And add a paragraph about the dual-license situation to make sure people
19099 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19100 to the OpenSSL toolkit.
19102 *Ralf S. Engelschall*
19104 * General source tree makefile cleanups: Made `making xxx in yyy...`
19105 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19106 Additionally cleaned up the `make links` target: Remove unnecessary
19107 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19108 to speed processing and no longer clutter the display with confusing
19109 stuff. Instead only the actually done links are displayed.
19111 *Ralf S. Engelschall*
19113 * Permit null encryption ciphersuites, used for authentication only. It used
19114 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19115 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19120 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19121 signed attributes when verifying signatures (this would break them),
19122 the detached data encoding was wrong and public keys obtained using
19123 X509_get_pubkey() weren't freed.
19127 * Add text documentation for the BUFFER functions. Also added a work around
19128 to a Win95 console bug. This was triggered by the password read stuff: the
19129 last character typed gets carried over to the next fread(). If you were
19130 generating a new cert request using 'req' for example then the last
19131 character of the passphrase would be CR which would then enter the first
19136 * Added the new 'Includes OpenSSL Cryptography Software' button as
19137 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19138 button and can be used by applications based on OpenSSL to show the
19139 relationship to the OpenSSL project.
19141 *Ralf S. Engelschall*
19143 * Remove confusing variables in function signatures in files
19144 ssl/ssl_lib.c and ssl/ssl.h.
19146 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19148 * Don't install bss_file.c under PREFIX/include/
19150 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19152 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19153 functions that return function pointers and has support for NT specific
19154 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19155 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19156 unsigned to signed types: this was killing the Win32 compile.
19160 * Add new certificate file to stack functions,
19161 SSL_add_dir_cert_subjects_to_stack() and
19162 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19163 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19164 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19165 This means that Apache-SSL and similar packages don't have to mess around
19166 to add as many CAs as they want to the preferred list.
19170 * Experiment with doxygen documentation. Currently only partially applied to
19172 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19173 openssl.doxy as the configuration file.
19177 * Get rid of remaining C++-style comments which strict C compilers hate.
19179 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19181 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19182 compiled in by default: it has problems with large keys.
19186 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19187 DH private keys and/or callback functions which directly correspond to
19188 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19189 is needed for applications which have to configure certificates on a
19190 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19192 For the RSA certificate situation is makes no difference, but
19193 for the DSA certificate situation this fixes the "no shared cipher"
19194 problem where the OpenSSL cipher selection procedure failed because the
19195 temporary keys were not overtaken from the context and the API provided
19196 no way to reconfigure them.
19197 The new functions now let applications reconfigure the stuff and they
19198 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19199 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19200 non-public-API function ssl_cert_instantiate() is used as a helper
19201 function and also to reduce code redundancy inside ssl_rsa.c.
19203 *Ralf S. Engelschall*
19205 * Move s_server -dcert and -dkey options out of the undocumented feature
19206 area because they are useful for the DSA situation and should be
19207 recognized by the users.
19209 *Ralf S. Engelschall*
19211 * Fix the cipher decision scheme for export ciphers: the export bits are
19212 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19213 SSL_EXP_MASK. So, the original variable has to be used instead of the
19214 already masked variable.
19216 *Richard Levitte <levitte@stacken.kth.se>*
19218 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19220 *Richard Levitte <levitte@stacken.kth.se>*
19222 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19223 from `int` to `unsigned int` because it is a length and initialized by
19224 EVP_DigestFinal() which expects an `unsigned int *`.
19226 *Richard Levitte <levitte@stacken.kth.se>*
19228 * Don't hard-code path to Perl interpreter on shebang line of Configure
19229 script. Instead use the usual Shell->Perl transition trick.
19231 *Ralf S. Engelschall*
19233 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19234 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19235 -noout -modulus` as it's already the case for `openssl rsa -noout
19236 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19237 currently the public key is printed (a decision which was already done by
19238 `openssl dsa -modulus` in the past) which serves a similar purpose.
19239 Additionally the NO_RSA no longer completely removes the whole -modulus
19240 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19243 *Ralf S. Engelschall*
19245 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19246 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19248 *Arne Ansper <arne@ats.cyber.ee>*
19250 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19251 to be added. Now both 'req' and 'ca' can use new objects defined in the
19256 * Add cool BIO that does syslog (or event log on NT).
19258 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19260 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19261 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19262 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19263 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19267 * Add preliminary config info for new extension code.
19271 * Make RSA_NO_PADDING really use no padding.
19273 *Ulf Moeller <ulf@fitug.de>*
19275 * Generate errors when private/public key check is done.
19279 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19280 for some CRL extensions and new objects added.
19284 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19285 key usage extension and fuller support for authority key id.
19289 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19290 padding method for RSA, which is recommended for new applications in PKCS
19291 #1 v2.0 (RFC 2437, October 1998).
19292 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19293 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19294 against Bleichbacher's attack on RSA.
19295 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19298 * Updates to the new SSL compression code
19300 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19302 * Fix so that the version number in the master secret, when passed
19303 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19304 (because the server will not accept higher), that the version number
19305 is 0x03,0x01, not 0x03,0x00
19307 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19309 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19310 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19311 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19315 * Support for RAW extensions where an arbitrary extension can be
19316 created by including its DER encoding. See `apps/openssl.cnf` for
19321 * Make sure latest Perl versions don't interpret some generated C array
19322 code as Perl array code in the crypto/err/err_genc.pl script.
19324 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19326 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19327 not many people have the assembler. Various Win32 compilation fixes and
19328 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19329 build instructions.
19333 * Modify configure script 'Configure' to automatically create crypto/date.h
19334 file under Win32 and also build pem.h from pem.org. New script
19335 util/mkfiles.pl to create the MINFO file on environments that can't do a
19336 'make files': perl util/mkfiles.pl >MINFO should work.
19340 * Major rework of DES function declarations, in the pursuit of correctness
19341 and purity. As a result, many evil casts evaporated, and some weirdness,
19342 too. You may find this causes warnings in your code. Zapping your evil
19343 casts will probably fix them. Mostly.
19347 * Fix for a typo in asn1.h. Bug fix to object creation script
19348 obj_dat.pl. It considered a zero in an object definition to mean
19349 "end of object": none of the objects in objects.h have any zeros
19350 so it wasn't spotted.
19352 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19354 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19355 Masking (CBCM). In the absence of test vectors, the best I have been able
19356 to do is check that the decrypt undoes the encrypt, so far. Send me test
19357 vectors if you have them.
19361 * Correct calculation of key length for export ciphers (too much space was
19362 allocated for null ciphers). This has not been tested!
19366 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19367 message is now correct (it understands "crypto" and "ssl" on its
19368 command line). There is also now an "update" option. This will update
19369 the util/ssleay.num and util/libeay.num files with any new functions.
19371 perl util/mkdef.pl crypto ssl update
19372 it will update them.
19376 * Overhauled the Perl interface:
19377 - ported BN stuff to OpenSSL's different BN library
19378 - made the perl/ source tree CVS-aware
19379 - renamed the package from SSLeay to OpenSSL (the files still contain
19380 their history because I've copied them in the repository)
19381 - removed obsolete files (the test scripts will be replaced
19382 by better Test::Harness variants in the future)
19384 *Ralf S. Engelschall*
19386 * First cut for a very conservative source tree cleanup:
19387 1. merge various obsolete readme texts into doc/ssleay.txt
19388 where we collect the old documents and readme texts.
19389 2. remove the first part of files where I'm already sure that we no
19390 longer need them because of three reasons: either they are just temporary
19391 files which were left by Eric or they are preserved original files where
19392 I've verified that the diff is also available in the CVS via "cvs diff
19393 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19394 the crypto/md/ stuff).
19396 *Ralf S. Engelschall*
19398 * More extension code. Incomplete support for subject and issuer alt
19399 name, issuer and authority key id. Change the i2v function parameters
19400 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19401 what that's for :-) Fix to ASN1 macro which messed up
19402 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19406 * Preliminary support for ENUMERATED type. This is largely copied from the
19411 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19413 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19415 * Make sure `make rehash` target really finds the `openssl` program.
19417 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19419 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19420 like to hear about it if this slows down other processors.
19424 * Add CygWin32 platform information to Configure script.
19426 *Alan Batie <batie@aahz.jf.intel.com>*
19428 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19430 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19432 * New program nseq to manipulate netscape certificate sequences
19436 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19441 * Fixes to BN code. Previously the default was to define BN_RECURSION
19442 but the BN code had some problems that would cause failures when
19443 doing certificate verification and some other functions.
19445 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19447 * Add ASN1 and PEM code to support netscape certificate sequences.
19451 * Add ASN1 and PEM code to support netscape certificate sequences.
19455 * Add several PKIX and private extended key usage OIDs.
19459 * Modify the 'ca' program to handle the new extension code. Modify
19460 openssl.cnf for new extension format, add comments.
19464 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19465 and add a sample to openssl.cnf so req -x509 now adds appropriate
19470 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19471 error code, add initial support to X509_print() and x509 application.
19475 * Takes a deep breath and start adding X509 V3 extension support code. Add
19476 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19477 stuff is currently isolated and isn't even compiled yet.
19481 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19482 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19483 Removed the versions check from X509 routines when loading extensions:
19484 this allows certain broken certificates that don't set the version
19485 properly to be processed.
19489 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19490 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19491 can still be regenerated with "make depend".
19495 * Spelling mistake in C version of CAST-128.
19497 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19499 * Changes to the error generation code. The perl script err-code.pl
19500 now reads in the old error codes and retains the old numbers, only
19501 adding new ones if necessary. It also only changes the .err files if new
19502 codes are added. The makefiles have been modified to only insert errors
19503 when needed (to avoid needlessly modifying header files). This is done
19504 by only inserting errors if the .err file is newer than the auto generated
19505 C file. To rebuild all the error codes from scratch (the old behaviour)
19506 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19507 or delete all the .err files.
19511 * CAST-128 was incorrectly implemented for short keys. The C version has
19512 been fixed, but is untested. The assembler versions are also fixed, but
19513 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19514 to regenerate it if needed.
19515 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19516 Hagino <itojun@kame.net>*
19518 * File was opened incorrectly in randfile.c.
19520 *Ulf Möller <ulf@fitug.de>*
19522 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19523 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19524 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19525 al: it's just almost always a UTCTime. Note this patch adds new error
19526 codes so do a "make errors" if there are problems.
19530 * Correct Linux 1 recognition in config.
19532 *Ulf Möller <ulf@fitug.de>*
19534 * Remove pointless MD5 hash when using DSA keys in ca.
19536 *Anonymous <nobody@replay.com>*
19538 * Generate an error if given an empty string as a cert directory. Also
19539 generate an error if handed NULL (previously returned 0 to indicate an
19540 error, but didn't set one).
19542 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19544 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19548 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19549 parameters. This was causing a warning which killed off the Win32 compile.
19553 * Remove C++ style comments from crypto/bn/bn_local.h.
19555 *Neil Costigan <neil.costigan@celocom.com>*
19557 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19558 based on a text string, looking up short and long names and finally
19559 "dot" format. The "dot" format stuff didn't work. Added new function
19560 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19561 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19562 OID is not part of the table.
19566 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19567 X509_LOOKUP_by_alias().
19571 * Sort openssl functions by name.
19575 * Get the `gendsa` command working and add it to the `list` command. Remove
19576 encryption from sample DSA keys (in case anyone is interested the password
19581 * Make *all* `*_free` functions accept a NULL pointer.
19583 *Frans Heymans <fheymans@isaserver.be>*
19585 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19588 *Anonymous <nobody@replay.com>*
19590 * s_server should send the CAfile as acceptable CAs, not its own cert.
19592 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19594 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19596 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19598 * Temp key "for export" tests were wrong in s3_srvr.c.
19600 *Anonymous <nobody@replay.com>*
19602 * Add prototype for temp key callback functions
19603 SSL_CTX_set_tmp_{rsa,dh}_callback().
19607 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19608 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19612 * X509_name_add_entry() freed the wrong thing after an error.
19614 *Arne Ansper <arne@ats.cyber.ee>*
19616 * rsa_eay.c would attempt to free a NULL context.
19618 *Arne Ansper <arne@ats.cyber.ee>*
19620 * BIO_s_socket() had a broken should_retry() on Windoze.
19622 *Arne Ansper <arne@ats.cyber.ee>*
19624 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19626 *Arne Ansper <arne@ats.cyber.ee>*
19628 * Make sure the already existing X509_STORE->depth variable is initialized
19629 in X509_STORE_new(), but document the fact that this variable is still
19630 unused in the certificate verification process.
19632 *Ralf S. Engelschall*
19634 * Fix the various library and `apps/` files to free up pkeys obtained from
19635 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19639 * Fix reference counting in X509_PUBKEY_get(). This makes
19640 demos/maurice/example2.c work, amongst others, probably.
19642 *Steve Henson and Ben Laurie*
19644 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19645 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19646 are no longer created. This way we have a single and consistent command
19647 line interface `openssl <command>`, similar to `cvs <command>`.
19649 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19651 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19652 BIT STRING wrapper always have zero unused bits.
19656 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19660 * Make the top-level INSTALL documentation easier to understand.
19664 * Makefiles updated to exit if an error occurs in a sub-directory
19665 make (including if user presses ^C) [Paul Sutton]
19667 * Make Montgomery context stuff explicit in RSA data structure.
19671 * Fix build order of pem and err to allow for generated pem.h.
19675 * Fix renumbering bug in X509_NAME_delete_entry().
19679 * Enhanced the err-ins.pl script so it makes the error library number
19680 global and can add a library name. This is needed for external ASN1 and
19681 other error libraries.
19685 * Fixed sk_insert which never worked properly.
19689 * Fix ASN1 macros so they can handle indefinite length constructed
19690 EXPLICIT tags. Some non standard certificates use these: they can now
19695 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19696 into a single doc/ssleay.txt bundle. This way the information is still
19697 preserved but no longer messes up this directory. Now it's new room for
19698 the new set of documentation files.
19700 *Ralf S. Engelschall*
19702 * SETs were incorrectly DER encoded. This was a major pain, because they
19703 shared code with SEQUENCEs, which aren't coded the same. This means that
19704 almost everything to do with SETs or SEQUENCEs has either changed name or
19705 number of arguments.
19707 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19709 * Fix test data to work with the above.
19713 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19714 was already fixed by Eric for 0.9.1 it seems.
19716 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19718 * Autodetect FreeBSD3.
19722 * Fix various bugs in Configure. This affects the following platforms:
19726 unixware-2.0-pentium
19731 * Eliminate generated files from CVS. Reorder tests to regenerate files
19732 before they are needed.
19736 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19740 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19742 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19743 changed SSLeay to OpenSSL in version strings.
19745 *Ralf S. Engelschall*
19747 * Some fixups to the top-level documents.
19751 * Fixed the nasty bug where rsaref.h was not found under compile-time
19752 because the symlink to include/ was missing.
19754 *Ralf S. Engelschall*
19756 * Incorporated the popular no-RSA/DSA-only patches
19757 which allow to compile an RSA-free SSLeay.
19759 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19761 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19762 when "ssleay" is still not found.
19764 *Ralf S. Engelschall*
19766 * Added more platforms to Configure: Cray T3E, HPUX 11,
19768 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19770 * Updated the README file.
19772 *Ralf S. Engelschall*
19774 * Added various .cvsignore files in the CVS repository subdirs
19775 to make a "cvs update" really silent.
19777 *Ralf S. Engelschall*
19779 * Recompiled the error-definition header files and added
19780 missing symbols to the Win32 linker tables.
19782 *Ralf S. Engelschall*
19784 * Cleaned up the top-level documents;
19785 o new files: CHANGES and LICENSE
19786 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19787 o merged COPYRIGHT into LICENSE
19788 o removed obsolete TODO file
19789 o renamed MICROSOFT to INSTALL.W32
19791 *Ralf S. Engelschall*
19793 * Removed dummy files from the 0.9.1b source tree:
19794 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19795 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19796 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19797 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19798 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19800 *Ralf S. Engelschall*
19802 * Added various platform portability fixes.
19806 * The Genesis of the OpenSSL rpject:
19807 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19808 Young and Tim J. Hudson created while they were working for C2Net until
19811 *The OpenSSL Project*
19813 ### Changes between 0.9.0b and 0.9.1b [not released]
19815 * Updated a few CA certificates under certs/
19819 * Changed some BIGNUM api stuff.
19823 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19824 DGUX x86, Linux Alpha, etc.
19828 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19829 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19834 * Add -strparse option to asn1pars program which parses nested
19837 *Dr Stephen Henson <shenson@bigfoot.com>*
19839 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19843 * DSA fix for "ca" program.
19847 * Added "-genkey" option to "dsaparam" program.
19851 * Added RIPE MD160 (rmd160) message digest.
19855 * Added -a (all) option to "ssleay version" command.
19859 * Added PLATFORM define which is the id given to Configure.
19863 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19867 * Extended the ASN.1 parser routines.
19871 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19875 * Added a BN_CTX to the BN library.
19879 * Fixed the weak key values in DES library
19883 * Changed API in EVP library for cipher aliases.
19887 * Added support for RC2/64bit cipher.
19891 * Converted the lhash library to the crypto/mem.c functions.
19895 * Added more recognized ASN.1 object ids.
19899 * Added more RSA padding checks for SSL/TLS.
19903 * Added BIO proxy/filter functionality.
19907 * Added extra_certs to SSL_CTX which can be used
19908 send extra CA certificates to the client in the CA cert chain sending
19909 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19913 * Now Fortezza is denied in the authentication phase because
19914 this is key exchange mechanism is not supported by SSLeay at all.
19918 * Additional PKCS1 checks.
19922 * Support the string "TLSv1" for all TLS v1 ciphers.
19926 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19927 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19931 * Fixed a few memory leaks.
19935 * Fixed various code and comment typos.
19939 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19940 bytes sent in the client random.
19942 *Edward Bishop <ebishop@spyglass.com>*
19946 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
19947 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
19948 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
19949 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
19950 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
19951 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
19952 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
19953 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
19954 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
19955 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
19956 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
19957 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
19958 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19959 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
19960 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19961 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19962 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19963 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19964 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19965 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19966 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19967 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19968 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19969 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19970 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19971 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19972 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19973 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19974 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19975 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19976 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19977 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19978 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19979 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19980 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19981 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19982 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19983 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19984 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19985 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19986 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19987 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19988 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19989 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19990 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19991 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19992 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19993 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19994 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19995 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19996 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19997 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19998 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19999 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
20000 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
20001 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
20002 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
20003 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
20004 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
20005 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
20006 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
20007 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
20008 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
20009 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
20010 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
20011 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
20012 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
20013 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
20014 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
20015 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
20016 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
20017 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
20018 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
20019 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
20020 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
20021 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
20022 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
20023 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
20024 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
20025 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
20026 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
20027 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
20028 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
20029 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
20030 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
20031 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
20032 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
20033 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
20034 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
20035 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
20036 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
20037 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
20038 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
20039 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
20040 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
20041 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
20042 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
20043 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
20044 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
20045 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
20046 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
20047 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
20048 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
20049 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
20050 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
20051 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
20052 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
20053 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20054 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20055 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20056 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20057 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20058 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20059 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20060 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20061 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20062 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20063 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20064 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20065 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20066 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20067 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20068 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20069 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20070 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20071 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20072 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20073 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20074 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20075 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20076 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20077 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20078 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20079 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20080 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20081 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20082 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20083 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20084 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20085 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20086 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20087 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20088 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20089 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20090 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20091 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20092 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20093 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20094 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20095 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20096 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20097 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20098 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20099 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20100 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20101 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20102 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20103 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20104 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20105 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20106 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20107 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20108 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20109 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20110 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20111 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20112 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20113 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20114 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20115 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20116 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20117 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20118 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20119 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20120 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20121 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655